Split (1)

train · 174k rows

Search is not available for this dataset

id string	submitter string	authors string	title string	comments string	journal-ref string	doi string	report-no string	categories string	license string	abstract string	versions list	update_date timestamp[s]	authors_parsed sequence	prompt string
2403.16526	Haiqiao Wang	Haiqiao Wang, Zhuoyuan Wang, Dong Ni, Yi Wang	ModeTv2: GPU-accelerated Motion Decomposition Transformer for Pairwise Optimization in Medical Image Registration	null	null	null	null	cs.CV	http://creativecommons.org/licenses/by-sa/4.0/	Deformable image registration plays a crucial role in medical imaging, aiding in disease diagnosis and image-guided interventions. Traditional iterative methods are slow, while deep learning (DL) accelerates solutions but faces usability and precision challenges. This study introduces a pyramid network with the enhanced motion decomposition Transformer (ModeTv2) operator, showcasing superior pairwise optimization (PO) akin to traditional methods. We re-implement ModeT operator with CUDA extensions to enhance its computational efficiency. We further propose RegHead module which refines deformation fields, improves the realism of deformation and reduces parameters. By adopting the PO, the proposed network balances accuracy, efficiency, and generalizability. Extensive experiments on three public brain MRI datasets and one abdominal CT dataset demonstrate the network's suitability for PO, providing a DL model with enhanced usability and interpretability. The code is publicly available at https://github.com/ZAX130/ModeTv2.	[ { "version": "v1", "created": "Mon, 25 Mar 2024 08:09:22 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 01:56:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Haiqiao", "" ], [ "Wang", "Zhuoyuan", "" ], [ "Ni", "Dong", "" ], [ "Wang", "Yi", "" ] ]	TITLE: ModeTv2: GPU-accelerated Motion Decomposition Transformer for Pairwise Optimization in Medical Image Registration ABSTRACT: Deformable image registration plays a crucial role in medical imaging, aiding in disease diagnosis and image-guided interventions. Traditional iterative methods are slow, while deep learning (DL) accelerates solutions but faces usability and precision challenges. This study introduces a pyramid network with the enhanced motion decomposition Transformer (ModeTv2) operator, showcasing superior pairwise optimization (PO) akin to traditional methods. We re-implement ModeT operator with CUDA extensions to enhance its computational efficiency. We further propose RegHead module which refines deformation fields, improves the realism of deformation and reduces parameters. By adopting the PO, the proposed network balances accuracy, efficiency, and generalizability. Extensive experiments on three public brain MRI datasets and one abdominal CT dataset demonstrate the network's suitability for PO, providing a DL model with enhanced usability and interpretability. The code is publicly available at https://github.com/ZAX130/ModeTv2.
2403.18771	Yukyung Lee	Yukyung Lee, Joonghoon Kim, Jaehee Kim, Hyowon Cho, Jaewook Kang, Pilsung Kang, Najoung Kim	CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklists	Extended version currently under review (Workshop version: HEAL at CHI 2024)	null	null	null	cs.CL	http://creativecommons.org/licenses/by-nc-nd/4.0/	Existing LLM-as-a-Judge approaches for evaluating text generation suffer from rating inconsistencies, with low agreement and high rating variance across different evaluator models. We attribute this to subjective evaluation criteria combined with Likert scale scoring in existing protocols. To address this issue, we introduce CheckEval, a checklist-based evaluation framework that improves rating reliability via decomposed binary questions. Through experiments with 12 evaluator models across multiple datasets, we first demonstrate that CheckEval strongly correlates with human judgments, improving the average correlation with human judgments by 0.10. More importantly, CheckEval dramatically improves the average agreement across evaluator models by 0.45 and reduces the score variance. CheckEval scores furthermore have the benefit of being more interpretable because it decomposes evaluation criteria into traceable binary decisions, allowing analyses of specific attributes driving quality judgments.	[ { "version": "v1", "created": "Wed, 27 Mar 2024 17:20:39 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 00:07:06 GMT" } ]	2025-03-18T00:00:00	[ [ "Lee", "Yukyung", "" ], [ "Kim", "Joonghoon", "" ], [ "Kim", "Jaehee", "" ], [ "Cho", "Hyowon", "" ], [ "Kang", "Jaewook", "" ], [ "Kang", "Pilsung", "" ], [ "Kim", "Najoung", "" ] ]	TITLE: CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklists ABSTRACT: Existing LLM-as-a-Judge approaches for evaluating text generation suffer from rating inconsistencies, with low agreement and high rating variance across different evaluator models. We attribute this to subjective evaluation criteria combined with Likert scale scoring in existing protocols. To address this issue, we introduce CheckEval, a checklist-based evaluation framework that improves rating reliability via decomposed binary questions. Through experiments with 12 evaluator models across multiple datasets, we first demonstrate that CheckEval strongly correlates with human judgments, improving the average correlation with human judgments by 0.10. More importantly, CheckEval dramatically improves the average agreement across evaluator models by 0.45 and reduces the score variance. CheckEval scores furthermore have the benefit of being more interpretable because it decomposes evaluation criteria into traceable binary decisions, allowing analyses of specific attributes driving quality judgments.
2404.00521	Yao Ni	Yao Ni, Piotr Koniusz	CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization	Accepted by CVPR 2024. 26 pages. Code: https://github.com/MaxwellYaoNi/CHAIN	null	null	null	cs.LG cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Generative Adversarial Networks (GANs) significantly advanced image generation but their performance heavily depends on abundant training data. In scenarios with limited data, GANs often struggle with discriminator overfitting and unstable training. Batch Normalization (BN), despite being known for enhancing generalization and training stability, has rarely been used in the discriminator of Data-Efficient GANs. Our work addresses this gap by identifying a critical flaw in BN: the tendency for gradient explosion during the centering and scaling steps. To tackle this issue, we present CHAIN (lipsCHitz continuity constrAIned Normalization), which replaces the conventional centering step with zero-mean regularization and integrates a Lipschitz continuity constraint in the scaling step. CHAIN further enhances GAN training by adaptively interpolating the normalized and unnormalized features, effectively avoiding discriminator overfitting. Our theoretical analyses firmly establishes CHAIN's effectiveness in reducing gradients in latent features and weights, improving stability and generalization in GAN training. Empirical evidence supports our theory. CHAIN achieves state-of-the-art results in data-limited scenarios on CIFAR-10/100, ImageNet, five low-shot and seven high-resolution few-shot image datasets. Code: https://github.com/MaxwellYaoNi/CHAIN	[ { "version": "v1", "created": "Sun, 31 Mar 2024 01:41:36 GMT" }, { "version": "v2", "created": "Tue, 2 Apr 2024 07:15:34 GMT" }, { "version": "v3", "created": "Sun, 7 Apr 2024 15:04:47 GMT" }, { "version": "v4", "created": "Sat, 1 Jun 2024 16:22:54 GMT" }, { "version": "v5", "created": "Sat, 2 Nov 2024 03:14:15 GMT" }, { "version": "v6", "created": "Sat, 15 Mar 2025 06:11:42 GMT" } ]	2025-03-18T00:00:00	[ [ "Ni", "Yao", "" ], [ "Koniusz", "Piotr", "" ] ]	TITLE: CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization ABSTRACT: Generative Adversarial Networks (GANs) significantly advanced image generation but their performance heavily depends on abundant training data. In scenarios with limited data, GANs often struggle with discriminator overfitting and unstable training. Batch Normalization (BN), despite being known for enhancing generalization and training stability, has rarely been used in the discriminator of Data-Efficient GANs. Our work addresses this gap by identifying a critical flaw in BN: the tendency for gradient explosion during the centering and scaling steps. To tackle this issue, we present CHAIN (lipsCHitz continuity constrAIned Normalization), which replaces the conventional centering step with zero-mean regularization and integrates a Lipschitz continuity constraint in the scaling step. CHAIN further enhances GAN training by adaptively interpolating the normalized and unnormalized features, effectively avoiding discriminator overfitting. Our theoretical analyses firmly establishes CHAIN's effectiveness in reducing gradients in latent features and weights, improving stability and generalization in GAN training. Empirical evidence supports our theory. CHAIN achieves state-of-the-art results in data-limited scenarios on CIFAR-10/100, ImageNet, five low-shot and seven high-resolution few-shot image datasets. Code: https://github.com/MaxwellYaoNi/CHAIN
2404.05583	Yue-Hua Han	Yue-Hua Han, Tai-Ming Huang, Kai-Lung Hua, Jun-Cheng Chen	Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model	null	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Generative models have enabled the creation of highly realistic facial-synthetic images, raising significant concerns due to their potential for misuse. Despite rapid advancements in the field of deepfake detection, developing efficient approaches to leverage foundation models for improved generalizability to unseen forgery samples remains challenging. To address this challenge, we propose a novel side-network-based decoder that extracts spatial and temporal cues using the CLIP image encoder for generalized video-based Deepfake detection. Additionally, we introduce Facial Component Guidance (FCG) to enhance spatial learning generalizability by encouraging the model to focus on key facial regions. By leveraging the generic features of a vision-language foundation model, our approach demonstrates promising generalizability on challenging Deepfake datasets while also exhibiting superiority in training data efficiency, parameter efficiency, and model robustness.	[ { "version": "v1", "created": "Mon, 8 Apr 2024 14:58:52 GMT" }, { "version": "v2", "created": "Wed, 5 Jun 2024 06:29:37 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 17:10:58 GMT" } ]	2025-03-18T00:00:00	[ [ "Han", "Yue-Hua", "" ], [ "Huang", "Tai-Ming", "" ], [ "Hua", "Kai-Lung", "" ], [ "Chen", "Jun-Cheng", "" ] ]	TITLE: Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model ABSTRACT: Generative models have enabled the creation of highly realistic facial-synthetic images, raising significant concerns due to their potential for misuse. Despite rapid advancements in the field of deepfake detection, developing efficient approaches to leverage foundation models for improved generalizability to unseen forgery samples remains challenging. To address this challenge, we propose a novel side-network-based decoder that extracts spatial and temporal cues using the CLIP image encoder for generalized video-based Deepfake detection. Additionally, we introduce Facial Component Guidance (FCG) to enhance spatial learning generalizability by encouraging the model to focus on key facial regions. By leveraging the generic features of a vision-language foundation model, our approach demonstrates promising generalizability on challenging Deepfake datasets while also exhibiting superiority in training data efficiency, parameter efficiency, and model robustness.
2404.10757	Yuyang Li	Yu-Yang Li, Yu Bai, Cunshi Wang, Mengwei Qu, Ziteng Lu, Roberto Soria, Jifeng Liu	Deep Learning and LLM-based Methods Applied to Stellar Lightcurve Classification	35 pages, 20 figures	Intell Comput. 2025;4:0110	10.34133/icomputing.0110	null	astro-ph.IM astro-ph.SR cs.CL cs.LG	http://creativecommons.org/licenses/by/4.0/	Light curves serve as a valuable source of information on stellar formation and evolution. With the rapid advancement of machine learning techniques, it can be effectively processed to extract astronomical patterns and information. In this study, we present a comprehensive evaluation of deep-learning and large language model (LLM) based models for the automatic classification of variable star light curves, based on large datasets from the Kepler and K2 missions. Special emphasis is placed on Cepheids, RR Lyrae, and eclipsing binaries, examining the influence of observational cadence and phase distribution on classification precision. Employing AutoDL optimization, we achieve striking performance with the 1D-Convolution+BiLSTM architecture and the Swin Transformer, hitting accuracies of 94\% and 99\% correspondingly, with the latter demonstrating a notable 83\% accuracy in discerning the elusive Type II Cepheids-comprising merely 0.02\% of the total dataset.We unveil StarWhisper LightCurve (LC), an innovative Series comprising three LLM-based models: LLM, multimodal large language model (MLLM), and Large Audio Language Model (LALM). Each model is fine-tuned with strategic prompt engineering and customized training methods to explore the emergent abilities of these models for astronomical data. Remarkably, StarWhisper LC Series exhibit high accuracies around 90\%, significantly reducing the need for explicit feature engineering, thereby paving the way for streamlined parallel data processing and the progression of multifaceted multimodal models in astronomical applications. The study furnishes two detailed catalogs illustrating the impacts of phase and sampling intervals on deep learning classification accuracy, showing that a substantial decrease of up to 14\% in observation duration and 21\% in sampling points can be realized without compromising accuracy by more than 10\%.	[ { "version": "v1", "created": "Tue, 16 Apr 2024 17:35:25 GMT" }, { "version": "v2", "created": "Mon, 24 Feb 2025 00:25:01 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Yu-Yang", "" ], [ "Bai", "Yu", "" ], [ "Wang", "Cunshi", "" ], [ "Qu", "Mengwei", "" ], [ "Lu", "Ziteng", "" ], [ "Soria", "Roberto", "" ], [ "Liu", "Jifeng", "" ] ]	TITLE: Deep Learning and LLM-based Methods Applied to Stellar Lightcurve Classification ABSTRACT: Light curves serve as a valuable source of information on stellar formation and evolution. With the rapid advancement of machine learning techniques, it can be effectively processed to extract astronomical patterns and information. In this study, we present a comprehensive evaluation of deep-learning and large language model (LLM) based models for the automatic classification of variable star light curves, based on large datasets from the Kepler and K2 missions. Special emphasis is placed on Cepheids, RR Lyrae, and eclipsing binaries, examining the influence of observational cadence and phase distribution on classification precision. Employing AutoDL optimization, we achieve striking performance with the 1D-Convolution+BiLSTM architecture and the Swin Transformer, hitting accuracies of 94\% and 99\% correspondingly, with the latter demonstrating a notable 83\% accuracy in discerning the elusive Type II Cepheids-comprising merely 0.02\% of the total dataset.We unveil StarWhisper LightCurve (LC), an innovative Series comprising three LLM-based models: LLM, multimodal large language model (MLLM), and Large Audio Language Model (LALM). Each model is fine-tuned with strategic prompt engineering and customized training methods to explore the emergent abilities of these models for astronomical data. Remarkably, StarWhisper LC Series exhibit high accuracies around 90\%, significantly reducing the need for explicit feature engineering, thereby paving the way for streamlined parallel data processing and the progression of multifaceted multimodal models in astronomical applications. The study furnishes two detailed catalogs illustrating the impacts of phase and sampling intervals on deep learning classification accuracy, showing that a substantial decrease of up to 14\% in observation duration and 21\% in sampling points can be realized without compromising accuracy by more than 10\%.
2404.15786	Christian Ledig	Sebastian Doerrich, Francesco Di Salvo, Julius Brockmann, Christian Ledig	Rethinking model prototyping through the MedMNIST+ dataset collection	null	Scientific Reports 15, 7669 (2025)	10.1038/s41598-025-92156-9	null	eess.IV cs.CV cs.LG	http://creativecommons.org/licenses/by/4.0/	The integration of deep learning based systems in clinical practice is often impeded by challenges rooted in limited and heterogeneous medical datasets. In addition, the field has increasingly prioritized marginal performance gains on a few, narrowly scoped benchmarks over clinical applicability, slowing down meaningful algorithmic progress. This trend often results in excessive fine-tuning of existing methods on selected datasets rather than fostering clinically relevant innovations. In response, this work introduces a comprehensive benchmark for the MedMNIST+ dataset collection, designed to diversify the evaluation landscape across several imaging modalities, anatomical regions, classification tasks and sample sizes. We systematically reassess commonly used Convolutional Neural Networks (CNNs) and Vision Transformer (ViT) architectures across distinct medical datasets, training methodologies, and input resolutions to validate and refine existing assumptions about model effectiveness and development. Our findings suggest that computationally efficient training schemes and modern foundation models offer viable alternatives to costly end-to-end training. Additionally, we observe that higher image resolutions do not consistently improve performance beyond a certain threshold. This highlights the potential benefits of using lower resolutions, particularly in prototyping stages, to reduce computational demands without sacrificing accuracy. Notably, our analysis reaffirms the competitiveness of CNNs compared to ViTs, emphasizing the importance of comprehending the intrinsic capabilities of different architectures. Finally, by establishing a standardized evaluation framework, we aim to enhance transparency, reproducibility, and comparability within the MedMNIST+ dataset collection. Code is available at https://github.com/sdoerrich97/rethinking-model-prototyping-MedMNISTPlus .	[ { "version": "v1", "created": "Wed, 24 Apr 2024 10:19:25 GMT" }, { "version": "v2", "created": "Tue, 7 May 2024 20:49:46 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 12:01:18 GMT" } ]	2025-03-18T00:00:00	[ [ "Doerrich", "Sebastian", "" ], [ "Di Salvo", "Francesco", "" ], [ "Brockmann", "Julius", "" ], [ "Ledig", "Christian", "" ] ]	TITLE: Rethinking model prototyping through the MedMNIST+ dataset collection ABSTRACT: The integration of deep learning based systems in clinical practice is often impeded by challenges rooted in limited and heterogeneous medical datasets. In addition, the field has increasingly prioritized marginal performance gains on a few, narrowly scoped benchmarks over clinical applicability, slowing down meaningful algorithmic progress. This trend often results in excessive fine-tuning of existing methods on selected datasets rather than fostering clinically relevant innovations. In response, this work introduces a comprehensive benchmark for the MedMNIST+ dataset collection, designed to diversify the evaluation landscape across several imaging modalities, anatomical regions, classification tasks and sample sizes. We systematically reassess commonly used Convolutional Neural Networks (CNNs) and Vision Transformer (ViT) architectures across distinct medical datasets, training methodologies, and input resolutions to validate and refine existing assumptions about model effectiveness and development. Our findings suggest that computationally efficient training schemes and modern foundation models offer viable alternatives to costly end-to-end training. Additionally, we observe that higher image resolutions do not consistently improve performance beyond a certain threshold. This highlights the potential benefits of using lower resolutions, particularly in prototyping stages, to reduce computational demands without sacrificing accuracy. Notably, our analysis reaffirms the competitiveness of CNNs compared to ViTs, emphasizing the importance of comprehending the intrinsic capabilities of different architectures. Finally, by establishing a standardized evaluation framework, we aim to enhance transparency, reproducibility, and comparability within the MedMNIST+ dataset collection. Code is available at https://github.com/sdoerrich97/rethinking-model-prototyping-MedMNISTPlus .
2404.16367	Kabir Ahuja	Kabir Ahuja, Vidhisha Balachandran, Madhur Panwar, Tianxing He, Noah A. Smith, Navin Goyal, Yulia Tsvetkov	Learning Syntax Without Planting Trees: Understanding Hierarchical Generalization in Transformers	Accepted in TACL Code now available: https://github.com/kabirahuja2431/transformers-hg	null	null	null	cs.CL cs.LG	http://creativecommons.org/licenses/by/4.0/	Transformers trained on natural language data have been shown to learn its hierarchical structure and generalize to sentences with unseen syntactic structures without explicitly encoding any structural bias. In this work, we investigate sources of inductive bias in transformer models and their training that could cause such generalization behavior to emerge. We extensively experiment with transformer models trained on multiple synthetic datasets and with different training objectives and show that while other objectives e.g. sequence-to-sequence modeling, prefix language modeling, often failed to lead to hierarchical generalization, models trained with the language modeling objective consistently learned to generalize hierarchically. We then conduct pruning experiments to study how transformers trained with the language modeling objective encode hierarchical structure. When pruned, we find joint existence of subnetworks within the model with different generalization behaviors (subnetworks corresponding to hierarchical structure and linear order). Finally, we take a Bayesian perspective to further uncover transformers' preference for hierarchical generalization: We establish a correlation between whether transformers generalize hierarchically on a dataset and whether the simplest explanation of that dataset is provided by a hierarchical grammar compared to regular grammars exhibiting linear generalization.	[ { "version": "v1", "created": "Thu, 25 Apr 2024 07:10:29 GMT" }, { "version": "v2", "created": "Fri, 31 May 2024 23:47:15 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 05:23:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Ahuja", "Kabir", "" ], [ "Balachandran", "Vidhisha", "" ], [ "Panwar", "Madhur", "" ], [ "He", "Tianxing", "" ], [ "Smith", "Noah A.", "" ], [ "Goyal", "Navin", "" ], [ "Tsvetkov", "Yulia", "" ] ]	TITLE: Learning Syntax Without Planting Trees: Understanding Hierarchical Generalization in Transformers ABSTRACT: Transformers trained on natural language data have been shown to learn its hierarchical structure and generalize to sentences with unseen syntactic structures without explicitly encoding any structural bias. In this work, we investigate sources of inductive bias in transformer models and their training that could cause such generalization behavior to emerge. We extensively experiment with transformer models trained on multiple synthetic datasets and with different training objectives and show that while other objectives e.g. sequence-to-sequence modeling, prefix language modeling, often failed to lead to hierarchical generalization, models trained with the language modeling objective consistently learned to generalize hierarchically. We then conduct pruning experiments to study how transformers trained with the language modeling objective encode hierarchical structure. When pruned, we find joint existence of subnetworks within the model with different generalization behaviors (subnetworks corresponding to hierarchical structure and linear order). Finally, we take a Bayesian perspective to further uncover transformers' preference for hierarchical generalization: We establish a correlation between whether transformers generalize hierarchically on a dataset and whether the simplest explanation of that dataset is provided by a hierarchical grammar compared to regular grammars exhibiting linear generalization.
2404.16820	Chuhan Zhang	Olivia Wiles, Chuhan Zhang, Isabela Albuquerque, Ivana Kaji\'c, Su Wang, Emanuele Bugliarello, Yasumasa Onoe, Pinelopi Papalampidi, Ira Ktena, Chris Knutsen, Cyrus Rashtchian, Anant Nawalgaria, Jordi Pont-Tuset, Aida Nematzadeh	Revisiting Text-to-Image Evaluation with Gecko: On Metrics, Prompts, and Human Ratings	Accepted to ICLR 2025 (Spotlight)	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	While text-to-image (T2I) generative models have become ubiquitous, they do not necessarily generate images that align with a given prompt. While previous work has evaluated T2I alignment by proposing metrics, benchmarks, and templates for collecting human judgements, the quality of these components is not systematically measured. Human-rated prompt sets are generally small and the reliability of the ratings -- and thereby the prompt set used to compare models -- is not evaluated. We address this gap by performing an extensive study evaluating auto-eval metrics and human templates. We provide three main contributions: (1) We introduce a comprehensive skills-based benchmark that can discriminate models across different human templates. This skills-based benchmark categorises prompts into sub-skills, allowing a practitioner to pinpoint not only which skills are challenging, but at what level of complexity a skill becomes challenging. (2) We gather human ratings across four templates and four T2I models for a total of >100K annotations. This allows us to understand where differences arise due to inherent ambiguity in the prompt and where they arise due to differences in metric and model quality. (3) Finally, we introduce a new QA-based auto-eval metric that is better correlated with human ratings than existing metrics for our new dataset, across different human templates, and on TIFA160.	[ { "version": "v1", "created": "Thu, 25 Apr 2024 17:58:43 GMT" }, { "version": "v2", "created": "Tue, 18 Feb 2025 21:18:48 GMT" }, { "version": "v3", "created": "Sat, 1 Mar 2025 22:41:18 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 15:53:14 GMT" } ]	2025-03-18T00:00:00	[ [ "Wiles", "Olivia", "" ], [ "Zhang", "Chuhan", "" ], [ "Albuquerque", "Isabela", "" ], [ "Kajić", "Ivana", "" ], [ "Wang", "Su", "" ], [ "Bugliarello", "Emanuele", "" ], [ "Onoe", "Yasumasa", "" ], [ "Papalampidi", "Pinelopi", "" ], [ "Ktena", "Ira", "" ], [ "Knutsen", "Chris", "" ], [ "Rashtchian", "Cyrus", "" ], [ "Nawalgaria", "Anant", "" ], [ "Pont-Tuset", "Jordi", "" ], [ "Nematzadeh", "Aida", "" ] ]	TITLE: Revisiting Text-to-Image Evaluation with Gecko: On Metrics, Prompts, and Human Ratings ABSTRACT: While text-to-image (T2I) generative models have become ubiquitous, they do not necessarily generate images that align with a given prompt. While previous work has evaluated T2I alignment by proposing metrics, benchmarks, and templates for collecting human judgements, the quality of these components is not systematically measured. Human-rated prompt sets are generally small and the reliability of the ratings -- and thereby the prompt set used to compare models -- is not evaluated. We address this gap by performing an extensive study evaluating auto-eval metrics and human templates. We provide three main contributions: (1) We introduce a comprehensive skills-based benchmark that can discriminate models across different human templates. This skills-based benchmark categorises prompts into sub-skills, allowing a practitioner to pinpoint not only which skills are challenging, but at what level of complexity a skill becomes challenging. (2) We gather human ratings across four templates and four T2I models for a total of >100K annotations. This allows us to understand where differences arise due to inherent ambiguity in the prompt and where they arise due to differences in metric and model quality. (3) Finally, we introduce a new QA-based auto-eval metric that is better correlated with human ratings than existing metrics for our new dataset, across different human templates, and on TIFA160.
2404.17092	Weiran Chen	Weiran Chen, Qi Xu	Robust and Efficient Adversarial Defense in SNNs via Image Purification and Joint Detection	null	null	10.1109/ICASSP49660.2025.10888581	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Spiking Neural Networks (SNNs) aim to bridge the gap between neuroscience and machine learning by emulating the structure of the human nervous system. However, like convolutional neural networks, SNNs are vulnerable to adversarial attacks. To tackle the challenge, we propose a biologically inspired methodology to enhance the robustness of SNNs, drawing insights from the visual masking effect and filtering theory. First, an end-to-end SNN-based image purification model is proposed to defend against adversarial attacks, including a noise extraction network and a non-blind denoising network. The former network extracts noise features from noisy images, while the latter component employs a residual U-Net structure to reconstruct high-quality noisy images and generate clean images. Simultaneously, a multi-level firing SNN based on Squeeze-and-Excitation Network is introduced to improve the robustness of the classifier. Crucially, the proposed image purification network serves as a pre-processing module, avoiding modifications to classifiers. Unlike adversarial training, our method is highly flexible and can be seamlessly integrated with other defense strategies. Experimental results on various datasets demonstrate that the proposed methodology outperforms state-of-the-art baselines in terms of defense effectiveness, training time, and resource consumption.	[ { "version": "v1", "created": "Fri, 26 Apr 2024 00:57:06 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 05:06:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Chen", "Weiran", "" ], [ "Xu", "Qi", "" ] ]	TITLE: Robust and Efficient Adversarial Defense in SNNs via Image Purification and Joint Detection ABSTRACT: Spiking Neural Networks (SNNs) aim to bridge the gap between neuroscience and machine learning by emulating the structure of the human nervous system. However, like convolutional neural networks, SNNs are vulnerable to adversarial attacks. To tackle the challenge, we propose a biologically inspired methodology to enhance the robustness of SNNs, drawing insights from the visual masking effect and filtering theory. First, an end-to-end SNN-based image purification model is proposed to defend against adversarial attacks, including a noise extraction network and a non-blind denoising network. The former network extracts noise features from noisy images, while the latter component employs a residual U-Net structure to reconstruct high-quality noisy images and generate clean images. Simultaneously, a multi-level firing SNN based on Squeeze-and-Excitation Network is introduced to improve the robustness of the classifier. Crucially, the proposed image purification network serves as a pre-processing module, avoiding modifications to classifiers. Unlike adversarial training, our method is highly flexible and can be seamlessly integrated with other defense strategies. Experimental results on various datasets demonstrate that the proposed methodology outperforms state-of-the-art baselines in terms of defense effectiveness, training time, and resource consumption.
2405.00604	Theodor Westny Mr	Theodor Westny and Bj\"orn Olofsson and Erik Frisk	Toward Unified Practices in Trajectory Prediction Research on Drone Datasets	https://github.com/westny/dronalize	null	null	null	cs.RO cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The availability of high-quality datasets is crucial for the development of behavior prediction algorithms in autonomous vehicles. This paper highlights the need to standardize the use of certain datasets for motion forecasting research to simplify comparative analysis and proposes a set of tools and practices to achieve this. Drawing on extensive experience and a comprehensive review of current literature, we summarize our proposals for preprocessing, visualization, and evaluation in the form of an open-sourced toolbox designed for researchers working on trajectory prediction problems. The clear specification of necessary preprocessing steps and evaluation metrics is intended to alleviate development efforts and facilitate the comparison of results across different studies. The toolbox is available at: https://github.com/westny/dronalize.	[ { "version": "v1", "created": "Wed, 1 May 2024 16:17:39 GMT" }, { "version": "v2", "created": "Tue, 24 Sep 2024 09:18:59 GMT" }, { "version": "v3", "created": "Fri, 14 Mar 2025 22:13:49 GMT" } ]	2025-03-18T00:00:00	[ [ "Westny", "Theodor", "" ], [ "Olofsson", "Björn", "" ], [ "Frisk", "Erik", "" ] ]	TITLE: Toward Unified Practices in Trajectory Prediction Research on Drone Datasets ABSTRACT: The availability of high-quality datasets is crucial for the development of behavior prediction algorithms in autonomous vehicles. This paper highlights the need to standardize the use of certain datasets for motion forecasting research to simplify comparative analysis and proposes a set of tools and practices to achieve this. Drawing on extensive experience and a comprehensive review of current literature, we summarize our proposals for preprocessing, visualization, and evaluation in the form of an open-sourced toolbox designed for researchers working on trajectory prediction problems. The clear specification of necessary preprocessing steps and evaluation metrics is intended to alleviate development efforts and facilitate the comparison of results across different studies. The toolbox is available at: https://github.com/westny/dronalize.
2405.01217	Chenying Liu	Chenying Liu, Conrad Albrecht, Yi Wang, Xiao Xiang Zhu	CromSS: Cross-modal pre-training with noisy labels for remote sensing image segmentation	The 1st short version was accepted as an oral presentation by ICLR 2024 ML4RS workshop. The 2nd extended version was accepted by IEEE TGRS	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We explore the potential of large-scale noisily labeled data to enhance feature learning by pretraining semantic segmentation models within a multi-modal framework for geospatial applications. We propose a novel Cross-modal Sample Selection (CromSS) method, a weakly supervised pretraining strategy designed to improve feature representations through cross-modal consistency and noise mitigation techniques. Unlike conventional pretraining approaches, CromSS exploits massive amounts of noisy and easy-to-come-by labels for improved feature learning beneficial to semantic segmentation tasks. We investigate middle and late fusion strategies to optimize the multi-modal pretraining architecture design. We also introduce a cross-modal sample selection module to mitigate the adverse effects of label noise, which employs a cross-modal entangling strategy to refine the estimated confidence masks within each modality to guide the sampling process. Additionally, we introduce a spatial-temporal label smoothing technique to counteract overconfidence for enhanced robustness against noisy labels. To validate our approach, we assembled the multi-modal dataset, NoLDO-S12, which consists of a large-scale noisy label subset from Google's Dynamic World (DW) dataset for pretraining and two downstream subsets with high-quality labels from Google DW and OpenStreetMap (OSM) for transfer learning. Experimental results on two downstream tasks and the publicly available DFC2020 dataset demonstrate that when effectively utilized, the low-cost noisy labels can significantly enhance feature learning for segmentation tasks. All data, code, and pretrained weights will be made publicly available.	[ { "version": "v1", "created": "Thu, 2 May 2024 11:58:06 GMT" }, { "version": "v2", "created": "Mon, 3 Mar 2025 07:38:09 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 07:26:04 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Chenying", "" ], [ "Albrecht", "Conrad", "" ], [ "Wang", "Yi", "" ], [ "Zhu", "Xiao Xiang", "" ] ]	TITLE: CromSS: Cross-modal pre-training with noisy labels for remote sensing image segmentation ABSTRACT: We explore the potential of large-scale noisily labeled data to enhance feature learning by pretraining semantic segmentation models within a multi-modal framework for geospatial applications. We propose a novel Cross-modal Sample Selection (CromSS) method, a weakly supervised pretraining strategy designed to improve feature representations through cross-modal consistency and noise mitigation techniques. Unlike conventional pretraining approaches, CromSS exploits massive amounts of noisy and easy-to-come-by labels for improved feature learning beneficial to semantic segmentation tasks. We investigate middle and late fusion strategies to optimize the multi-modal pretraining architecture design. We also introduce a cross-modal sample selection module to mitigate the adverse effects of label noise, which employs a cross-modal entangling strategy to refine the estimated confidence masks within each modality to guide the sampling process. Additionally, we introduce a spatial-temporal label smoothing technique to counteract overconfidence for enhanced robustness against noisy labels. To validate our approach, we assembled the multi-modal dataset, NoLDO-S12, which consists of a large-scale noisy label subset from Google's Dynamic World (DW) dataset for pretraining and two downstream subsets with high-quality labels from Google DW and OpenStreetMap (OSM) for transfer learning. Experimental results on two downstream tasks and the publicly available DFC2020 dataset demonstrate that when effectively utilized, the low-cost noisy labels can significantly enhance feature learning for segmentation tasks. All data, code, and pretrained weights will be made publicly available.
2405.10948	Guankun Wang	Guankun Wang, Long Bai, Wan Jun Nah, Jie Wang, Zhaoxi Zhang, Zhen Chen, Jinlin Wu, Mobarakol Islam, Hongbin Liu, and Hongliang Ren	Surgical-LVLM: Learning to Adapt Large Vision-Language Model for Grounded Visual Question Answering in Robotic Surgery	The manuscript is accepted by ICLR 2025 FM-Wild Workshop	null	null	null	cs.CV cs.AI cs.RO eess.IV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Recent advancements in Surgical Visual Question Answering (Surgical-VQA) and related region grounding have shown great promise for robotic and medical applications, addressing the critical need for automated methods in personalized surgical mentorship. However, existing models primarily provide simple structured answers and struggle with complex scenarios due to their limited capability in recognizing long-range dependencies and aligning multimodal information. In this paper, we introduce Surgical-LVLM, a novel personalized large vision-language model tailored for complex surgical scenarios. Leveraging the pre-trained large vision-language model and specialized Visual Perception LoRA (VP-LoRA) blocks, our model excels in understanding complex visual-language tasks within surgical contexts. In addressing the visual grounding task, we propose the Token-Interaction (TIT) module, which strengthens the interaction between the grounding module and the language responses of the Large Visual Language Model (LVLM) after projecting them into the latent space. We demonstrate the effectiveness of Surgical-LVLM on several benchmarks, including EndoVis-17-VQLA, EndoVis-18-VQLA, and a newly introduced EndoVis Conversations dataset, which sets new performance standards. Our work contributes to advancing the field of automated surgical mentorship by providing a context-aware solution.	[ { "version": "v1", "created": "Fri, 22 Mar 2024 08:38:27 GMT" }, { "version": "v2", "created": "Fri, 7 Mar 2025 01:02:22 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 02:23:30 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Guankun", "" ], [ "Bai", "Long", "" ], [ "Nah", "Wan Jun", "" ], [ "Wang", "Jie", "" ], [ "Zhang", "Zhaoxi", "" ], [ "Chen", "Zhen", "" ], [ "Wu", "Jinlin", "" ], [ "Islam", "Mobarakol", "" ], [ "Liu", "Hongbin", "" ], [ "Ren", "Hongliang", "" ] ]	TITLE: Surgical-LVLM: Learning to Adapt Large Vision-Language Model for Grounded Visual Question Answering in Robotic Surgery ABSTRACT: Recent advancements in Surgical Visual Question Answering (Surgical-VQA) and related region grounding have shown great promise for robotic and medical applications, addressing the critical need for automated methods in personalized surgical mentorship. However, existing models primarily provide simple structured answers and struggle with complex scenarios due to their limited capability in recognizing long-range dependencies and aligning multimodal information. In this paper, we introduce Surgical-LVLM, a novel personalized large vision-language model tailored for complex surgical scenarios. Leveraging the pre-trained large vision-language model and specialized Visual Perception LoRA (VP-LoRA) blocks, our model excels in understanding complex visual-language tasks within surgical contexts. In addressing the visual grounding task, we propose the Token-Interaction (TIT) module, which strengthens the interaction between the grounding module and the language responses of the Large Visual Language Model (LVLM) after projecting them into the latent space. We demonstrate the effectiveness of Surgical-LVLM on several benchmarks, including EndoVis-17-VQLA, EndoVis-18-VQLA, and a newly introduced EndoVis Conversations dataset, which sets new performance standards. Our work contributes to advancing the field of automated surgical mentorship by providing a context-aware solution.
2405.16868	Tianhang Wang	Tianhang Wang, Fan Lu, Zehan Zheng, Zhijun Li, Guang Chen, Changjun Jiang	RCDN: Towards Robust Camera-Insensitivity Collaborative Perception via Dynamic Feature-based 3D Neural Modeling	null	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Collaborative perception is dedicated to tackling the constraints of single-agent perception, such as occlusions, based on the multiple agents' multi-view sensor inputs. However, most existing works assume an ideal condition that all agents' multi-view cameras are continuously available. In reality, cameras may be highly noisy, obscured or even failed during the collaboration. In this work, we introduce a new robust camera-insensitivity problem: how to overcome the issues caused by the failed camera perspectives, while stabilizing high collaborative performance with low calibration cost? To address above problems, we propose RCDN, a Robust Camera-insensitivity collaborative perception with a novel Dynamic feature-based 3D Neural modeling mechanism. The key intuition of RCDN is to construct collaborative neural rendering field representations to recover failed perceptual messages sent by multiple agents. To better model collaborative neural rendering field, RCDN first establishes a geometry BEV feature based time-invariant static field with other agents via fast hash grid modeling. Based on the static background field, the proposed time-varying dynamic field can model corresponding motion vectors for foregrounds with appropriate positions. To validate RCDN, we create OPV2V-N, a new large-scale dataset with manual labelling under different camera failed scenarios. Extensive experiments conducted on OPV2V-N show that RCDN can be ported to other baselines and improve their robustness in extreme camera-insensitivity settings.	[ { "version": "v1", "created": "Mon, 27 May 2024 06:35:55 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 06:27:08 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Tianhang", "" ], [ "Lu", "Fan", "" ], [ "Zheng", "Zehan", "" ], [ "Li", "Zhijun", "" ], [ "Chen", "Guang", "" ], [ "Jiang", "Changjun", "" ] ]	TITLE: RCDN: Towards Robust Camera-Insensitivity Collaborative Perception via Dynamic Feature-based 3D Neural Modeling ABSTRACT: Collaborative perception is dedicated to tackling the constraints of single-agent perception, such as occlusions, based on the multiple agents' multi-view sensor inputs. However, most existing works assume an ideal condition that all agents' multi-view cameras are continuously available. In reality, cameras may be highly noisy, obscured or even failed during the collaboration. In this work, we introduce a new robust camera-insensitivity problem: how to overcome the issues caused by the failed camera perspectives, while stabilizing high collaborative performance with low calibration cost? To address above problems, we propose RCDN, a Robust Camera-insensitivity collaborative perception with a novel Dynamic feature-based 3D Neural modeling mechanism. The key intuition of RCDN is to construct collaborative neural rendering field representations to recover failed perceptual messages sent by multiple agents. To better model collaborative neural rendering field, RCDN first establishes a geometry BEV feature based time-invariant static field with other agents via fast hash grid modeling. Based on the static background field, the proposed time-varying dynamic field can model corresponding motion vectors for foregrounds with appropriate positions. To validate RCDN, we create OPV2V-N, a new large-scale dataset with manual labelling under different camera failed scenarios. Extensive experiments conducted on OPV2V-N show that RCDN can be ported to other baselines and improve their robustness in extreme camera-insensitivity settings.
2405.17035	Harshit Varma	Harshit Varma, Dheeraj Nagaraj, Karthikeyan Shanmugam	Glauber Generative Model: Discrete Diffusion Models via Binary Classification	ICLR 2025	null	null	null	cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We introduce the Glauber Generative Model (GGM), a new class of discrete diffusion models, to obtain new samples from a distribution given samples from a discrete space. GGM deploys a discrete Markov chain called the heat bath dynamics (or the Glauber dynamics) to denoise a sequence of noisy tokens to a sample from a joint distribution of discrete tokens. Our novel conceptual framework provides an exact reduction of the task of learning the denoising Markov chain to solving a class of binary classification tasks. More specifically, the model learns to classify a given token in a noisy sequence as signal or noise. In contrast, prior works on discrete diffusion models either solve regression problems to learn importance ratios, or minimize loss functions given by variational approximations. We apply GGM to language modeling and image generation, where images are discretized using image tokenizers like VQGANs. We show that it outperforms existing discrete diffusion models in language generation, and demonstrates strong performance for image generation without using dataset-specific image tokenizers. We also show that our model is capable of performing well in zero-shot control settings like text and image infilling.	[ { "version": "v1", "created": "Mon, 27 May 2024 10:42:13 GMT" }, { "version": "v2", "created": "Thu, 27 Jun 2024 05:09:57 GMT" }, { "version": "v3", "created": "Tue, 27 Aug 2024 13:05:33 GMT" }, { "version": "v4", "created": "Sun, 16 Mar 2025 09:13:20 GMT" } ]	2025-03-18T00:00:00	[ [ "Varma", "Harshit", "" ], [ "Nagaraj", "Dheeraj", "" ], [ "Shanmugam", "Karthikeyan", "" ] ]	TITLE: Glauber Generative Model: Discrete Diffusion Models via Binary Classification ABSTRACT: We introduce the Glauber Generative Model (GGM), a new class of discrete diffusion models, to obtain new samples from a distribution given samples from a discrete space. GGM deploys a discrete Markov chain called the heat bath dynamics (or the Glauber dynamics) to denoise a sequence of noisy tokens to a sample from a joint distribution of discrete tokens. Our novel conceptual framework provides an exact reduction of the task of learning the denoising Markov chain to solving a class of binary classification tasks. More specifically, the model learns to classify a given token in a noisy sequence as signal or noise. In contrast, prior works on discrete diffusion models either solve regression problems to learn importance ratios, or minimize loss functions given by variational approximations. We apply GGM to language modeling and image generation, where images are discretized using image tokenizers like VQGANs. We show that it outperforms existing discrete diffusion models in language generation, and demonstrates strong performance for image generation without using dataset-specific image tokenizers. We also show that our model is capable of performing well in zero-shot control settings like text and image infilling.
2405.18684	Mohammadjavad Matinkia	Mohammadjavad Matinkia, Nilanjan Ray	Learning Diffeomorphism for Image Registration with Time-Continuous Networks using Semigroup Regularization	27 pages, 11 figures	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Diffeomorphic image registration (DIR) is a fundamental task in 3D medical image analysis that seeks topology-preserving deformations between image pairs. To ensure diffeomorphism, a common approach is to model the deformation field as the flow map solution of a differential equation, which is solved using efficient schemes such as scaling and squaring along with multiple smoothness regularization terms. In this paper, we propose a novel learning-based approach for diffeomorphic 3D image registration that models diffeomorphisms in a continuous-time framework using only a single regularization term, without requiring additional integration. We exploit the semigroup property-a fundamental characteristic of flow maps-as the sole form of regularization, ensuring temporally continuous diffeomorphic flows between image pairs. Leveraging this property, we prove that our formulation directly learns the flow map solution of an ODE, ensuring continuous inverse and cycle consistencies without explicit enforcement, while eliminating additional integration schemes and regularization terms. To achieve time-continuous diffeomorphisms, we employ time-embedded UNets, an architecture commonly used in diffusion models. Our results demonstrate that modeling diffeomorphism continuously in time improves registration performance. Experimental results on four public datasets demonstrate the superiority of our model over state-of-the-art diffeomorphic methods. Additionally, comparison to several recent non-diffeomorphic deformable image registration methods shows that our method achieves competitive Dice scores while significantly improving topology preservation.	[ { "version": "v1", "created": "Wed, 29 May 2024 01:25:43 GMT" }, { "version": "v2", "created": "Thu, 14 Nov 2024 04:13:08 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 21:22:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Matinkia", "Mohammadjavad", "" ], [ "Ray", "Nilanjan", "" ] ]	TITLE: Learning Diffeomorphism for Image Registration with Time-Continuous Networks using Semigroup Regularization ABSTRACT: Diffeomorphic image registration (DIR) is a fundamental task in 3D medical image analysis that seeks topology-preserving deformations between image pairs. To ensure diffeomorphism, a common approach is to model the deformation field as the flow map solution of a differential equation, which is solved using efficient schemes such as scaling and squaring along with multiple smoothness regularization terms. In this paper, we propose a novel learning-based approach for diffeomorphic 3D image registration that models diffeomorphisms in a continuous-time framework using only a single regularization term, without requiring additional integration. We exploit the semigroup property-a fundamental characteristic of flow maps-as the sole form of regularization, ensuring temporally continuous diffeomorphic flows between image pairs. Leveraging this property, we prove that our formulation directly learns the flow map solution of an ODE, ensuring continuous inverse and cycle consistencies without explicit enforcement, while eliminating additional integration schemes and regularization terms. To achieve time-continuous diffeomorphisms, we employ time-embedded UNets, an architecture commonly used in diffusion models. Our results demonstrate that modeling diffeomorphism continuously in time improves registration performance. Experimental results on four public datasets demonstrate the superiority of our model over state-of-the-art diffeomorphic methods. Additionally, comparison to several recent non-diffeomorphic deformable image registration methods shows that our method achieves competitive Dice scores while significantly improving topology preservation.
2406.00430	Jianxiang Feng	Zhi Zheng, Qian Feng, Hang Li, Alois Knoll, Jianxiang Feng	Evaluating Uncertainty-based Failure Detection for Closed-Loop LLM Planners	Accepted at ICRA 2024 Workshop on Back to the Future: Robot Learning Going Probabilistic. Website: https://sites.google.com/view/konwloop/home	null	null	null	cs.RO cs.AI	http://creativecommons.org/licenses/by/4.0/	Recently, Large Language Models (LLMs) have witnessed remarkable performance as zero-shot task planners for robotic manipulation tasks. However, the open-loop nature of previous works makes LLM-based planning error-prone and fragile. On the other hand, failure detection approaches for closed-loop planning are often limited by task-specific heuristics or following an unrealistic assumption that the prediction is trustworthy all the time. As a general-purpose reasoning machine, LLMs or Multimodal Large Language Models (MLLMs) are promising for detecting failures. However, However, the appropriateness of the aforementioned assumption diminishes due to the notorious hullucination problem. In this work, we attempt to mitigate these issues by introducing a framework for closed-loop LLM-based planning called KnowLoop, backed by an uncertainty-based MLLMs failure detector, which is agnostic to any used MLLMs or LLMs. Specifically, we evaluate three different ways for quantifying the uncertainty of MLLMs, namely token probability, entropy, and self-explained confidence as primary metrics based on three carefully designed representative prompting strategies. With a self-collected dataset including various manipulation tasks and an LLM-based robot system, our experiments demonstrate that token probability and entropy are more reflective compared to self-explained confidence. By setting an appropriate threshold to filter out uncertain predictions and seek human help actively, the accuracy of failure detection can be significantly enhanced. This improvement boosts the effectiveness of closed-loop planning and the overall success rate of tasks.	[ { "version": "v1", "created": "Sat, 1 Jun 2024 12:52:06 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 17:21:09 GMT" } ]	2025-03-18T00:00:00	[ [ "Zheng", "Zhi", "" ], [ "Feng", "Qian", "" ], [ "Li", "Hang", "" ], [ "Knoll", "Alois", "" ], [ "Feng", "Jianxiang", "" ] ]	TITLE: Evaluating Uncertainty-based Failure Detection for Closed-Loop LLM Planners ABSTRACT: Recently, Large Language Models (LLMs) have witnessed remarkable performance as zero-shot task planners for robotic manipulation tasks. However, the open-loop nature of previous works makes LLM-based planning error-prone and fragile. On the other hand, failure detection approaches for closed-loop planning are often limited by task-specific heuristics or following an unrealistic assumption that the prediction is trustworthy all the time. As a general-purpose reasoning machine, LLMs or Multimodal Large Language Models (MLLMs) are promising for detecting failures. However, However, the appropriateness of the aforementioned assumption diminishes due to the notorious hullucination problem. In this work, we attempt to mitigate these issues by introducing a framework for closed-loop LLM-based planning called KnowLoop, backed by an uncertainty-based MLLMs failure detector, which is agnostic to any used MLLMs or LLMs. Specifically, we evaluate three different ways for quantifying the uncertainty of MLLMs, namely token probability, entropy, and self-explained confidence as primary metrics based on three carefully designed representative prompting strategies. With a self-collected dataset including various manipulation tasks and an LLM-based robot system, our experiments demonstrate that token probability and entropy are more reflective compared to self-explained confidence. By setting an appropriate threshold to filter out uncertain predictions and seek human help actively, the accuracy of failure detection can be significantly enhanced. This improvement boosts the effectiveness of closed-loop planning and the overall success rate of tasks.
2406.02923	Malyaban Bal	Malyaban Bal and Abhronil Sengupta	P-SpikeSSM: Harnessing Probabilistic Spiking State Space Models for Long-Range Dependency Tasks	Accepted at ICLR 2025	null	null	null	cs.NE	http://creativecommons.org/licenses/by/4.0/	Spiking neural networks (SNNs) are posited as a computationally efficient and biologically plausible alternative to conventional neural architectures, with their core computational framework primarily using the leaky integrate-and-fire (LIF) neuron model. However, the limited hidden state representation of LIF neurons, characterized by a scalar membrane potential, and sequential spike generation process, poses challenges for effectively developing scalable spiking models to address long-range dependencies in sequence learning tasks. In this study, we develop a scalable probabilistic spiking learning framework for long-range dependency tasks leveraging the fundamentals of state space models. Unlike LIF neurons that rely on the deterministic Heaviside function for a sequential process of spike generation, we introduce a SpikeSampler layer that samples spikes stochastically based on an SSM-based neuronal model while allowing parallel computations. To address non-differentiability of the spiking operation and enable effective training, we also propose a surrogate function tailored for the stochastic nature of the SpikeSampler layer. To enhance inter-neuron communication, we introduce the SpikeMixer block, which integrates spikes from neuron populations in each layer. This is followed by a ClampFuse layer, incorporating a residual connection to capture complex dependencies, enabling scalability of the model. Our models attain state-of-the-art performance among SNN models across diverse long-range dependency tasks, encompassing the Long Range Arena benchmark, permuted sequential MNIST, and the Speech Command dataset and demonstrate sparse spiking pattern highlighting its computational efficiency.	[ { "version": "v1", "created": "Wed, 5 Jun 2024 04:23:11 GMT" }, { "version": "v2", "created": "Thu, 3 Oct 2024 18:55:14 GMT" }, { "version": "v3", "created": "Thu, 20 Feb 2025 18:44:10 GMT" }, { "version": "v4", "created": "Mon, 3 Mar 2025 06:02:04 GMT" }, { "version": "v5", "created": "Mon, 17 Mar 2025 01:02:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Bal", "Malyaban", "" ], [ "Sengupta", "Abhronil", "" ] ]	TITLE: P-SpikeSSM: Harnessing Probabilistic Spiking State Space Models for Long-Range Dependency Tasks ABSTRACT: Spiking neural networks (SNNs) are posited as a computationally efficient and biologically plausible alternative to conventional neural architectures, with their core computational framework primarily using the leaky integrate-and-fire (LIF) neuron model. However, the limited hidden state representation of LIF neurons, characterized by a scalar membrane potential, and sequential spike generation process, poses challenges for effectively developing scalable spiking models to address long-range dependencies in sequence learning tasks. In this study, we develop a scalable probabilistic spiking learning framework for long-range dependency tasks leveraging the fundamentals of state space models. Unlike LIF neurons that rely on the deterministic Heaviside function for a sequential process of spike generation, we introduce a SpikeSampler layer that samples spikes stochastically based on an SSM-based neuronal model while allowing parallel computations. To address non-differentiability of the spiking operation and enable effective training, we also propose a surrogate function tailored for the stochastic nature of the SpikeSampler layer. To enhance inter-neuron communication, we introduce the SpikeMixer block, which integrates spikes from neuron populations in each layer. This is followed by a ClampFuse layer, incorporating a residual connection to capture complex dependencies, enabling scalability of the model. Our models attain state-of-the-art performance among SNN models across diverse long-range dependency tasks, encompassing the Long Range Arena benchmark, permuted sequential MNIST, and the Speech Command dataset and demonstrate sparse spiking pattern highlighting its computational efficiency.
2406.04419	Md Atik Ahamed	Md Atik Ahamed, Qiang Cheng	TSCMamba: Mamba Meets Multi-View Learning for Time Series Classification	null	null	null	null	cs.LG	http://creativecommons.org/licenses/by-nc-nd/4.0/	Multivariate time series classification (TSC) is critical for various applications in fields such as healthcare and finance. While various approaches for TSC have been explored, important properties of time series, such as shift equivariance and inversion invariance, are largely underexplored by existing works. To fill this gap, we propose a novel multi-view approach to capture patterns with properties like shift equivariance. Our method integrates diverse features, including spectral, temporal, local, and global features, to obtain rich, complementary contexts for TSC. We use continuous wavelet transform to capture time-frequency features that remain consistent even when the input is shifted in time. These features are fused with temporal convolutional or multilayer perceptron features to provide complex local and global contextual information. We utilize the Mamba state space model for efficient and scalable sequence modeling and capturing long-range dependencies in time series. Moreover, we introduce a new scanning scheme for Mamba, called tango scanning, to effectively model sequence relationships and leverage inversion invariance, thereby enhancing our model's generalization and robustness. Experiments on two sets of benchmark datasets (10+20 datasets) demonstrate our approach's effectiveness, achieving average accuracy improvements of 4.01-6.45\% and 7.93\% respectively, over leading TSC models such as TimesNet and TSLANet.	[ { "version": "v1", "created": "Thu, 6 Jun 2024 18:05:10 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 17:40:41 GMT" } ]	2025-03-18T00:00:00	[ [ "Ahamed", "Md Atik", "" ], [ "Cheng", "Qiang", "" ] ]	TITLE: TSCMamba: Mamba Meets Multi-View Learning for Time Series Classification ABSTRACT: Multivariate time series classification (TSC) is critical for various applications in fields such as healthcare and finance. While various approaches for TSC have been explored, important properties of time series, such as shift equivariance and inversion invariance, are largely underexplored by existing works. To fill this gap, we propose a novel multi-view approach to capture patterns with properties like shift equivariance. Our method integrates diverse features, including spectral, temporal, local, and global features, to obtain rich, complementary contexts for TSC. We use continuous wavelet transform to capture time-frequency features that remain consistent even when the input is shifted in time. These features are fused with temporal convolutional or multilayer perceptron features to provide complex local and global contextual information. We utilize the Mamba state space model for efficient and scalable sequence modeling and capturing long-range dependencies in time series. Moreover, we introduce a new scanning scheme for Mamba, called tango scanning, to effectively model sequence relationships and leverage inversion invariance, thereby enhancing our model's generalization and robustness. Experiments on two sets of benchmark datasets (10+20 datasets) demonstrate our approach's effectiveness, achieving average accuracy improvements of 4.01-6.45\% and 7.93\% respectively, over leading TSC models such as TimesNet and TSLANet.
2406.04927	Georgios Efstathiadis	Georgios Efstathiadis, Vijay Yadav, Anzar Abbas	LLM-based speaker diarization correction: A generalizable approach	null	Speech Communication, Volume 170, 2025, Page 103224	10.1016/j.specom.2025.103224	null	eess.AS cs.CL	http://creativecommons.org/licenses/by-nc-nd/4.0/	Speaker diarization is necessary for interpreting conversations transcribed using automated speech recognition (ASR) tools. Despite significant developments in diarization methods, diarization accuracy remains an issue. Here, we investigate the use of large language models (LLMs) for diarization correction as a post-processing step. LLMs were fine-tuned using the Fisher corpus, a large dataset of transcribed conversations. The ability of the models to improve diarization accuracy in a holdout dataset from the Fisher corpus as well as an independent dataset was measured. We report that fine-tuned LLMs can markedly improve diarization accuracy. However, model performance is constrained to transcripts produced using the same ASR tool as the transcripts used for fine-tuning, limiting generalizability. To address this constraint, an ensemble model was developed by combining weights from three separate models, each fine-tuned using transcripts from a different ASR tool. The ensemble model demonstrated better overall performance than each of the ASR-specific models, suggesting that a generalizable and ASR-agnostic approach may be achievable. We have made the weights of these models publicly available on HuggingFace at https://huggingface.co/bklynhlth.	[ { "version": "v1", "created": "Fri, 7 Jun 2024 13:33:22 GMT" }, { "version": "v2", "created": "Fri, 13 Sep 2024 20:42:20 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 13:34:07 GMT" } ]	2025-03-18T00:00:00	[ [ "Efstathiadis", "Georgios", "" ], [ "Yadav", "Vijay", "" ], [ "Abbas", "Anzar", "" ] ]	TITLE: LLM-based speaker diarization correction: A generalizable approach ABSTRACT: Speaker diarization is necessary for interpreting conversations transcribed using automated speech recognition (ASR) tools. Despite significant developments in diarization methods, diarization accuracy remains an issue. Here, we investigate the use of large language models (LLMs) for diarization correction as a post-processing step. LLMs were fine-tuned using the Fisher corpus, a large dataset of transcribed conversations. The ability of the models to improve diarization accuracy in a holdout dataset from the Fisher corpus as well as an independent dataset was measured. We report that fine-tuned LLMs can markedly improve diarization accuracy. However, model performance is constrained to transcripts produced using the same ASR tool as the transcripts used for fine-tuning, limiting generalizability. To address this constraint, an ensemble model was developed by combining weights from three separate models, each fine-tuned using transcripts from a different ASR tool. The ensemble model demonstrated better overall performance than each of the ASR-specific models, suggesting that a generalizable and ASR-agnostic approach may be achievable. We have made the weights of these models publicly available on HuggingFace at https://huggingface.co/bklynhlth.
2406.08920	Swapnil Bhosale	Swapnil Bhosale, Haosen Yang, Diptesh Kanojia, Jiankang Deng, Xiatian Zhu	AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis	Accepted to NeurIPS 2024	null	null	null	cs.SD cs.AI eess.AS	http://creativecommons.org/licenses/by-nc-sa/4.0/	Novel view acoustic synthesis (NVAS) aims to render binaural audio at any target viewpoint, given a mono audio emitted by a sound source at a 3D scene. Existing methods have proposed NeRF-based implicit models to exploit visual cues as a condition for synthesizing binaural audio. However, in addition to low efficiency originating from heavy NeRF rendering, these methods all have a limited ability of characterizing the entire scene environment such as room geometry, material properties, and the spatial relation between the listener and sound source. To address these issues, we propose a novel Audio-Visual Gaussian Splatting (AV-GS) model. To obtain a material-aware and geometry-aware condition for audio synthesis, we learn an explicit point-based scene representation with an audio-guidance parameter on locally initialized Gaussian points, taking into account the space relation from the listener and sound source. To make the visual scene model audio adaptive, we propose a point densification and pruning strategy to optimally distribute the Gaussian points, with the per-point contribution in sound propagation (e.g., more points needed for texture-less wall surfaces as they affect sound path diversion). Extensive experiments validate the superiority of our AV-GS over existing alternatives on the real-world RWAS and simulation-based SoundSpaces datasets.	[ { "version": "v1", "created": "Thu, 13 Jun 2024 08:34:12 GMT" }, { "version": "v2", "created": "Fri, 14 Jun 2024 06:38:50 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 19:43:03 GMT" } ]	2025-03-18T00:00:00	[ [ "Bhosale", "Swapnil", "" ], [ "Yang", "Haosen", "" ], [ "Kanojia", "Diptesh", "" ], [ "Deng", "Jiankang", "" ], [ "Zhu", "Xiatian", "" ] ]	TITLE: AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis ABSTRACT: Novel view acoustic synthesis (NVAS) aims to render binaural audio at any target viewpoint, given a mono audio emitted by a sound source at a 3D scene. Existing methods have proposed NeRF-based implicit models to exploit visual cues as a condition for synthesizing binaural audio. However, in addition to low efficiency originating from heavy NeRF rendering, these methods all have a limited ability of characterizing the entire scene environment such as room geometry, material properties, and the spatial relation between the listener and sound source. To address these issues, we propose a novel Audio-Visual Gaussian Splatting (AV-GS) model. To obtain a material-aware and geometry-aware condition for audio synthesis, we learn an explicit point-based scene representation with an audio-guidance parameter on locally initialized Gaussian points, taking into account the space relation from the listener and sound source. To make the visual scene model audio adaptive, we propose a point densification and pruning strategy to optimally distribute the Gaussian points, with the per-point contribution in sound propagation (e.g., more points needed for texture-less wall surfaces as they affect sound path diversion). Extensive experiments validate the superiority of our AV-GS over existing alternatives on the real-world RWAS and simulation-based SoundSpaces datasets.
2406.11601	Weronika Ormaniec	Weronika Ormaniec, Scott Sussex, Lars Lorch, Bernhard Sch\"olkopf, Andreas Krause	Standardizing Structural Causal Models	Added additional benchmarks, including PC algorithm, GES, GOLEM. Evaluated Var-sortability and R2-sortability of the heuristics for mitigating variance accumulation	null	null	null	cs.LG stat.ML	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Synthetic datasets generated by structural causal models (SCMs) are commonly used for benchmarking causal structure learning algorithms. However, the variances and pairwise correlations in SCM data tend to increase along the causal ordering. Several popular algorithms exploit these artifacts, possibly leading to conclusions that do not generalize to real-world settings. Existing metrics like $\operatorname{Var}$-sortability and $\operatorname{R^2}$-sortability quantify these patterns, but they do not provide tools to remedy them. To address this, we propose internally-standardized structural causal models (iSCMs), a modification of SCMs that introduces a standardization operation at each variable during the generative process. By construction, iSCMs are not $\operatorname{Var}$-sortable. We also find empirical evidence that they are mostly not $\operatorname{R^2}$-sortable for commonly-used graph families. Moreover, contrary to the post-hoc standardization of data generated by standard SCMs, we prove that linear iSCMs are less identifiable from prior knowledge on the weights and do not collapse to deterministic relationships in large systems, which may make iSCMs a useful model in causal inference beyond the benchmarking problem studied here. Our code is publicly available at: https://github.com/werkaaa/iscm.	[ { "version": "v1", "created": "Mon, 17 Jun 2024 14:52:21 GMT" }, { "version": "v2", "created": "Thu, 10 Oct 2024 21:14:49 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 14:26:33 GMT" } ]	2025-03-18T00:00:00	[ [ "Ormaniec", "Weronika", "" ], [ "Sussex", "Scott", "" ], [ "Lorch", "Lars", "" ], [ "Schölkopf", "Bernhard", "" ], [ "Krause", "Andreas", "" ] ]	TITLE: Standardizing Structural Causal Models ABSTRACT: Synthetic datasets generated by structural causal models (SCMs) are commonly used for benchmarking causal structure learning algorithms. However, the variances and pairwise correlations in SCM data tend to increase along the causal ordering. Several popular algorithms exploit these artifacts, possibly leading to conclusions that do not generalize to real-world settings. Existing metrics like $\operatorname{Var}$-sortability and $\operatorname{R^2}$-sortability quantify these patterns, but they do not provide tools to remedy them. To address this, we propose internally-standardized structural causal models (iSCMs), a modification of SCMs that introduces a standardization operation at each variable during the generative process. By construction, iSCMs are not $\operatorname{Var}$-sortable. We also find empirical evidence that they are mostly not $\operatorname{R^2}$-sortable for commonly-used graph families. Moreover, contrary to the post-hoc standardization of data generated by standard SCMs, we prove that linear iSCMs are less identifiable from prior knowledge on the weights and do not collapse to deterministic relationships in large systems, which may make iSCMs a useful model in causal inference beyond the benchmarking problem studied here. Our code is publicly available at: https://github.com/werkaaa/iscm.
2406.13378	Zidong Cao	Zidong Cao, Jinjing Zhu, Weiming Zhang, Hao Ai, Haotian Bai, Hengshuang Zhao, Lin Wang	PanDA: Towards Panoramic Depth Anything with Unlabeled Panoramas and Mobius Spatial Augmentation	16 pages, 18 figures, accepted by CVPR 2025	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Recently, Depth Anything Models (DAMs) - a type of depth foundation models - have demonstrated impressive zero-shot capabilities across diverse perspective images. Despite its success, it remains an open question regarding DAMs' performance on panorama images that enjoy a large field-of-view (180x360) but suffer from spherical distortions. To address this gap, we conduct an empirical analysis to evaluate the performance of DAMs on panoramic images and identify their limitations. For this, we undertake comprehensive experiments to assess the performance of DAMs from three key factors: panoramic representations, 360 camera positions for capturing scenarios, and spherical spatial transformations. This way, we reveal some key findings, e.g., DAMs are sensitive to spatial transformations. We then propose a semi-supervised learning (SSL) framework to learn a panoramic DAM, dubbed PanDA. Under the umbrella of SSL, PanDA first learns a teacher model by fine-tuning DAM through joint training on synthetic indoor and outdoor panoramic datasets. Then, a student model is trained using large-scale unlabeled data, leveraging pseudo-labels generated by the teacher model. To enhance PanDA's generalization capability, M"obius transformation-based spatial augmentation (MTSA) is proposed to impose consistency regularization between the predicted depth maps from the original and spatially transformed ones. This subtly improves the student model's robustness to various spatial transformations, even under severe distortions. Extensive experiments demonstrate that PanDA exhibits remarkable zero-shot capability across diverse scenes, and outperforms the data-specific panoramic depth estimation methods on two popular real-world benchmarks.	[ { "version": "v1", "created": "Wed, 19 Jun 2024 09:19:06 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 09:07:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Cao", "Zidong", "" ], [ "Zhu", "Jinjing", "" ], [ "Zhang", "Weiming", "" ], [ "Ai", "Hao", "" ], [ "Bai", "Haotian", "" ], [ "Zhao", "Hengshuang", "" ], [ "Wang", "Lin", "" ] ]	TITLE: PanDA: Towards Panoramic Depth Anything with Unlabeled Panoramas and Mobius Spatial Augmentation ABSTRACT: Recently, Depth Anything Models (DAMs) - a type of depth foundation models - have demonstrated impressive zero-shot capabilities across diverse perspective images. Despite its success, it remains an open question regarding DAMs' performance on panorama images that enjoy a large field-of-view (180x360) but suffer from spherical distortions. To address this gap, we conduct an empirical analysis to evaluate the performance of DAMs on panoramic images and identify their limitations. For this, we undertake comprehensive experiments to assess the performance of DAMs from three key factors: panoramic representations, 360 camera positions for capturing scenarios, and spherical spatial transformations. This way, we reveal some key findings, e.g., DAMs are sensitive to spatial transformations. We then propose a semi-supervised learning (SSL) framework to learn a panoramic DAM, dubbed PanDA. Under the umbrella of SSL, PanDA first learns a teacher model by fine-tuning DAM through joint training on synthetic indoor and outdoor panoramic datasets. Then, a student model is trained using large-scale unlabeled data, leveraging pseudo-labels generated by the teacher model. To enhance PanDA's generalization capability, M"obius transformation-based spatial augmentation (MTSA) is proposed to impose consistency regularization between the predicted depth maps from the original and spatially transformed ones. This subtly improves the student model's robustness to various spatial transformations, even under severe distortions. Extensive experiments demonstrate that PanDA exhibits remarkable zero-shot capability across diverse scenes, and outperforms the data-specific panoramic depth estimation methods on two popular real-world benchmarks.
2406.17503	Fu Feng	Fu Feng, Yucheng Xie, Jing Wang, Xin Geng	WAVE: Weight Templates for Adaptive Initialization of Variable-sized Models	null	null	null	null	cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The growing complexity of model parameters underscores the significance of pre-trained models. However, deployment constraints often necessitate models of varying sizes, exposing limitations in the conventional pre-training and fine-tuning paradigm, particularly when target model sizes are incompatible with pre-trained ones. To address this challenge, we propose WAVE, a novel approach that reformulates variable-sized model initialization from a multi-task perspective, where initializing each model size is treated as a distinct task. WAVE employs shared, size-agnostic weight templates alongside size-specific weight scalers to achieve consistent initialization across various model sizes. These weight templates, constructed within the Learngene framework, integrate knowledge from pre-trained models through a distillation process constrained by Kronecker-based rules. Target models are then initialized by concatenating and weighting these templates, with adaptive connection rules established by lightweight weight scalers, whose parameters are learned from minimal training data. Extensive experiments demonstrate the efficiency of WAVE, achieving state-of-the-art performance in initializing models of various depth and width. The knowledge encapsulated in weight templates is also task-agnostic, allowing for seamless transfer across diverse downstream datasets. Code will be made available at https://github.com/fu-feng/WAVE.	[ { "version": "v1", "created": "Tue, 25 Jun 2024 12:43:33 GMT" }, { "version": "v2", "created": "Mon, 15 Jul 2024 06:41:13 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 17:21:38 GMT" } ]	2025-03-18T00:00:00	[ [ "Feng", "Fu", "" ], [ "Xie", "Yucheng", "" ], [ "Wang", "Jing", "" ], [ "Geng", "Xin", "" ] ]	TITLE: WAVE: Weight Templates for Adaptive Initialization of Variable-sized Models ABSTRACT: The growing complexity of model parameters underscores the significance of pre-trained models. However, deployment constraints often necessitate models of varying sizes, exposing limitations in the conventional pre-training and fine-tuning paradigm, particularly when target model sizes are incompatible with pre-trained ones. To address this challenge, we propose WAVE, a novel approach that reformulates variable-sized model initialization from a multi-task perspective, where initializing each model size is treated as a distinct task. WAVE employs shared, size-agnostic weight templates alongside size-specific weight scalers to achieve consistent initialization across various model sizes. These weight templates, constructed within the Learngene framework, integrate knowledge from pre-trained models through a distillation process constrained by Kronecker-based rules. Target models are then initialized by concatenating and weighting these templates, with adaptive connection rules established by lightweight weight scalers, whose parameters are learned from minimal training data. Extensive experiments demonstrate the efficiency of WAVE, achieving state-of-the-art performance in initializing models of various depth and width. The knowledge encapsulated in weight templates is also task-agnostic, allowing for seamless transfer across diverse downstream datasets. Code will be made available at https://github.com/fu-feng/WAVE.
2406.18333	Hossein Ranjbar	Hossein Ranjbar, Alireza Taheri	Continuous Sign Language Recognition Using Intra-inter Gloss Attention	null	null	10.1007/s11042-025-20721-5	null	cs.CV cs.AI	http://creativecommons.org/licenses/by-nc-nd/4.0/	Many continuous sign language recognition (CSLR) studies adopt transformer-based architectures for sequence modeling due to their powerful capacity for capturing global contexts. Nevertheless, vanilla self-attention, which serves as the core module of the transformer, calculates a weighted average over all time steps; therefore, the local temporal semantics of sign videos may not be fully exploited. In this study, we introduce a novel module in sign language recognition studies, called intra-inter gloss attention module, to leverage the relationships among frames within glosses and the semantic and grammatical dependencies between glosses in the video. In the intra-gloss attention module, the video is divided into equally sized chunks and a self-attention mechanism is applied within each chunk. This localized self-attention significantly reduces complexity and eliminates noise introduced by considering non-relative frames. In the inter-gloss attention module, we first aggregate the chunk-level features within each gloss chunk by average pooling along the temporal dimension. Subsequently, multi-head self-attention is applied to all chunk-level features. Given the non-significance of the signer-environment interaction, we utilize segmentation to remove the background of the videos. This enables the proposed model to direct its focus toward the signer. Experimental results on the PHOENIX-2014 benchmark dataset demonstrate that our method can effectively extract sign language features in an end-to-end manner without any prior knowledge, improve the accuracy of CSLR, and achieve the word error rate (WER) of 20.4 on the test set which is a competitive result compare to the state-of-the-art which uses additional supervisions.	[ { "version": "v1", "created": "Wed, 26 Jun 2024 13:21:08 GMT" } ]	2025-03-18T00:00:00	[ [ "Ranjbar", "Hossein", "" ], [ "Taheri", "Alireza", "" ] ]	TITLE: Continuous Sign Language Recognition Using Intra-inter Gloss Attention ABSTRACT: Many continuous sign language recognition (CSLR) studies adopt transformer-based architectures for sequence modeling due to their powerful capacity for capturing global contexts. Nevertheless, vanilla self-attention, which serves as the core module of the transformer, calculates a weighted average over all time steps; therefore, the local temporal semantics of sign videos may not be fully exploited. In this study, we introduce a novel module in sign language recognition studies, called intra-inter gloss attention module, to leverage the relationships among frames within glosses and the semantic and grammatical dependencies between glosses in the video. In the intra-gloss attention module, the video is divided into equally sized chunks and a self-attention mechanism is applied within each chunk. This localized self-attention significantly reduces complexity and eliminates noise introduced by considering non-relative frames. In the inter-gloss attention module, we first aggregate the chunk-level features within each gloss chunk by average pooling along the temporal dimension. Subsequently, multi-head self-attention is applied to all chunk-level features. Given the non-significance of the signer-environment interaction, we utilize segmentation to remove the background of the videos. This enables the proposed model to direct its focus toward the signer. Experimental results on the PHOENIX-2014 benchmark dataset demonstrate that our method can effectively extract sign language features in an end-to-end manner without any prior knowledge, improve the accuracy of CSLR, and achieve the word error rate (WER) of 20.4 on the test set which is a competitive result compare to the state-of-the-art which uses additional supervisions.
2406.18345	Yi Ding	Yi Ding, Chengxuan Tong, Shuailei Zhang, Muyun Jiang, Yong Li, Kevin Lim Jun Liang, Cuntai Guan	EmT: A Novel Transformer for Generalized Cross-subject EEG Emotion Recognition	12 pages, 9 figures. This work has been accepted by IEEE TNNLS	null	10.1109/TNNLS.2025.3552603	null	cs.LG eess.SP	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Integrating prior knowledge of neurophysiology into neural network architecture enhances the performance of emotion decoding. While numerous techniques emphasize learning spatial and short-term temporal patterns, there has been limited emphasis on capturing the vital long-term contextual information associated with emotional cognitive processes. In order to address this discrepancy, we introduce a novel transformer model called emotion transformer (EmT). EmT is designed to excel in both generalized cross-subject EEG emotion classification and regression tasks. In EmT, EEG signals are transformed into a temporal graph format, creating a sequence of EEG feature graphs using a temporal graph construction module (TGC). A novel residual multi-view pyramid GCN module (RMPG) is then proposed to learn dynamic graph representations for each EEG feature graph within the series, and the learned representations of each graph are fused into one token. Furthermore, we design a temporal contextual transformer module (TCT) with two types of token mixers to learn the temporal contextual information. Finally, the task-specific output module (TSO) generates the desired outputs. Experiments on four publicly available datasets show that EmT achieves higher results than the baseline methods for both EEG emotion classification and regression tasks. The code is available at https://github.com/yi-ding-cs/EmT.	[ { "version": "v1", "created": "Wed, 26 Jun 2024 13:42:11 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 05:17:27 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 02:22:04 GMT" } ]	2025-03-18T00:00:00	[ [ "Ding", "Yi", "" ], [ "Tong", "Chengxuan", "" ], [ "Zhang", "Shuailei", "" ], [ "Jiang", "Muyun", "" ], [ "Li", "Yong", "" ], [ "Liang", "Kevin Lim Jun", "" ], [ "Guan", "Cuntai", "" ] ]	TITLE: EmT: A Novel Transformer for Generalized Cross-subject EEG Emotion Recognition ABSTRACT: Integrating prior knowledge of neurophysiology into neural network architecture enhances the performance of emotion decoding. While numerous techniques emphasize learning spatial and short-term temporal patterns, there has been limited emphasis on capturing the vital long-term contextual information associated with emotional cognitive processes. In order to address this discrepancy, we introduce a novel transformer model called emotion transformer (EmT). EmT is designed to excel in both generalized cross-subject EEG emotion classification and regression tasks. In EmT, EEG signals are transformed into a temporal graph format, creating a sequence of EEG feature graphs using a temporal graph construction module (TGC). A novel residual multi-view pyramid GCN module (RMPG) is then proposed to learn dynamic graph representations for each EEG feature graph within the series, and the learned representations of each graph are fused into one token. Furthermore, we design a temporal contextual transformer module (TCT) with two types of token mixers to learn the temporal contextual information. Finally, the task-specific output module (TSO) generates the desired outputs. Experiments on four publicly available datasets show that EmT achieves higher results than the baseline methods for both EEG emotion classification and regression tasks. The code is available at https://github.com/yi-ding-cs/EmT.
2406.18894	Vasileios Kouliaridis	Vasileios Kouliaridis, Georgios Karopoulos, Georgios Kambourakis	Assessing the Effectiveness of LLMs in Android Application Vulnerability Analysis	null	null	10.1007/978-3-031-85593-1_9	null	cs.CR	http://creativecommons.org/licenses/by/4.0/	The increasing frequency of attacks on Android applications coupled with the recent popularity of large language models (LLMs) necessitates a comprehensive understanding of the capabilities of the latter in identifying potential vulnerabilities, which is key to mitigate the overall risk. To this end, the work at hand compares the ability of nine state-of-the-art LLMs to detect Android code vulnerabilities listed in the latest Open Worldwide Application Security Project (OWASP) Mobile Top 10. Each LLM was evaluated against an open dataset of over 100 vulnerable code samples, including obfuscated ones, assessing each model's ability to identify key vulnerabilities. Our analysis reveals the strengths and weaknesses of each LLM, identifying important factors that contribute to their performance. Additionally, we offer insights into context augmentation with retrieval-augmented generation (RAG) for detecting Android code vulnerabilities, which in turn may propel secure application development. Finally, while the reported findings regarding code vulnerability analysis show promise, they also reveal significant discrepancies among the different LLMs.	[ { "version": "v1", "created": "Thu, 27 Jun 2024 05:14:34 GMT" } ]	2025-03-18T00:00:00	[ [ "Kouliaridis", "Vasileios", "" ], [ "Karopoulos", "Georgios", "" ], [ "Kambourakis", "Georgios", "" ] ]	TITLE: Assessing the Effectiveness of LLMs in Android Application Vulnerability Analysis ABSTRACT: The increasing frequency of attacks on Android applications coupled with the recent popularity of large language models (LLMs) necessitates a comprehensive understanding of the capabilities of the latter in identifying potential vulnerabilities, which is key to mitigate the overall risk. To this end, the work at hand compares the ability of nine state-of-the-art LLMs to detect Android code vulnerabilities listed in the latest Open Worldwide Application Security Project (OWASP) Mobile Top 10. Each LLM was evaluated against an open dataset of over 100 vulnerable code samples, including obfuscated ones, assessing each model's ability to identify key vulnerabilities. Our analysis reveals the strengths and weaknesses of each LLM, identifying important factors that contribute to their performance. Additionally, we offer insights into context augmentation with retrieval-augmented generation (RAG) for detecting Android code vulnerabilities, which in turn may propel secure application development. Finally, while the reported findings regarding code vulnerability analysis show promise, they also reveal significant discrepancies among the different LLMs.
2407.03605	Xiaoxia Liu	Xiaoxia Liu, Shijie Yu, Jian Lu, Xiaojun Chen	Orthogonal Constrained Minimization with Tensor $\ell_{2,p}$ Regularization for HSI Denoising and Destriping	null	null	null	null	math.OC cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Hyperspectral images (HSIs) are often contaminated by a mixture of noises such as Gaussian noise, dead lines, stripes, and so on. In this paper, we propose a novel approach for HSI denoising and destriping, called NLTL2p, which consists of an orthogonal constrained minimization model and an iterative algorithm with convergence guarantees. The model of the proposed NLTL2p approach is built based on a new sparsity-enhanced Nonlocal Low-rank Tensor regularization and a tensor $\ell_{2,p}$ norm with $p\in(0,1)$. The low-rank constraints for HSI denoising utilize the spatial nonlocal self-similarity and spectral correlation of HSIs and are formulated based on independent higher-order singular value decomposition with sparsity enhancement on its core tensor to prompt more low-rankness. The tensor $\ell_{2,p}$ norm for HSI destriping is extended from the matrix $\ell_{2,p}$ norm. A proximal block coordinate descent algorithm is proposed in the NLTL2p approach to solve the resulting nonconvex nonsmooth minimization with orthogonal constraints. We show any accumulation point of the sequence generated by the proposed algorithm converges to a first-order stationary point, which is defined using three equalities of substationarity, symmetry, and feasibility for orthogonal constraints. In the numerical experiments, we compare the proposed method with state-of-the-art methods including a deep learning based method, and test the methods on both simulated and real HSI datasets. Our proposed NLTL2p method demonstrates outperformance in terms of metrics such as mean peak signal-to-noise ratio as well as visual quality.	[ { "version": "v1", "created": "Thu, 4 Jul 2024 03:33:19 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 03:13:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Xiaoxia", "" ], [ "Yu", "Shijie", "" ], [ "Lu", "Jian", "" ], [ "Chen", "Xiaojun", "" ] ]	TITLE: Orthogonal Constrained Minimization with Tensor $\ell_{2,p}$ Regularization for HSI Denoising and Destriping ABSTRACT: Hyperspectral images (HSIs) are often contaminated by a mixture of noises such as Gaussian noise, dead lines, stripes, and so on. In this paper, we propose a novel approach for HSI denoising and destriping, called NLTL2p, which consists of an orthogonal constrained minimization model and an iterative algorithm with convergence guarantees. The model of the proposed NLTL2p approach is built based on a new sparsity-enhanced Nonlocal Low-rank Tensor regularization and a tensor $\ell_{2,p}$ norm with $p\in(0,1)$. The low-rank constraints for HSI denoising utilize the spatial nonlocal self-similarity and spectral correlation of HSIs and are formulated based on independent higher-order singular value decomposition with sparsity enhancement on its core tensor to prompt more low-rankness. The tensor $\ell_{2,p}$ norm for HSI destriping is extended from the matrix $\ell_{2,p}$ norm. A proximal block coordinate descent algorithm is proposed in the NLTL2p approach to solve the resulting nonconvex nonsmooth minimization with orthogonal constraints. We show any accumulation point of the sequence generated by the proposed algorithm converges to a first-order stationary point, which is defined using three equalities of substationarity, symmetry, and feasibility for orthogonal constraints. In the numerical experiments, we compare the proposed method with state-of-the-art methods including a deep learning based method, and test the methods on both simulated and real HSI datasets. Our proposed NLTL2p method demonstrates outperformance in terms of metrics such as mean peak signal-to-noise ratio as well as visual quality.
2407.05649	Tongzhou Liao	Tongzhou Liao, Barnab\'as P\'oczos	Greener GRASS: Enhancing GNNs with Encoding, Rewiring, and Attention	Published as a conference paper at ICLR 2025	null	null	null	cs.LG cs.AI cs.NE	http://creativecommons.org/licenses/by/4.0/	Graph Neural Networks (GNNs) have become important tools for machine learning on graph-structured data. In this paper, we explore the synergistic combination of graph encoding, graph rewiring, and graph attention, by introducing Graph Attention with Stochastic Structures (GRASS), a novel GNN architecture. GRASS utilizes relative random walk probabilities (RRWP) encoding and a novel decomposed variant (D-RRWP) to efficiently capture structural information. It rewires the input graph by superimposing a random regular graph to enhance long-range information propagation. It also employs a novel additive attention mechanism tailored for graph-structured data. Our empirical evaluations demonstrate that GRASS achieves state-of-the-art performance on multiple benchmark datasets, including a 20.3% reduction in mean absolute error on the ZINC dataset.	[ { "version": "v1", "created": "Mon, 8 Jul 2024 06:21:56 GMT" }, { "version": "v2", "created": "Thu, 18 Jul 2024 07:30:43 GMT" }, { "version": "v3", "created": "Wed, 9 Oct 2024 16:32:11 GMT" }, { "version": "v4", "created": "Sun, 2 Mar 2025 11:37:49 GMT" }, { "version": "v5", "created": "Fri, 14 Mar 2025 23:47:53 GMT" } ]	2025-03-18T00:00:00	[ [ "Liao", "Tongzhou", "" ], [ "Póczos", "Barnabás", "" ] ]	TITLE: Greener GRASS: Enhancing GNNs with Encoding, Rewiring, and Attention ABSTRACT: Graph Neural Networks (GNNs) have become important tools for machine learning on graph-structured data. In this paper, we explore the synergistic combination of graph encoding, graph rewiring, and graph attention, by introducing Graph Attention with Stochastic Structures (GRASS), a novel GNN architecture. GRASS utilizes relative random walk probabilities (RRWP) encoding and a novel decomposed variant (D-RRWP) to efficiently capture structural information. It rewires the input graph by superimposing a random regular graph to enhance long-range information propagation. It also employs a novel additive attention mechanism tailored for graph-structured data. Our empirical evaluations demonstrate that GRASS achieves state-of-the-art performance on multiple benchmark datasets, including a 20.3% reduction in mean absolute error on the ZINC dataset.
2407.05782	Ioannis Tsiamas	Ioannis Tsiamas, Santiago Pascual, Chunghsin Yeh, Joan Serr\`a	Sequential Contrastive Audio-Visual Learning	ICASSP 2025. Version 1 contains more details	null	null	null	cs.SD cs.CV cs.LG cs.MM eess.AS	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Contrastive learning has emerged as a powerful technique in audio-visual representation learning, leveraging the natural co-occurrence of audio and visual modalities in webscale video datasets. However, conventional contrastive audio-visual learning (CAV) methodologies often rely on aggregated representations derived through temporal aggregation, neglecting the intrinsic sequential nature of the data. This oversight raises concerns regarding the ability of standard approaches to capture and utilize fine-grained information within sequences. In response to this limitation, we propose sequential contrastive audiovisual learning (SCAV), which contrasts examples based on their non-aggregated representation space using multidimensional sequential distances. Audio-visual retrieval experiments with the VGGSound and Music datasets demonstrate the effectiveness of SCAV, with up to 3.5x relative improvements in recall against traditional aggregation-based contrastive learning and other previously proposed methods, which utilize more parameters and data. We also show that models trained with SCAV exhibit a significant degree of flexibility regarding the metric employed for retrieval, allowing us to use a hybrid retrieval approach that is both effective and efficient.	[ { "version": "v1", "created": "Mon, 8 Jul 2024 09:45:20 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 13:36:14 GMT" } ]	2025-03-18T00:00:00	[ [ "Tsiamas", "Ioannis", "" ], [ "Pascual", "Santiago", "" ], [ "Yeh", "Chunghsin", "" ], [ "Serrà", "Joan", "" ] ]	TITLE: Sequential Contrastive Audio-Visual Learning ABSTRACT: Contrastive learning has emerged as a powerful technique in audio-visual representation learning, leveraging the natural co-occurrence of audio and visual modalities in webscale video datasets. However, conventional contrastive audio-visual learning (CAV) methodologies often rely on aggregated representations derived through temporal aggregation, neglecting the intrinsic sequential nature of the data. This oversight raises concerns regarding the ability of standard approaches to capture and utilize fine-grained information within sequences. In response to this limitation, we propose sequential contrastive audiovisual learning (SCAV), which contrasts examples based on their non-aggregated representation space using multidimensional sequential distances. Audio-visual retrieval experiments with the VGGSound and Music datasets demonstrate the effectiveness of SCAV, with up to 3.5x relative improvements in recall against traditional aggregation-based contrastive learning and other previously proposed methods, which utilize more parameters and data. We also show that models trained with SCAV exhibit a significant degree of flexibility regarding the metric employed for retrieval, allowing us to use a hybrid retrieval approach that is both effective and efficient.
2407.08227	Catarina Moreira	Chihcheng Hsieh, Catarina Moreira, Isabel Blanco Nobre, Sandra Costa Sousa, Chun Ouyang, Margot Brereton, Joaquim Jorge and Jacinto C. Nascimento	DALL-M: Context-Aware Clinical Data Augmentation with LLMs	null	null	null	null	cs.AI cs.IR cs.LG	http://creativecommons.org/licenses/by-nc-nd/4.0/	X-ray images are vital in medical diagnostics, but their effectiveness is limited without clinical context. Radiologists often find chest X-rays insufficient for diagnosing underlying diseases, necessitating the integration of structured clinical features with radiology reports. To address this, we introduce DALL-M, a novel framework that enhances clinical datasets by generating contextual synthetic data. DALL-M augments structured patient data, including vital signs (e.g., heart rate, oxygen saturation), radiology findings (e.g., lesion presence), and demographic factors. It integrates this tabular data with contextual knowledge extracted from radiology reports and domain-specific resources (e.g., Radiopaedia, Wikipedia), ensuring clinical consistency and reliability. DALL-M follows a three-phase process: (i) clinical context storage, (ii) expert query generation, and (iii) context-aware feature augmentation. Using large language models (LLMs), it generates both contextual synthetic values for existing clinical features and entirely new, clinically relevant features. Applied to 799 cases from the MIMIC-IV dataset, DALL-M expanded the original 9 clinical features to 91. Empirical validation with machine learning models (including Decision Trees, Random Forests, XGBoost, and TabNET) demonstrated a 16.5% improvement in F1 score and a 25% increase in Precision and Recall. DALL-M bridges an important gap in clinical data augmentation by preserving data integrity while enhancing predictive modeling in healthcare. Our results show that integrating LLM-generated synthetic features significantly improves model performance, making DALL-M a scalable and practical approach for AI-driven medical diagnostics.	[ { "version": "v1", "created": "Thu, 11 Jul 2024 07:01:50 GMT" }, { "version": "v2", "created": "Mon, 7 Oct 2024 09:51:46 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 06:25:38 GMT" } ]	2025-03-18T00:00:00	[ [ "Hsieh", "Chihcheng", "" ], [ "Moreira", "Catarina", "" ], [ "Nobre", "Isabel Blanco", "" ], [ "Sousa", "Sandra Costa", "" ], [ "Ouyang", "Chun", "" ], [ "Brereton", "Margot", "" ], [ "Jorge", "Joaquim", "" ], [ "Nascimento", "Jacinto C.", "" ] ]	TITLE: DALL-M: Context-Aware Clinical Data Augmentation with LLMs ABSTRACT: X-ray images are vital in medical diagnostics, but their effectiveness is limited without clinical context. Radiologists often find chest X-rays insufficient for diagnosing underlying diseases, necessitating the integration of structured clinical features with radiology reports. To address this, we introduce DALL-M, a novel framework that enhances clinical datasets by generating contextual synthetic data. DALL-M augments structured patient data, including vital signs (e.g., heart rate, oxygen saturation), radiology findings (e.g., lesion presence), and demographic factors. It integrates this tabular data with contextual knowledge extracted from radiology reports and domain-specific resources (e.g., Radiopaedia, Wikipedia), ensuring clinical consistency and reliability. DALL-M follows a three-phase process: (i) clinical context storage, (ii) expert query generation, and (iii) context-aware feature augmentation. Using large language models (LLMs), it generates both contextual synthetic values for existing clinical features and entirely new, clinically relevant features. Applied to 799 cases from the MIMIC-IV dataset, DALL-M expanded the original 9 clinical features to 91. Empirical validation with machine learning models (including Decision Trees, Random Forests, XGBoost, and TabNET) demonstrated a 16.5% improvement in F1 score and a 25% increase in Precision and Recall. DALL-M bridges an important gap in clinical data augmentation by preserving data integrity while enhancing predictive modeling in healthcare. Our results show that integrating LLM-generated synthetic features significantly improves model performance, making DALL-M a scalable and practical approach for AI-driven medical diagnostics.
2407.09295	Yulong Yang	Yulong Yang, Xinshan Yang, Shuaidong Li, Chenhao Lin, Zhengyu Zhao, Chao Shen, Tianwei Zhang	Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents	Preprint. Work in progress	null	null	null	cs.CR	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The integration of Large Language Models (LLMs) and Multi-modal Large Language Models (MLLMs) into mobile GUI agents has significantly enhanced user efficiency and experience. However, this advancement also introduces potential security vulnerabilities that have yet to be thoroughly explored. In this paper, we present a systematic security investigation of multi-modal mobile GUI agents, addressing this critical gap in the existing literature. Our contributions are twofold: (1) we propose a novel threat modeling methodology, leading to the discovery and feasibility analysis of 34 previously unreported attacks, and (2) we design an attack framework to systematically construct and evaluate these threats. Through a combination of real-world case studies and extensive dataset-driven experiments, we validate the severity and practicality of those attacks, highlighting the pressing need for robust security measures in mobile GUI systems.	[ { "version": "v1", "created": "Fri, 12 Jul 2024 14:30:05 GMT" }, { "version": "v2", "created": "Wed, 17 Jul 2024 13:36:56 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 07:13:53 GMT" } ]	2025-03-18T00:00:00	[ [ "Yang", "Yulong", "" ], [ "Yang", "Xinshan", "" ], [ "Li", "Shuaidong", "" ], [ "Lin", "Chenhao", "" ], [ "Zhao", "Zhengyu", "" ], [ "Shen", "Chao", "" ], [ "Zhang", "Tianwei", "" ] ]	TITLE: Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents ABSTRACT: The integration of Large Language Models (LLMs) and Multi-modal Large Language Models (MLLMs) into mobile GUI agents has significantly enhanced user efficiency and experience. However, this advancement also introduces potential security vulnerabilities that have yet to be thoroughly explored. In this paper, we present a systematic security investigation of multi-modal mobile GUI agents, addressing this critical gap in the existing literature. Our contributions are twofold: (1) we propose a novel threat modeling methodology, leading to the discovery and feasibility analysis of 34 previously unreported attacks, and (2) we design an attack framework to systematically construct and evaluate these threats. Through a combination of real-world case studies and extensive dataset-driven experiments, we validate the severity and practicality of those attacks, highlighting the pressing need for robust security measures in mobile GUI systems.
2407.14500	Rongkun Zheng	Rongkun Zheng, Lu Qi, Xi Chen, Yi Wang, Kun Wang, Yu Qiao, Hengshuang Zhao	ViLLa: Video Reasoning Segmentation with Large Language Model	15 pages,7 figures	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Recent efforts in video reasoning segmentation (VRS) integrate large language models (LLMs) with perception models to localize and track objects via textual instructions, achieving barely satisfactory results in simple scenarios. However, they struggled to discriminate and deduce the objects from user queries in more real-world scenes featured by long durations, multiple objects, rapid motion, and heavy occlusions. In this work, we analyze the underlying causes of these limitations, and present ViLLa: Video reasoning segmentation with Large Language Model. Remarkably, our ViLLa manages to tackle these challenges through multiple core innovations: (1) a context synthesizer that dynamically encodes the user intent with video contexts for accurate reasoning, resolving ambiguities in complex queries, and (2) a hierarchical temporal synchronizer that disentangles multi-object interactions across complex temporal scenarios by modelling multi-object interactions at local and global temporal scales. To enable efficient processing of long videos, ViLLa incorporates (3) a key segment sampler that adaptively partitions long videos into shorter but semantically dense segments for less redundancy. What's more, to promote research in this unexplored area, we construct a VRS benchmark, VideoReasonSeg, featuring different complex scenarios. Our model also exhibits impressive state-of-the-art results on VideoReasonSeg, Ref-YouTube-VOS, Ref-DAVIS17, MeViS, and ReVOS. Both quantitative and qualitative experiments demonstrate that our method effectively enhances video reasoning segmentation capabilities for multimodal LLMs. The code and dataset will be available at https://github.com/rkzheng99/ViLLa.	[ { "version": "v1", "created": "Thu, 18 Jul 2024 17:59:17 GMT" }, { "version": "v2", "created": "Mon, 29 Jul 2024 13:32:14 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 14:39:54 GMT" } ]	2025-03-18T00:00:00	[ [ "Zheng", "Rongkun", "" ], [ "Qi", "Lu", "" ], [ "Chen", "Xi", "" ], [ "Wang", "Yi", "" ], [ "Wang", "Kun", "" ], [ "Qiao", "Yu", "" ], [ "Zhao", "Hengshuang", "" ] ]	TITLE: ViLLa: Video Reasoning Segmentation with Large Language Model ABSTRACT: Recent efforts in video reasoning segmentation (VRS) integrate large language models (LLMs) with perception models to localize and track objects via textual instructions, achieving barely satisfactory results in simple scenarios. However, they struggled to discriminate and deduce the objects from user queries in more real-world scenes featured by long durations, multiple objects, rapid motion, and heavy occlusions. In this work, we analyze the underlying causes of these limitations, and present ViLLa: Video reasoning segmentation with Large Language Model. Remarkably, our ViLLa manages to tackle these challenges through multiple core innovations: (1) a context synthesizer that dynamically encodes the user intent with video contexts for accurate reasoning, resolving ambiguities in complex queries, and (2) a hierarchical temporal synchronizer that disentangles multi-object interactions across complex temporal scenarios by modelling multi-object interactions at local and global temporal scales. To enable efficient processing of long videos, ViLLa incorporates (3) a key segment sampler that adaptively partitions long videos into shorter but semantically dense segments for less redundancy. What's more, to promote research in this unexplored area, we construct a VRS benchmark, VideoReasonSeg, featuring different complex scenarios. Our model also exhibits impressive state-of-the-art results on VideoReasonSeg, Ref-YouTube-VOS, Ref-DAVIS17, MeViS, and ReVOS. Both quantitative and qualitative experiments demonstrate that our method effectively enhances video reasoning segmentation capabilities for multimodal LLMs. The code and dataset will be available at https://github.com/rkzheng99/ViLLa.
2407.14850	Nizhuan Wang	Yueyang Li, Weiming Zeng, Wenhao Dong, Di Han, Lei Chen, Hongyu Chen, Zijian Kang, Shengyu Gong, Hongjie Yan, Wai Ting Siok, and Nizhuan Wang	A Tale of Single-channel Electroencephalogram: Devices, Datasets, Signal Processing, Applications, and Future Directions	null	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Single-channel electroencephalogram (EEG) is a cost-effective, comfortable, and non-invasive method for monitoring brain activity, widely adopted by researchers, consumers, and clinicians. The increasing number and proportion of articles on single-channel EEG underscore its growing potential. This paper provides a comprehensive review of single-channel EEG, focusing on development trends, devices, datasets, signal processing methods, recent applications, and future directions. Definitions of bipolar and unipolar configurations in single-channel EEG are clarified to guide future advancements. Applications mainly span sleep staging, emotion recognition, educational research, and clinical diagnosis. Ongoing advancements of single-channel EEG in AI-based EEG generation techniques suggest potential parity or superiority over multichannel EEG performance.	[ { "version": "v1", "created": "Sat, 20 Jul 2024 11:36:17 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 02:13:30 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Yueyang", "" ], [ "Zeng", "Weiming", "" ], [ "Dong", "Wenhao", "" ], [ "Han", "Di", "" ], [ "Chen", "Lei", "" ], [ "Chen", "Hongyu", "" ], [ "Kang", "Zijian", "" ], [ "Gong", "Shengyu", "" ], [ "Yan", "Hongjie", "" ], [ "Siok", "Wai Ting", "" ], [ "Wang", "Nizhuan", "" ] ]	TITLE: A Tale of Single-channel Electroencephalogram: Devices, Datasets, Signal Processing, Applications, and Future Directions ABSTRACT: Single-channel electroencephalogram (EEG) is a cost-effective, comfortable, and non-invasive method for monitoring brain activity, widely adopted by researchers, consumers, and clinicians. The increasing number and proportion of articles on single-channel EEG underscore its growing potential. This paper provides a comprehensive review of single-channel EEG, focusing on development trends, devices, datasets, signal processing methods, recent applications, and future directions. Definitions of bipolar and unipolar configurations in single-channel EEG are clarified to guide future advancements. Applications mainly span sleep staging, emotion recognition, educational research, and clinical diagnosis. Ongoing advancements of single-channel EEG in AI-based EEG generation techniques suggest potential parity or superiority over multichannel EEG performance.
2407.16008	Jiaming Shen	Jiaming Shen, Ran Xu, Yennie Jun, Zhen Qin, Tianqi Liu, Carl Yang, Yi Liang, Simon Baumgartner, Michael Bendersky	Boosting Reward Model with Preference-Conditional Multi-Aspect Synthetic Data Generation	ICLR 2025 SSI-FM version	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Reward models (RMs) are crucial for aligning large language models (LLMs) with human preferences. They are trained using preference datasets where each example consists of one input prompt, two responses, and a preference label. As curating a high-quality human labeled preference dataset is both time-consuming and expensive, people often rely on existing powerful LLMs for preference label generation. This can potentially introduce noise and impede RM training. In this work, we present RMBoost, a novel synthetic preference data generation paradigm to boost reward model quality. Unlike traditional methods, which generate two responses before obtaining the preference label, RMBoost first generates one response and selects a preference label, followed by generating the second more (or less) preferred response conditioned on the pre-selected preference label and the first response. This approach offers two main advantages. First, RMBoost reduces labeling noise since preference pairs are constructed intentionally. Second, RMBoost facilitates the creation of more diverse responses by incorporating various quality aspects (e.g., helpfulness, relevance, completeness) into the prompts. We conduct extensive experiments across three diverse datasets and demonstrate that RMBoost outperforms other synthetic preference data generation techniques and significantly boosts the performance of four distinct reward models.	[ { "version": "v1", "created": "Mon, 22 Jul 2024 19:21:55 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 20:08:08 GMT" } ]	2025-03-18T00:00:00	[ [ "Shen", "Jiaming", "" ], [ "Xu", "Ran", "" ], [ "Jun", "Yennie", "" ], [ "Qin", "Zhen", "" ], [ "Liu", "Tianqi", "" ], [ "Yang", "Carl", "" ], [ "Liang", "Yi", "" ], [ "Baumgartner", "Simon", "" ], [ "Bendersky", "Michael", "" ] ]	TITLE: Boosting Reward Model with Preference-Conditional Multi-Aspect Synthetic Data Generation ABSTRACT: Reward models (RMs) are crucial for aligning large language models (LLMs) with human preferences. They are trained using preference datasets where each example consists of one input prompt, two responses, and a preference label. As curating a high-quality human labeled preference dataset is both time-consuming and expensive, people often rely on existing powerful LLMs for preference label generation. This can potentially introduce noise and impede RM training. In this work, we present RMBoost, a novel synthetic preference data generation paradigm to boost reward model quality. Unlike traditional methods, which generate two responses before obtaining the preference label, RMBoost first generates one response and selects a preference label, followed by generating the second more (or less) preferred response conditioned on the pre-selected preference label and the first response. This approach offers two main advantages. First, RMBoost reduces labeling noise since preference pairs are constructed intentionally. Second, RMBoost facilitates the creation of more diverse responses by incorporating various quality aspects (e.g., helpfulness, relevance, completeness) into the prompts. We conduct extensive experiments across three diverse datasets and demonstrate that RMBoost outperforms other synthetic preference data generation techniques and significantly boosts the performance of four distinct reward models.
2407.17390	Itamar Trainin	Itamar Trainin, Omri Abend	$T^5Score$: A Methodology for Automatically Assessing the Quality of LLM Generated Multi-Document Topic Sets	null	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Using LLMs for Multi-Document Topic Extraction has recently gained popularity due to their apparent high-quality outputs, expressiveness, and ease of use. However, most existing evaluation practices are not designed for LLM-generated topics and result in low inter-annotator agreement scores, hindering the reliable use of LLMs for the task. To address this, we introduce $T^5Score$, an evaluation methodology that decomposes the quality of a topic set into quantifiable aspects, measurable through easy-to-perform annotation tasks. This framing enables a convenient, manual or automatic, evaluation procedure resulting in a strong inter-annotator agreement score. To substantiate our methodology and claims, we perform extensive experimentation on multiple datasets and report the results.	[ { "version": "v1", "created": "Wed, 24 Jul 2024 16:14:15 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 08:21:34 GMT" } ]	2025-03-18T00:00:00	[ [ "Trainin", "Itamar", "" ], [ "Abend", "Omri", "" ] ]	TITLE: $T^5Score$: A Methodology for Automatically Assessing the Quality of LLM Generated Multi-Document Topic Sets ABSTRACT: Using LLMs for Multi-Document Topic Extraction has recently gained popularity due to their apparent high-quality outputs, expressiveness, and ease of use. However, most existing evaluation practices are not designed for LLM-generated topics and result in low inter-annotator agreement scores, hindering the reliable use of LLMs for the task. To address this, we introduce $T^5Score$, an evaluation methodology that decomposes the quality of a topic set into quantifiable aspects, measurable through easy-to-perform annotation tasks. This framing enables a convenient, manual or automatic, evaluation procedure resulting in a strong inter-annotator agreement score. To substantiate our methodology and claims, we perform extensive experimentation on multiple datasets and report the results.
2407.19675	Wulian Yun	Wulian Yun, Mengshi Qi, Fei Peng, Huadong Ma	Semi-Supervised Teacher-Reference-Student Architecture for Action Quality Assessment	To be published in ECCV2024	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Existing action quality assessment (AQA) methods often require a large number of label annotations for fully supervised learning, which are laborious and expensive. In practice, the labeled data are difficult to obtain because the AQA annotation process requires domain-specific expertise. In this paper, we propose a novel semi-supervised method, which can be utilized for better assessment of the AQA task by exploiting a large amount of unlabeled data and a small portion of labeled data. Differing from the traditional teacher-student network, we propose a teacher-reference-student architecture to learn both unlabeled and labeled data, where the teacher network and the reference network are used to generate pseudo-labels for unlabeled data to supervise the student network. Specifically, the teacher predicts pseudo-labels by capturing high-level features of unlabeled data. The reference network provides adequate supervision of the student network by referring to additional action information. Moreover, we introduce confidence memory to improve the reliability of pseudo-labels by storing the most accurate ever output of the teacher network and reference network. To validate our method, we conduct extensive experiments on three AQA benchmark datasets. Experimental results show that our method achieves significant improvements and outperforms existing semi-supervised AQA methods.	[ { "version": "v1", "created": "Mon, 29 Jul 2024 03:36:39 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 08:12:47 GMT" } ]	2025-03-18T00:00:00	[ [ "Yun", "Wulian", "" ], [ "Qi", "Mengshi", "" ], [ "Peng", "Fei", "" ], [ "Ma", "Huadong", "" ] ]	TITLE: Semi-Supervised Teacher-Reference-Student Architecture for Action Quality Assessment ABSTRACT: Existing action quality assessment (AQA) methods often require a large number of label annotations for fully supervised learning, which are laborious and expensive. In practice, the labeled data are difficult to obtain because the AQA annotation process requires domain-specific expertise. In this paper, we propose a novel semi-supervised method, which can be utilized for better assessment of the AQA task by exploiting a large amount of unlabeled data and a small portion of labeled data. Differing from the traditional teacher-student network, we propose a teacher-reference-student architecture to learn both unlabeled and labeled data, where the teacher network and the reference network are used to generate pseudo-labels for unlabeled data to supervise the student network. Specifically, the teacher predicts pseudo-labels by capturing high-level features of unlabeled data. The reference network provides adequate supervision of the student network by referring to additional action information. Moreover, we introduce confidence memory to improve the reliability of pseudo-labels by storing the most accurate ever output of the teacher network and reference network. To validate our method, we conduct extensive experiments on three AQA benchmark datasets. Experimental results show that our method achieves significant improvements and outperforms existing semi-supervised AQA methods.
2407.20361	Aditya Kulkarni	Aditya Kulkarni, Vivek Balachandran, Dinil Mon Divakaran and Tamal Das	From ML to LLM: Evaluating the Robustness of Phishing Webpage Detection Models against Adversarial Attacks	null	null	null	null	cs.CR	http://creativecommons.org/licenses/by-nc-nd/4.0/	Phishing attacks attempt to deceive users into stealing sensitive information, posing a significant cybersecurity threat. Advances in machine learning (ML) and deep learning (DL) have led to the development of numerous phishing webpage detection solutions, but these models remain vulnerable to adversarial attacks. Evaluating their robustness against adversarial phishing webpages is essential. Existing tools contain datasets of pre-designed phishing webpages for a limited number of brands, and lack diversity in phishing features. To address these challenges, we develop PhishOracle, a tool that generates adversarial phishing webpages by embedding diverse phishing features into legitimate webpages. We evaluate the robustness of three existing task-specific models -- Stack model, VisualPhishNet, and Phishpedia -- against PhishOracle-generated adversarial phishing webpages and observe a significant drop in their detection rates. In contrast, a multimodal large language model (MLLM)-based phishing detector demonstrates stronger robustness against these adversarial attacks but still is prone to evasion. Our findings highlight the vulnerability of phishing detection models to adversarial attacks, emphasizing the need for more robust detection approaches. Furthermore, we conduct a user study to evaluate whether PhishOracle-generated adversarial phishing webpages can deceive users. The results show that many of these phishing webpages evade not only existing detection models but also users. We also develop the PhishOracle web app, allowing users to input a legitimate URL, select relevant phishing features and generate a corresponding phishing webpage. All resources will be made publicly available on GitHub.	[ { "version": "v1", "created": "Mon, 29 Jul 2024 18:21:34 GMT" }, { "version": "v2", "created": "Wed, 18 Sep 2024 16:07:40 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 11:39:42 GMT" } ]	2025-03-18T00:00:00	[ [ "Kulkarni", "Aditya", "" ], [ "Balachandran", "Vivek", "" ], [ "Divakaran", "Dinil Mon", "" ], [ "Das", "Tamal", "" ] ]	TITLE: From ML to LLM: Evaluating the Robustness of Phishing Webpage Detection Models against Adversarial Attacks ABSTRACT: Phishing attacks attempt to deceive users into stealing sensitive information, posing a significant cybersecurity threat. Advances in machine learning (ML) and deep learning (DL) have led to the development of numerous phishing webpage detection solutions, but these models remain vulnerable to adversarial attacks. Evaluating their robustness against adversarial phishing webpages is essential. Existing tools contain datasets of pre-designed phishing webpages for a limited number of brands, and lack diversity in phishing features. To address these challenges, we develop PhishOracle, a tool that generates adversarial phishing webpages by embedding diverse phishing features into legitimate webpages. We evaluate the robustness of three existing task-specific models -- Stack model, VisualPhishNet, and Phishpedia -- against PhishOracle-generated adversarial phishing webpages and observe a significant drop in their detection rates. In contrast, a multimodal large language model (MLLM)-based phishing detector demonstrates stronger robustness against these adversarial attacks but still is prone to evasion. Our findings highlight the vulnerability of phishing detection models to adversarial attacks, emphasizing the need for more robust detection approaches. Furthermore, we conduct a user study to evaluate whether PhishOracle-generated adversarial phishing webpages can deceive users. The results show that many of these phishing webpages evade not only existing detection models but also users. We also develop the PhishOracle web app, allowing users to input a legitimate URL, select relevant phishing features and generate a corresponding phishing webpage. All resources will be made publicly available on GitHub.
2407.20640	Peng Ye	Bo Li, Wei Wang, Peng Ye	Improved Bounds for Pure Private Agnostic Learning: Item-Level and User-Level Privacy	Fix some typos	null	null	null	cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Machine Learning has made remarkable progress in a wide range of fields. In many scenarios, learning is performed on datasets involving sensitive information, in which privacy protection is essential for learning algorithms. In this work, we study pure private learning in the agnostic model -- a framework reflecting the learning process in practice. We examine the number of users required under item-level (where each user contributes one example) and user-level (where each user contributes multiple examples) privacy and derive several improved upper bounds. For item-level privacy, our algorithm achieves a near optimal bound for general concept classes. We extend this to the user-level setting, rendering a tighter upper bound than the one proved by Ghazi et al. (2023). Lastly, we consider the problem of learning thresholds under user-level privacy and present an algorithm with a nearly tight user complexity.	[ { "version": "v1", "created": "Tue, 30 Jul 2024 08:35:26 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 13:19:11 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Bo", "" ], [ "Wang", "Wei", "" ], [ "Ye", "Peng", "" ] ]	TITLE: Improved Bounds for Pure Private Agnostic Learning: Item-Level and User-Level Privacy ABSTRACT: Machine Learning has made remarkable progress in a wide range of fields. In many scenarios, learning is performed on datasets involving sensitive information, in which privacy protection is essential for learning algorithms. In this work, we study pure private learning in the agnostic model -- a framework reflecting the learning process in practice. We examine the number of users required under item-level (where each user contributes one example) and user-level (where each user contributes multiple examples) privacy and derive several improved upper bounds. For item-level privacy, our algorithm achieves a near optimal bound for general concept classes. We extend this to the user-level setting, rendering a tighter upper bound than the one proved by Ghazi et al. (2023). Lastly, we consider the problem of learning thresholds under user-level privacy and present an algorithm with a nearly tight user complexity.
2407.21368	Danfeng Guo	Danfeng Guo and Demetri Terzopoulos	Prompting Medical Large Vision-Language Models to Diagnose Pathologies by Visual Question Answering	Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://melba-journal.org/2025:004	Machine.Learning.for.Biomedical.Imaging. 3 (2025)	10.59275/j.melba.2025-1a8b	null	cs.CV cs.AI cs.CL cs.LG	http://creativecommons.org/licenses/by/4.0/	Large Vision-Language Models (LVLMs) have achieved significant success in recent years, and they have been extended to the medical domain. Although demonstrating satisfactory performance on medical Visual Question Answering (VQA) tasks, Medical LVLMs (MLVLMs) suffer from the hallucination problem, which makes them fail to diagnose complex pathologies. Moreover, they readily fail to learn minority pathologies due to imbalanced training data. We propose two prompting strategies for MLVLMs that reduce hallucination and improve VQA performance. In the first strategy, we provide a detailed explanation of the queried pathology. In the second strategy, we fine-tune a cheap, weak learner to achieve high performance on a specific metric, and textually provide its judgment to the MLVLM. Tested on the MIMIC-CXR-JPG and Chexpert datasets, our methods significantly improve the diagnostic F1 score, with the highest increase being 0.27. We also demonstrate that our prompting strategies can be extended to general LVLM domains. Based on POPE metrics, it effectively suppresses the false negative predictions of existing LVLMs and improves Recall by approximately 0.07.	[ { "version": "v1", "created": "Wed, 31 Jul 2024 06:34:38 GMT" }, { "version": "v2", "created": "Sat, 1 Mar 2025 06:14:00 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 00:27:45 GMT" } ]	2025-03-18T00:00:00	[ [ "Guo", "Danfeng", "" ], [ "Terzopoulos", "Demetri", "" ] ]	TITLE: Prompting Medical Large Vision-Language Models to Diagnose Pathologies by Visual Question Answering ABSTRACT: Large Vision-Language Models (LVLMs) have achieved significant success in recent years, and they have been extended to the medical domain. Although demonstrating satisfactory performance on medical Visual Question Answering (VQA) tasks, Medical LVLMs (MLVLMs) suffer from the hallucination problem, which makes them fail to diagnose complex pathologies. Moreover, they readily fail to learn minority pathologies due to imbalanced training data. We propose two prompting strategies for MLVLMs that reduce hallucination and improve VQA performance. In the first strategy, we provide a detailed explanation of the queried pathology. In the second strategy, we fine-tune a cheap, weak learner to achieve high performance on a specific metric, and textually provide its judgment to the MLVLM. Tested on the MIMIC-CXR-JPG and Chexpert datasets, our methods significantly improve the diagnostic F1 score, with the highest increase being 0.27. We also demonstrate that our prompting strategies can be extended to general LVLM domains. Based on POPE metrics, it effectively suppresses the false negative predictions of existing LVLMs and improves Recall by approximately 0.07.
2408.02032	Fushuo Huo	Fushuo Huo, Wenchao Xu, Zhong Zhang, Haozhao Wang, Zhicheng Chen, Peilin Zhao	Self-Introspective Decoding: Alleviating Hallucinations for Large Vision-Language Models	ICLR2025	null	null	null	cs.CV cs.AI	http://creativecommons.org/licenses/by/4.0/	While Large Vision-Language Models (LVLMs) have rapidly advanced in recent years, the prevalent issue known as the `hallucination' problem has emerged as a significant bottleneck, hindering their real-world deployments. Existing methods mitigate this issue mainly from two perspectives: One approach leverages extra knowledge like robust instruction tuning LVLMs with curated datasets or employing auxiliary analysis networks, which inevitable incur additional costs. Another approach, known as contrastive decoding, induces hallucinations by manually disturbing the vision or instruction raw inputs and mitigates them by contrasting the outputs of the disturbed and original LVLMs. However, these approaches rely on empirical holistic input disturbances and double the inference cost. To avoid these issues, we propose a simple yet effective method named Self-Introspective Decoding (SID). Our empirical investigation reveals that pretrained LVLMs can introspectively assess the importance of vision tokens based on preceding vision and text (both instruction and generated) tokens. We develop the Context and Text-aware Token Selection (CT2S) strategy, which preserves only unimportant vision tokens after early layers of LVLMs to adaptively amplify text-informed hallucination during the auto-regressive decoding. This approach ensures that multimodal knowledge absorbed in the early layers induces multimodal contextual rather than aimless hallucinations. Subsequently, the original token logits subtract the amplified vision-and-text association hallucinations, guiding LVLMs decoding faithfully. Extensive experiments illustrate SID generates less-hallucination and higher-quality texts across various metrics, without extra knowledge and much additional computation burdens.	[ { "version": "v1", "created": "Sun, 4 Aug 2024 13:50:17 GMT" }, { "version": "v2", "created": "Tue, 8 Oct 2024 12:26:40 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 06:51:13 GMT" } ]	2025-03-18T00:00:00	[ [ "Huo", "Fushuo", "" ], [ "Xu", "Wenchao", "" ], [ "Zhang", "Zhong", "" ], [ "Wang", "Haozhao", "" ], [ "Chen", "Zhicheng", "" ], [ "Zhao", "Peilin", "" ] ]	TITLE: Self-Introspective Decoding: Alleviating Hallucinations for Large Vision-Language Models ABSTRACT: While Large Vision-Language Models (LVLMs) have rapidly advanced in recent years, the prevalent issue known as the `hallucination' problem has emerged as a significant bottleneck, hindering their real-world deployments. Existing methods mitigate this issue mainly from two perspectives: One approach leverages extra knowledge like robust instruction tuning LVLMs with curated datasets or employing auxiliary analysis networks, which inevitable incur additional costs. Another approach, known as contrastive decoding, induces hallucinations by manually disturbing the vision or instruction raw inputs and mitigates them by contrasting the outputs of the disturbed and original LVLMs. However, these approaches rely on empirical holistic input disturbances and double the inference cost. To avoid these issues, we propose a simple yet effective method named Self-Introspective Decoding (SID). Our empirical investigation reveals that pretrained LVLMs can introspectively assess the importance of vision tokens based on preceding vision and text (both instruction and generated) tokens. We develop the Context and Text-aware Token Selection (CT2S) strategy, which preserves only unimportant vision tokens after early layers of LVLMs to adaptively amplify text-informed hallucination during the auto-regressive decoding. This approach ensures that multimodal knowledge absorbed in the early layers induces multimodal contextual rather than aimless hallucinations. Subsequently, the original token logits subtract the amplified vision-and-text association hallucinations, guiding LVLMs decoding faithfully. Extensive experiments illustrate SID generates less-hallucination and higher-quality texts across various metrics, without extra knowledge and much additional computation burdens.
2408.02833	Costantino Carugno	Costantino Carugno, Maurizio Ferrari Dacrema, Paolo Cremonesi	Adaptive Learning for Quantum Linear Regression	null	null	10.1109/QCE60285.2024.00186	null	quant-ph cs.LG	http://creativecommons.org/licenses/by-nc-nd/4.0/	The recent availability of quantum annealers as cloud-based services has enabled new ways to handle machine learning problems, and several relevant algorithms have been adapted to run on these devices. In a recent work, linear regression was formulated as a quadratic binary optimization problem that can be solved via quantum annealing. Although this approach promises a computational time advantage for large datasets, the quality of the solution is limited by the necessary use of a precision vector, used to approximate the real-numbered regression coefficients in the quantum formulation. In this work, we focus on the practical challenge of improving the precision vector encoding: instead of setting an array of generic values equal for all coefficients, we allow each one to be expressed by its specific precision, which is tuned with a simple adaptive algorithm. This approach is evaluated on synthetic datasets of increasing size, and linear regression is solved using the D-Wave Advantage quantum annealer, as well as classical solvers. To the best of our knowledge, this is the largest dataset ever evaluated for linear regression on a quantum annealer. The results show that our formulation is able to deliver improved solution quality in all instances, and could better exploit the potential of current quantum devices.	[ { "version": "v1", "created": "Mon, 5 Aug 2024 21:09:01 GMT" } ]	2025-03-18T00:00:00	[ [ "Carugno", "Costantino", "" ], [ "Dacrema", "Maurizio Ferrari", "" ], [ "Cremonesi", "Paolo", "" ] ]	TITLE: Adaptive Learning for Quantum Linear Regression ABSTRACT: The recent availability of quantum annealers as cloud-based services has enabled new ways to handle machine learning problems, and several relevant algorithms have been adapted to run on these devices. In a recent work, linear regression was formulated as a quadratic binary optimization problem that can be solved via quantum annealing. Although this approach promises a computational time advantage for large datasets, the quality of the solution is limited by the necessary use of a precision vector, used to approximate the real-numbered regression coefficients in the quantum formulation. In this work, we focus on the practical challenge of improving the precision vector encoding: instead of setting an array of generic values equal for all coefficients, we allow each one to be expressed by its specific precision, which is tuned with a simple adaptive algorithm. This approach is evaluated on synthetic datasets of increasing size, and linear regression is solved using the D-Wave Advantage quantum annealer, as well as classical solvers. To the best of our knowledge, this is the largest dataset ever evaluated for linear regression on a quantum annealer. The results show that our formulation is able to deliver improved solution quality in all instances, and could better exploit the potential of current quantum devices.
2408.04315	Wei Huo	Wei Huo, Changxin Liu, Kemi Ding, Karl Henrik Johansson, Ling Shi	Federated Cubic Regularized Newton Learning with Sparsification-amplified Differential Privacy	null	null	null	null	cs.LG cs.SY eess.SY	http://creativecommons.org/licenses/by/4.0/	This paper investigates the use of the cubic-regularized Newton method within a federated learning framework while addressing two major concerns that commonly arise in federated learning: privacy leakage and communication bottleneck. We introduce a federated learning algorithm called Differentially Private Federated Cubic Regularized Newton (DP-FCRN). By leveraging second-order techniques, our algorithm achieves lower iteration complexity compared to first-order methods. We also incorporate noise perturbation during local computations to ensure privacy. Furthermore, we employ sparsification in uplink transmission, which not only reduces the communication costs but also amplifies the privacy guarantee. Specifically, this approach reduces the necessary noise intensity without compromising privacy protection. We analyze the convergence properties of our algorithm and establish the privacy guarantee. Finally, we validate the effectiveness of the proposed algorithm through experiments on a benchmark dataset.	[ { "version": "v1", "created": "Thu, 8 Aug 2024 08:48:54 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 08:45:25 GMT" } ]	2025-03-18T00:00:00	[ [ "Huo", "Wei", "" ], [ "Liu", "Changxin", "" ], [ "Ding", "Kemi", "" ], [ "Johansson", "Karl Henrik", "" ], [ "Shi", "Ling", "" ] ]	TITLE: Federated Cubic Regularized Newton Learning with Sparsification-amplified Differential Privacy ABSTRACT: This paper investigates the use of the cubic-regularized Newton method within a federated learning framework while addressing two major concerns that commonly arise in federated learning: privacy leakage and communication bottleneck. We introduce a federated learning algorithm called Differentially Private Federated Cubic Regularized Newton (DP-FCRN). By leveraging second-order techniques, our algorithm achieves lower iteration complexity compared to first-order methods. We also incorporate noise perturbation during local computations to ensure privacy. Furthermore, we employ sparsification in uplink transmission, which not only reduces the communication costs but also amplifies the privacy guarantee. Specifically, this approach reduces the necessary noise intensity without compromising privacy protection. We analyze the convergence properties of our algorithm and establish the privacy guarantee. Finally, we validate the effectiveness of the proposed algorithm through experiments on a benchmark dataset.
2408.11470	Panfeng Liu	Panfeng Liu and Guoliang Qiu and Biaoshuai Tao and Kuan Yang	A Thorough Comparison Between Independent Cascade and Susceptible-Infected-Recovered Models	30 pages, 6 figures	null	null	null	cs.SI physics.soc-ph	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We study cascades in social networks with the independent cascade (IC) model and the Susceptible-Infected-recovered (SIR) model. The well-studied IC model fails to capture the feature of node recovery, and the SIR model is a variant of the IC model with the node recovery feature. In the SIR model, by computing the probability that a node successfully infects another before its recovery and viewing this probability as the corresponding IC parameter, the SIR model becomes an "out-going-edge-correlated" version of the IC model: the events of the infections along different out-going edges of a node become dependent in the SIR model, whereas these events are independent in the IC model. In this paper, we thoroughly compare the two models and examine the effect of this extra dependency in the SIR model. By a carefully designed coupling argument, we show that the seeds in the IC model have a stronger influence spread than their counterparts in the SIR model, and sometimes it can be significantly stronger. Specifically, we prove that, given the same network, the same seed sets, and the parameters of the two models being set based on the above-mentioned equivalence, the expected number of infected nodes at the end of the cascade for the IC model is weakly larger than that for the SIR model, and there are instances where this dominance is significant. We also study the influence maximization problem with the SIR model. We show that the above-mentioned difference in the two models yields different seed-selection strategies, which motivates the design of influence maximization algorithms specifically for the SIR model. We design efficient approximation algorithms with theoretical guarantees by adapting the reverse-reachable-set-based algorithms, commonly used for the IC model, to the SIR model. Finally, we conduct experimental studies over real-world datasets.	[ { "version": "v1", "created": "Wed, 21 Aug 2024 09:38:41 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 15:25:19 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Panfeng", "" ], [ "Qiu", "Guoliang", "" ], [ "Tao", "Biaoshuai", "" ], [ "Yang", "Kuan", "" ] ]	TITLE: A Thorough Comparison Between Independent Cascade and Susceptible-Infected-Recovered Models ABSTRACT: We study cascades in social networks with the independent cascade (IC) model and the Susceptible-Infected-recovered (SIR) model. The well-studied IC model fails to capture the feature of node recovery, and the SIR model is a variant of the IC model with the node recovery feature. In the SIR model, by computing the probability that a node successfully infects another before its recovery and viewing this probability as the corresponding IC parameter, the SIR model becomes an "out-going-edge-correlated" version of the IC model: the events of the infections along different out-going edges of a node become dependent in the SIR model, whereas these events are independent in the IC model. In this paper, we thoroughly compare the two models and examine the effect of this extra dependency in the SIR model. By a carefully designed coupling argument, we show that the seeds in the IC model have a stronger influence spread than their counterparts in the SIR model, and sometimes it can be significantly stronger. Specifically, we prove that, given the same network, the same seed sets, and the parameters of the two models being set based on the above-mentioned equivalence, the expected number of infected nodes at the end of the cascade for the IC model is weakly larger than that for the SIR model, and there are instances where this dominance is significant. We also study the influence maximization problem with the SIR model. We show that the above-mentioned difference in the two models yields different seed-selection strategies, which motivates the design of influence maximization algorithms specifically for the SIR model. We design efficient approximation algorithms with theoretical guarantees by adapting the reverse-reachable-set-based algorithms, commonly used for the IC model, to the SIR model. Finally, we conduct experimental studies over real-world datasets.
2408.12871	Xiaochen Zhou	Zhou Xiaochen, Liang Xingzhou, Zou Hui, Lu Yi, Qu Jingjing	DeepDiveAI: Identifying AI Related Documents in Large Scale Literature Data	null	null	null	null	cs.AI	http://creativecommons.org/licenses/by/4.0/	In this paper, we propose a method to automatically classify AI-related documents from large-scale literature databases, leading to the creation of an AI-related literature dataset, named DeepDiveAI. The dataset construction approach integrates expert knowledge with the capabilities of advanced models, structured across two global stages. In the first stage, expert-curated classification datasets are used to train an LSTM model, which classifies coarse AI related records from large-scale datasets. In the second stage, we use Qwen2.5 Plus to annotate a random 10% of the coarse AI-related records, which are then used to train a BERT binary classifier. This step further refines the coarse AI related record set to obtain the final DeepDiveAI dataset. Evaluation results demonstrate that the entire workflow can efficiently and accurately identify AI-related literature from large-scale datasets.	[ { "version": "v1", "created": "Fri, 23 Aug 2024 07:05:12 GMT" }, { "version": "v2", "created": "Wed, 28 Aug 2024 11:30:28 GMT" }, { "version": "v3", "created": "Tue, 8 Oct 2024 07:21:57 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 12:46:22 GMT" } ]	2025-03-18T00:00:00	[ [ "Xiaochen", "Zhou", "" ], [ "Xingzhou", "Liang", "" ], [ "Hui", "Zou", "" ], [ "Yi", "Lu", "" ], [ "Jingjing", "Qu", "" ] ]	TITLE: DeepDiveAI: Identifying AI Related Documents in Large Scale Literature Data ABSTRACT: In this paper, we propose a method to automatically classify AI-related documents from large-scale literature databases, leading to the creation of an AI-related literature dataset, named DeepDiveAI. The dataset construction approach integrates expert knowledge with the capabilities of advanced models, structured across two global stages. In the first stage, expert-curated classification datasets are used to train an LSTM model, which classifies coarse AI related records from large-scale datasets. In the second stage, we use Qwen2.5 Plus to annotate a random 10% of the coarse AI-related records, which are then used to train a BERT binary classifier. This step further refines the coarse AI related record set to obtain the final DeepDiveAI dataset. Evaluation results demonstrate that the entire workflow can efficiently and accurately identify AI-related literature from large-scale datasets.
2408.15185	Ghazal Alinezhad Noghre	Ghazal Alinezhad Noghre, Armin Danesh Pazho, Hamed Tabkhi	Human-Centric Video Anomaly Detection Through Spatio-Temporal Pose Tokenization and Transformer	null	null	null	null	cs.CV cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Video Anomaly Detection (VAD) presents a significant challenge in computer vision, particularly due to the unpredictable and infrequent nature of anomalous events, coupled with the diverse and dynamic environments in which they occur. Human-centric VAD, a specialized area within this domain, faces additional complexities, including variations in human behavior, potential biases in data, and substantial privacy concerns related to human subjects. These issues complicate the development of models that are both robust and generalizable. To address these challenges, recent advancements have focused on pose-based VAD, which leverages human pose as a high-level feature to mitigate privacy concerns, reduce appearance biases, and minimize background interference. In this paper, we introduce SPARTA, a novel transformer-based architecture designed specifically for human-centric pose-based VAD. SPARTA introduces an innovative Spatio-Temporal Pose and Relative Pose (ST-PRP) tokenization method that produces an enriched representation of human motion over time. This approach ensures that the transformer's attention mechanism captures both spatial and temporal patterns simultaneously, rather than focusing on only one aspect. The addition of the relative pose further emphasizes subtle deviations from normal human movements. The architecture's core, a novel Unified Encoder Twin Decoders (UETD) transformer, significantly improves the detection of anomalous behaviors in video data. Extensive evaluations across multiple benchmark datasets demonstrate that SPARTA consistently outperforms existing methods, establishing a new state-of-the-art in pose-based VAD.	[ { "version": "v1", "created": "Tue, 27 Aug 2024 16:40:14 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 14:05:49 GMT" } ]	2025-03-18T00:00:00	[ [ "Noghre", "Ghazal Alinezhad", "" ], [ "Pazho", "Armin Danesh", "" ], [ "Tabkhi", "Hamed", "" ] ]	TITLE: Human-Centric Video Anomaly Detection Through Spatio-Temporal Pose Tokenization and Transformer ABSTRACT: Video Anomaly Detection (VAD) presents a significant challenge in computer vision, particularly due to the unpredictable and infrequent nature of anomalous events, coupled with the diverse and dynamic environments in which they occur. Human-centric VAD, a specialized area within this domain, faces additional complexities, including variations in human behavior, potential biases in data, and substantial privacy concerns related to human subjects. These issues complicate the development of models that are both robust and generalizable. To address these challenges, recent advancements have focused on pose-based VAD, which leverages human pose as a high-level feature to mitigate privacy concerns, reduce appearance biases, and minimize background interference. In this paper, we introduce SPARTA, a novel transformer-based architecture designed specifically for human-centric pose-based VAD. SPARTA introduces an innovative Spatio-Temporal Pose and Relative Pose (ST-PRP) tokenization method that produces an enriched representation of human motion over time. This approach ensures that the transformer's attention mechanism captures both spatial and temporal patterns simultaneously, rather than focusing on only one aspect. The addition of the relative pose further emphasizes subtle deviations from normal human movements. The architecture's core, a novel Unified Encoder Twin Decoders (UETD) transformer, significantly improves the detection of anomalous behaviors in video data. Extensive evaluations across multiple benchmark datasets demonstrate that SPARTA consistently outperforms existing methods, establishing a new state-of-the-art in pose-based VAD.
2408.16444	Leandro Car\'isio Fernandes	Leandro Car\'isio Fernandes, Gustavo Bartz Guedes, Thiago Soares Laitz, Thales Sales Almeida, Rodrigo Nogueira, Roberto Lotufo, Jayr Pereira	SurveySum: A Dataset for Summarizing Multiple Scientific Articles into a Survey Section	15 pages, 6 figures, 1 table. Submitted to BRACIS 2024	null	10.1007/978-3-031-79032-4_30	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Document summarization is a task to shorten texts into concise and informative summaries. This paper introduces a novel dataset designed for summarizing multiple scientific articles into a section of a survey. Our contributions are: (1) SurveySum, a new dataset addressing the gap in domain-specific summarization tools; (2) two specific pipelines to summarize scientific articles into a section of a survey; and (3) the evaluation of these pipelines using multiple metrics to compare their performance. Our results highlight the importance of high-quality retrieval stages and the impact of different configurations on the quality of generated summaries.	[ { "version": "v1", "created": "Thu, 29 Aug 2024 11:13:23 GMT" } ]	2025-03-18T00:00:00	[ [ "Fernandes", "Leandro Carísio", "" ], [ "Guedes", "Gustavo Bartz", "" ], [ "Laitz", "Thiago Soares", "" ], [ "Almeida", "Thales Sales", "" ], [ "Nogueira", "Rodrigo", "" ], [ "Lotufo", "Roberto", "" ], [ "Pereira", "Jayr", "" ] ]	TITLE: SurveySum: A Dataset for Summarizing Multiple Scientific Articles into a Survey Section ABSTRACT: Document summarization is a task to shorten texts into concise and informative summaries. This paper introduces a novel dataset designed for summarizing multiple scientific articles into a section of a survey. Our contributions are: (1) SurveySum, a new dataset addressing the gap in domain-specific summarization tools; (2) two specific pipelines to summarize scientific articles into a section of a survey; and (3) the evaluation of these pipelines using multiple metrics to compare their performance. Our results highlight the importance of high-quality retrieval stages and the impact of different configurations on the quality of generated summaries.
2409.07896	Shun Zou	Shun Zou, Zhuo Zhang, Yi Zou, Guangwei Gao	MambaMIC: An Efficient Baseline for Microscopic Image Classification with State Space Models	7 pages, 4 figures	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In recent years, CNN and Transformer-based methods have made significant progress in Microscopic Image Classification (MIC). However, existing approaches still face the dilemma between global modeling and efficient computation. While the Selective State Space Model (SSM) can simulate long-range dependencies with linear complexity, it still encounters challenges in MIC, such as local pixel forgetting, channel redundancy, and lack of local perception. To address these issues, we propose a simple yet efficient vision backbone for MIC tasks, named MambaMIC. Specifically, we introduce a Local-Global dual-branch aggregation module: the MambaMIC Block, designed to effectively capture and fuse local connectivity and global dependencies. In the local branch, we use local convolutions to capture pixel similarity, mitigating local pixel forgetting and enhancing perception. In the global branch, SSM extracts global dependencies, while Locally Aware Enhanced Filter reduces channel redundancy and local pixel forgetting. Additionally, we design a Feature Modulation Interaction Aggregation Module for deep feature interaction and key feature re-localization. Extensive benchmarking shows that MambaMIC achieves state-of-the-art performance across five datasets. code is available at https://zs1314.github.io/MambaMIC	[ { "version": "v1", "created": "Thu, 12 Sep 2024 10:01:33 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 03:18:57 GMT" } ]	2025-03-18T00:00:00	[ [ "Zou", "Shun", "" ], [ "Zhang", "Zhuo", "" ], [ "Zou", "Yi", "" ], [ "Gao", "Guangwei", "" ] ]	TITLE: MambaMIC: An Efficient Baseline for Microscopic Image Classification with State Space Models ABSTRACT: In recent years, CNN and Transformer-based methods have made significant progress in Microscopic Image Classification (MIC). However, existing approaches still face the dilemma between global modeling and efficient computation. While the Selective State Space Model (SSM) can simulate long-range dependencies with linear complexity, it still encounters challenges in MIC, such as local pixel forgetting, channel redundancy, and lack of local perception. To address these issues, we propose a simple yet efficient vision backbone for MIC tasks, named MambaMIC. Specifically, we introduce a Local-Global dual-branch aggregation module: the MambaMIC Block, designed to effectively capture and fuse local connectivity and global dependencies. In the local branch, we use local convolutions to capture pixel similarity, mitigating local pixel forgetting and enhancing perception. In the global branch, SSM extracts global dependencies, while Locally Aware Enhanced Filter reduces channel redundancy and local pixel forgetting. Additionally, we design a Feature Modulation Interaction Aggregation Module for deep feature interaction and key feature re-localization. Extensive benchmarking shows that MambaMIC achieves state-of-the-art performance across five datasets. code is available at https://zs1314.github.io/MambaMIC
2409.08481	Zhuoyuan Li	Zhuoyuan Li, Junqi Liao, Chuanbo Tang, Haotian Zhang, Yuqi Li, Yifan Bian, Xihua Sheng, Xinmin Feng, Yao Li, Changsheng Gao, Li Li, Dong Liu, Feng Wu	USTC-TD: A Test Dataset and Benchmark for Image and Video Coding in 2020s	16 pages. Project Page: https://esakak.github.io/USTC-TD. Supplementary Material: https://zhuoyuanli1997.github.io/files/USTC-TD/sup.pdf	null	null	null	eess.IV cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Image/video coding has been a remarkable research area for both academia and industry for many years. Testing datasets, especially high-quality image/video datasets are desirable for the justified evaluation of coding-related research, practical applications, and standardization activities. We put forward a test dataset namely USTC-TD, which has been successfully adopted in the practical end-to-end image/video coding challenge of the IEEE International Conference on Visual Communications and Image Processing (VCIP) in 2022 and 2023. USTC-TD contains 40 images at 4K spatial resolution and 10 video sequences at 1080p spatial resolution, featuring various content due to the diverse environmental factors (e.g. scene type, texture, motion, view) and the designed imaging factors (e.g. illumination, lens, shadow). We quantitatively evaluate USTC-TD on different image/video features (spatial, temporal, color, lightness), and compare it with the previous image/video test datasets, which verifies its excellent compensation for the shortcomings of existing datasets. We also evaluate both classic standardized and recently learned image/video coding schemes on USTC-TD using objective quality metrics (PSNR, MS-SSIM, VMAF) and subjective quality metric (MOS), providing an extensive benchmark for these evaluated schemes. Based on the characteristics and specific design of the proposed test dataset, we analyze the benchmark performance and shed light on the future research and development of image/video coding. All the data are released online: https://esakak.github.io/USTC-TD.	[ { "version": "v1", "created": "Fri, 13 Sep 2024 02:13:11 GMT" }, { "version": "v2", "created": "Thu, 14 Nov 2024 05:13:21 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 02:09:46 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Zhuoyuan", "" ], [ "Liao", "Junqi", "" ], [ "Tang", "Chuanbo", "" ], [ "Zhang", "Haotian", "" ], [ "Li", "Yuqi", "" ], [ "Bian", "Yifan", "" ], [ "Sheng", "Xihua", "" ], [ "Feng", "Xinmin", "" ], [ "Li", "Yao", "" ], [ "Gao", "Changsheng", "" ], [ "Li", "Li", "" ], [ "Liu", "Dong", "" ], [ "Wu", "Feng", "" ] ]	TITLE: USTC-TD: A Test Dataset and Benchmark for Image and Video Coding in 2020s ABSTRACT: Image/video coding has been a remarkable research area for both academia and industry for many years. Testing datasets, especially high-quality image/video datasets are desirable for the justified evaluation of coding-related research, practical applications, and standardization activities. We put forward a test dataset namely USTC-TD, which has been successfully adopted in the practical end-to-end image/video coding challenge of the IEEE International Conference on Visual Communications and Image Processing (VCIP) in 2022 and 2023. USTC-TD contains 40 images at 4K spatial resolution and 10 video sequences at 1080p spatial resolution, featuring various content due to the diverse environmental factors (e.g. scene type, texture, motion, view) and the designed imaging factors (e.g. illumination, lens, shadow). We quantitatively evaluate USTC-TD on different image/video features (spatial, temporal, color, lightness), and compare it with the previous image/video test datasets, which verifies its excellent compensation for the shortcomings of existing datasets. We also evaluate both classic standardized and recently learned image/video coding schemes on USTC-TD using objective quality metrics (PSNR, MS-SSIM, VMAF) and subjective quality metric (MOS), providing an extensive benchmark for these evaluated schemes. Based on the characteristics and specific design of the proposed test dataset, we analyze the benchmark performance and shed light on the future research and development of image/video coding. All the data are released online: https://esakak.github.io/USTC-TD.
2409.09021	Soumitra Kundu	Soumitra Kundu and Gargi Panda and Saumik Bhattacharya and Aurobinda Routray and Rajlakshmi Guha	INN-PAR: Invertible Neural Network for PPG to ABP Reconstruction	ICASSP 2025	null	10.1109/ICASSP49660.2025.10888915	null	cs.LG cs.HC	http://creativecommons.org/licenses/by/4.0/	Non-invasive and continuous blood pressure (BP) monitoring is essential for the early prevention of many cardiovascular diseases. Estimating arterial blood pressure (ABP) from photoplethysmography (PPG) has emerged as a promising solution. However, existing deep learning approaches for PPG-to-ABP reconstruction (PAR) encounter certain information loss, impacting the precision of the reconstructed signal. To overcome this limitation, we introduce an invertible neural network for PPG to ABP reconstruction (INN-PAR), which employs a series of invertible blocks to jointly learn the mapping between PPG and its gradient with the ABP signal and its gradient. INN-PAR efficiently captures both forward and inverse mappings simultaneously, thereby preventing information loss. By integrating signal gradients into the learning process, INN-PAR enhances the network's ability to capture essential high-frequency details, leading to more accurate signal reconstruction. Moreover, we propose a multi-scale convolution module (MSCM) within the invertible block, enabling the model to learn features across multiple scales effectively. We have experimented on two benchmark datasets, which show that INN-PAR significantly outperforms the state-of-the-art methods in both waveform reconstruction and BP measurement accuracy. Codes can be found at: https://github.com/soumitra1992/INNPAR-PPG2ABP.	[ { "version": "v1", "created": "Fri, 13 Sep 2024 17:48:48 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 17:28:15 GMT" } ]	2025-03-18T00:00:00	[ [ "Kundu", "Soumitra", "" ], [ "Panda", "Gargi", "" ], [ "Bhattacharya", "Saumik", "" ], [ "Routray", "Aurobinda", "" ], [ "Guha", "Rajlakshmi", "" ] ]	TITLE: INN-PAR: Invertible Neural Network for PPG to ABP Reconstruction ABSTRACT: Non-invasive and continuous blood pressure (BP) monitoring is essential for the early prevention of many cardiovascular diseases. Estimating arterial blood pressure (ABP) from photoplethysmography (PPG) has emerged as a promising solution. However, existing deep learning approaches for PPG-to-ABP reconstruction (PAR) encounter certain information loss, impacting the precision of the reconstructed signal. To overcome this limitation, we introduce an invertible neural network for PPG to ABP reconstruction (INN-PAR), which employs a series of invertible blocks to jointly learn the mapping between PPG and its gradient with the ABP signal and its gradient. INN-PAR efficiently captures both forward and inverse mappings simultaneously, thereby preventing information loss. By integrating signal gradients into the learning process, INN-PAR enhances the network's ability to capture essential high-frequency details, leading to more accurate signal reconstruction. Moreover, we propose a multi-scale convolution module (MSCM) within the invertible block, enabling the model to learn features across multiple scales effectively. We have experimented on two benchmark datasets, which show that INN-PAR significantly outperforms the state-of-the-art methods in both waveform reconstruction and BP measurement accuracy. Codes can be found at: https://github.com/soumitra1992/INNPAR-PPG2ABP.
2409.10687	Ruchik Mishra	Ruchik Mishra, Andrew Frye, Madan Mohan Rayguru, Dan O. Popa	Personalized Speech Emotion Recognition in Human-Robot Interaction using Vision Transformers	This work has been accepted for the IEEE Robotics and Automation Letters (RA-L)	null	null	null	eess.AS cs.HC cs.RO cs.SD	http://creativecommons.org/licenses/by/4.0/	Emotions are an essential element in verbal communication, so understanding individuals' affect during a human-robot interaction (HRI) becomes imperative. This paper investigates the application of vision transformer models, namely ViT (Vision Transformers) and BEiT (BERT Pre-Training of Image Transformers) pipelines, for Speech Emotion Recognition (SER) in HRI. The focus is to generalize the SER models for individual speech characteristics by fine-tuning these models on benchmark datasets and exploiting ensemble methods. For this purpose, we collected audio data from different human subjects having pseudo-naturalistic conversations with the NAO robot. We then fine-tuned our ViT and BEiT-based models and tested these models on unseen speech samples from the participants. In the results, we show that fine-tuning vision transformers on benchmark datasets and and then using either these already fine-tuned models or ensembling ViT/BEiT models gets us the highest classification accuracies per individual when it comes to identifying four primary emotions from their speech: neutral, happy, sad, and angry, as compared to fine-tuning vanilla-ViTs or BEiTs.	[ { "version": "v1", "created": "Mon, 16 Sep 2024 19:34:34 GMT" }, { "version": "v2", "created": "Fri, 22 Nov 2024 23:26:24 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 14:58:30 GMT" } ]	2025-03-18T00:00:00	[ [ "Mishra", "Ruchik", "" ], [ "Frye", "Andrew", "" ], [ "Rayguru", "Madan Mohan", "" ], [ "Popa", "Dan O.", "" ] ]	TITLE: Personalized Speech Emotion Recognition in Human-Robot Interaction using Vision Transformers ABSTRACT: Emotions are an essential element in verbal communication, so understanding individuals' affect during a human-robot interaction (HRI) becomes imperative. This paper investigates the application of vision transformer models, namely ViT (Vision Transformers) and BEiT (BERT Pre-Training of Image Transformers) pipelines, for Speech Emotion Recognition (SER) in HRI. The focus is to generalize the SER models for individual speech characteristics by fine-tuning these models on benchmark datasets and exploiting ensemble methods. For this purpose, we collected audio data from different human subjects having pseudo-naturalistic conversations with the NAO robot. We then fine-tuned our ViT and BEiT-based models and tested these models on unseen speech samples from the participants. In the results, we show that fine-tuning vision transformers on benchmark datasets and and then using either these already fine-tuned models or ensembling ViT/BEiT models gets us the highest classification accuracies per individual when it comes to identifying four primary emotions from their speech: neutral, happy, sad, and angry, as compared to fine-tuning vanilla-ViTs or BEiTs.
2409.10831	Phillip Long	Phillip Long, Zachary Novack, Taylor Berg-Kirkpatrick, Julian McAuley	PDMX: A Large-Scale Public Domain MusicXML Dataset for Symbolic Music Processing	Accepted to 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)	null	10.1109/ICASSP49660.2025.10890217	null	cs.SD cs.AI cs.LG cs.MM eess.AS	http://creativecommons.org/licenses/by/4.0/	The recent explosion of generative AI-Music systems has raised numerous concerns over data copyright, licensing music from musicians, and the conflict between open-source AI and large prestige companies. Such issues highlight the need for publicly available, copyright-free musical data, in which there is a large shortage, particularly for symbolic music data. To alleviate this issue, we present PDMX: a large-scale open-source dataset of over 250K public domain MusicXML scores collected from the score-sharing forum MuseScore, making it the largest available copyright-free symbolic music dataset to our knowledge. PDMX additionally includes a wealth of both tag and user interaction metadata, allowing us to efficiently analyze the dataset and filter for high quality user-generated scores. Given the additional metadata afforded by our data collection process, we conduct multitrack music generation experiments evaluating how different representative subsets of PDMX lead to different behaviors in downstream models, and how user-rating statistics can be used as an effective measure of data quality. Examples can be found at https://pnlong.github.io/PDMX.demo/.	[ { "version": "v1", "created": "Tue, 17 Sep 2024 01:48:42 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 03:08:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Long", "Phillip", "" ], [ "Novack", "Zachary", "" ], [ "Berg-Kirkpatrick", "Taylor", "" ], [ "McAuley", "Julian", "" ] ]	TITLE: PDMX: A Large-Scale Public Domain MusicXML Dataset for Symbolic Music Processing ABSTRACT: The recent explosion of generative AI-Music systems has raised numerous concerns over data copyright, licensing music from musicians, and the conflict between open-source AI and large prestige companies. Such issues highlight the need for publicly available, copyright-free musical data, in which there is a large shortage, particularly for symbolic music data. To alleviate this issue, we present PDMX: a large-scale open-source dataset of over 250K public domain MusicXML scores collected from the score-sharing forum MuseScore, making it the largest available copyright-free symbolic music dataset to our knowledge. PDMX additionally includes a wealth of both tag and user interaction metadata, allowing us to efficiently analyze the dataset and filter for high quality user-generated scores. Given the additional metadata afforded by our data collection process, we conduct multitrack music generation experiments evaluating how different representative subsets of PDMX lead to different behaviors in downstream models, and how user-rating statistics can be used as an effective measure of data quality. Examples can be found at https://pnlong.github.io/PDMX.demo/.
2409.13477	Chinmay Surendra Rao	Chinmay Rao, Matthias van Osch, Nicola Pezzotti, Jeroen de Bresser, Laurens Beljaards, Jakob Meineke, Elwin de Weerdt, Huangling Lu, Mariya Doneva, and Marius Staring	A Plug-and-Play Method for Guided Multi-contrast MRI Reconstruction based on Content/Style Modeling	This work has been submitted to the IEEE for possible publication	null	null	null	eess.IV cs.CV physics.med-ph	http://creativecommons.org/licenses/by/4.0/	Since multiple MRI contrasts of the same anatomy contain redundant information, one contrast can be used as a prior for guiding the reconstruction of an undersampled subsequent contrast. To this end, several learning-based guided reconstruction methods have been proposed. However, a key challenge is the requirement of large paired training datasets comprising raw data and aligned reference images. We propose a modular two-stage approach for guided reconstruction addressing this issue, which additionally provides an explanatory framework for the multi-contrast problem in terms of the shared and non-shared generative factors underlying two given contrasts. A content/style model of two-contrast image data is learned from a largely unpaired image-domain dataset and is subsequently applied as a plug-and-play operator in iterative reconstruction. The disentanglement of content and style allows explicit representation of contrast-independent and contrast-specific factors. Based on this, incorporating prior information into the reconstruction reduces to simply replacing the aliased content of the image estimate with high-quality content derived from the reference scan. Combining this component with a data consistency step and introducing a general corrective process for the content yields an iterative scheme. We name this novel approach PnP-MUNIT. Various aspects like interpretability and convergence are explored via simulations. Furthermore, its practicality is demonstrated on the NYU fastMRI DICOM dataset and two in-house multi-coil raw datasets, obtaining up to 32.6% more acceleration over learning-based non-guided reconstruction for a given SSIM. In a radiological task, PnP-MUNIT allowed 33.3% more acceleration over clinical reconstruction at diagnostic quality.	[ { "version": "v1", "created": "Fri, 20 Sep 2024 13:08:51 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 23:39:10 GMT" } ]	2025-03-18T00:00:00	[ [ "Rao", "Chinmay", "" ], [ "van Osch", "Matthias", "" ], [ "Pezzotti", "Nicola", "" ], [ "de Bresser", "Jeroen", "" ], [ "Beljaards", "Laurens", "" ], [ "Meineke", "Jakob", "" ], [ "de Weerdt", "Elwin", "" ], [ "Lu", "Huangling", "" ], [ "Doneva", "Mariya", "" ], [ "Staring", "Marius", "" ] ]	TITLE: A Plug-and-Play Method for Guided Multi-contrast MRI Reconstruction based on Content/Style Modeling ABSTRACT: Since multiple MRI contrasts of the same anatomy contain redundant information, one contrast can be used as a prior for guiding the reconstruction of an undersampled subsequent contrast. To this end, several learning-based guided reconstruction methods have been proposed. However, a key challenge is the requirement of large paired training datasets comprising raw data and aligned reference images. We propose a modular two-stage approach for guided reconstruction addressing this issue, which additionally provides an explanatory framework for the multi-contrast problem in terms of the shared and non-shared generative factors underlying two given contrasts. A content/style model of two-contrast image data is learned from a largely unpaired image-domain dataset and is subsequently applied as a plug-and-play operator in iterative reconstruction. The disentanglement of content and style allows explicit representation of contrast-independent and contrast-specific factors. Based on this, incorporating prior information into the reconstruction reduces to simply replacing the aliased content of the image estimate with high-quality content derived from the reference scan. Combining this component with a data consistency step and introducing a general corrective process for the content yields an iterative scheme. We name this novel approach PnP-MUNIT. Various aspects like interpretability and convergence are explored via simulations. Furthermore, its practicality is demonstrated on the NYU fastMRI DICOM dataset and two in-house multi-coil raw datasets, obtaining up to 32.6% more acceleration over learning-based non-guided reconstruction for a given SSIM. In a radiological task, PnP-MUNIT allowed 33.3% more acceleration over clinical reconstruction at diagnostic quality.
2409.14876	Shilong Yang	Shilong Yang, Chulong Zhang, Qi Zang, Juan Yu, Liang Zeng, Xiao Luo, Yexuan Xing, Xin Pan, Qi Li, Xiaokun Liang, Yaoqin Xie	Mammo-Clustering: A Multi-views Tri-level Information Fusion Context Clustering Framework for Localization and Classification in Mammography	10 pages, 6 figures	null	null	null	cs.CV cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Breast cancer is a significant global health issue, and the diagnosis of breast imaging has always been challenging. Mammography images typically have extremely high resolution, with lesions occupying only a very small area. Down-sampling in neural networks can easily lead to the loss of microcalcifications or subtle structures, making it difficult for traditional neural network architectures to address these issues. To tackle these challenges, we propose a Context Clustering Network with triple information fusion. Firstly, compared to CNNs or transformers, we find that Context clustering methods (1) are more computationally efficient and (2) can more easily associate structural or pathological features, making them suitable for the clinical tasks of mammography. Secondly, we propose a triple information fusion mechanism that integrates global information, feature-based local information, and patch-based local information. The proposed approach is rigorously evaluated on two public datasets, Vindr-Mammo and CBIS-DDSM, using five independent splits to ensure statistical robustness. Our method achieves an AUC of 0.828 on Vindr-Mammo and 0.805 on CBIS-DDSM, outperforming the next best method by 3.1% and 2.4%, respectively. These improvements are statistically significant (p<0.05), underscoring the benefits of Context Clustering Network with triple information fusion. Overall, our Context Clustering framework demonstrates strong potential as a scalable and cost-effective solution for large-scale mammography screening, enabling more efficient and accurate breast cancer detection. Access to our method is available at https://github.com/Sohyu1/Mammo_Clustering.	[ { "version": "v1", "created": "Mon, 23 Sep 2024 10:17:13 GMT" }, { "version": "v2", "created": "Sun, 16 Feb 2025 16:00:00 GMT" }, { "version": "v3", "created": "Sun, 2 Mar 2025 17:27:04 GMT" }, { "version": "v4", "created": "Sat, 15 Mar 2025 07:30:53 GMT" } ]	2025-03-18T00:00:00	[ [ "Yang", "Shilong", "" ], [ "Zhang", "Chulong", "" ], [ "Zang", "Qi", "" ], [ "Yu", "Juan", "" ], [ "Zeng", "Liang", "" ], [ "Luo", "Xiao", "" ], [ "Xing", "Yexuan", "" ], [ "Pan", "Xin", "" ], [ "Li", "Qi", "" ], [ "Liang", "Xiaokun", "" ], [ "Xie", "Yaoqin", "" ] ]	TITLE: Mammo-Clustering: A Multi-views Tri-level Information Fusion Context Clustering Framework for Localization and Classification in Mammography ABSTRACT: Breast cancer is a significant global health issue, and the diagnosis of breast imaging has always been challenging. Mammography images typically have extremely high resolution, with lesions occupying only a very small area. Down-sampling in neural networks can easily lead to the loss of microcalcifications or subtle structures, making it difficult for traditional neural network architectures to address these issues. To tackle these challenges, we propose a Context Clustering Network with triple information fusion. Firstly, compared to CNNs or transformers, we find that Context clustering methods (1) are more computationally efficient and (2) can more easily associate structural or pathological features, making them suitable for the clinical tasks of mammography. Secondly, we propose a triple information fusion mechanism that integrates global information, feature-based local information, and patch-based local information. The proposed approach is rigorously evaluated on two public datasets, Vindr-Mammo and CBIS-DDSM, using five independent splits to ensure statistical robustness. Our method achieves an AUC of 0.828 on Vindr-Mammo and 0.805 on CBIS-DDSM, outperforming the next best method by 3.1% and 2.4%, respectively. These improvements are statistically significant (p<0.05), underscoring the benefits of Context Clustering Network with triple information fusion. Overall, our Context Clustering framework demonstrates strong potential as a scalable and cost-effective solution for large-scale mammography screening, enabling more efficient and accurate breast cancer detection. Access to our method is available at https://github.com/Sohyu1/Mammo_Clustering.
2409.18896	Denys Iliash	Denys Iliash, Hanxiao Jiang, Yiming Zhang, Manolis Savva, Angel X. Chang	S2O: Static to Openable Enhancement for Articulated 3D Objects	null	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Despite much progress in large 3D datasets there are currently few interactive 3D object datasets, and their scale is limited due to the manual effort required in their construction. We introduce the static to openable (S2O) task which creates interactive articulated 3D objects from static counterparts through openable part detection, motion prediction, and interior geometry completion. We formulate a unified framework to tackle this task, and curate a challenging dataset of openable 3D objects that serves as a test bed for systematic evaluation. Our experiments benchmark methods from prior work, extended and improved methods, and simple yet effective heuristics for the S2O task. We find that turning static 3D objects into interactively openable counterparts is possible but that all methods struggle to generalize to realistic settings of the task, and we highlight promising future work directions. Our work enables efficient creation of interactive 3D objects for robotic manipulation and embodied AI tasks.	[ { "version": "v1", "created": "Fri, 27 Sep 2024 16:34:13 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 19:13:28 GMT" } ]	2025-03-18T00:00:00	[ [ "Iliash", "Denys", "" ], [ "Jiang", "Hanxiao", "" ], [ "Zhang", "Yiming", "" ], [ "Savva", "Manolis", "" ], [ "Chang", "Angel X.", "" ] ]	TITLE: S2O: Static to Openable Enhancement for Articulated 3D Objects ABSTRACT: Despite much progress in large 3D datasets there are currently few interactive 3D object datasets, and their scale is limited due to the manual effort required in their construction. We introduce the static to openable (S2O) task which creates interactive articulated 3D objects from static counterparts through openable part detection, motion prediction, and interior geometry completion. We formulate a unified framework to tackle this task, and curate a challenging dataset of openable 3D objects that serves as a test bed for systematic evaluation. Our experiments benchmark methods from prior work, extended and improved methods, and simple yet effective heuristics for the S2O task. We find that turning static 3D objects into interactively openable counterparts is possible but that all methods struggle to generalize to realistic settings of the task, and we highlight promising future work directions. Our work enables efficient creation of interactive 3D objects for robotic manipulation and embodied AI tasks.
2409.19583	Jun Liu	Jun Liu, Geng Yuan, Weihao Zeng, Hao Tang, Wenbin Zhang, Xue Lin, XiaoLin Xu, Dong Huang, and Yanzhi Wang	Brain Tumor Classification on MRI in Light of Molecular Markers	ICAI'22 - The 24th International Conference on Artificial Intelligence, The 2022 World Congress in Computer Science, Computer Engineering, & Applied Computing (CSCE'22), Las Vegas, USA. The paper acceptance rate 17% for regular papers. The publication of the CSCE 2022 conference proceedings has been delayed due to the pandemic	Springer Nature - Book Series: Transactions on Computational Science & Computational Intelligence, 2022	null	null	eess.IV cs.CV cs.LG q-bio.QM	http://creativecommons.org/licenses/by/4.0/	In research findings, co-deletion of the 1p/19q gene is associated with clinical outcomes in low-grade gliomas. The ability to predict 1p19q status is critical for treatment planning and patient follow-up. This study aims to utilize a specially MRI-based convolutional neural network for brain cancer detection. Although public networks such as RestNet and AlexNet can effectively diagnose brain cancers using transfer learning, the model includes quite a few weights that have nothing to do with medical images. As a result, the diagnostic results are unreliable by the transfer learning model. To deal with the problem of trustworthiness, we create the model from the ground up, rather than depending on a pre-trained model. To enable flexibility, we combined convolution stacking with a dropout and full connect operation, it improved performance by reducing overfitting. During model training, we also supplement the given dataset and inject Gaussian noise. We use three--fold cross-validation to train the best selection model. Comparing InceptionV3, VGG16, and MobileNetV2 fine-tuned with pre-trained models, our model produces better results. On an validation set of 125 codeletion vs. 31 not codeletion images, the proposed network achieves 96.37\% percent F1-score, 97.46\% percent precision, and 96.34\% percent recall when classifying 1p/19q codeletion and not codeletion images.	[ { "version": "v1", "created": "Sun, 29 Sep 2024 07:04:26 GMT" }, { "version": "v2", "created": "Mon, 10 Mar 2025 17:01:47 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 18:50:23 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Jun", "" ], [ "Yuan", "Geng", "" ], [ "Zeng", "Weihao", "" ], [ "Tang", "Hao", "" ], [ "Zhang", "Wenbin", "" ], [ "Lin", "Xue", "" ], [ "Xu", "XiaoLin", "" ], [ "Huang", "Dong", "" ], [ "Wang", "Yanzhi", "" ] ]	TITLE: Brain Tumor Classification on MRI in Light of Molecular Markers ABSTRACT: In research findings, co-deletion of the 1p/19q gene is associated with clinical outcomes in low-grade gliomas. The ability to predict 1p19q status is critical for treatment planning and patient follow-up. This study aims to utilize a specially MRI-based convolutional neural network for brain cancer detection. Although public networks such as RestNet and AlexNet can effectively diagnose brain cancers using transfer learning, the model includes quite a few weights that have nothing to do with medical images. As a result, the diagnostic results are unreliable by the transfer learning model. To deal with the problem of trustworthiness, we create the model from the ground up, rather than depending on a pre-trained model. To enable flexibility, we combined convolution stacking with a dropout and full connect operation, it improved performance by reducing overfitting. During model training, we also supplement the given dataset and inject Gaussian noise. We use three--fold cross-validation to train the best selection model. Comparing InceptionV3, VGG16, and MobileNetV2 fine-tuned with pre-trained models, our model produces better results. On an validation set of 125 codeletion vs. 31 not codeletion images, the proposed network achieves 96.37\% percent F1-score, 97.46\% percent precision, and 96.34\% percent recall when classifying 1p/19q codeletion and not codeletion images.
2409.19917	Hongjie Fang	Jingjing Chen, Hongjie Fang, Hao-Shu Fang and Cewu Lu	Towards Effective Utilization of Mixed-Quality Demonstrations in Robotic Manipulation via Segment-Level Selection and Optimization	ICRA 2025. Project website: https://tonyfang.net/s2i/	null	null	null	cs.RO	http://creativecommons.org/licenses/by-nc-sa/4.0/	Data is crucial for robotic manipulation, as it underpins the development of robotic systems for complex tasks. While high-quality, diverse datasets enhance the performance and adaptability of robotic manipulation policies, collecting extensive expert-level data is resource-intensive. Consequently, many current datasets suffer from quality inconsistencies due to operator variability, highlighting the need for methods to utilize mixed-quality data effectively. To mitigate these issues, we propose "Select Segments to Imitate" (S2I), a framework that selects and optimizes mixed-quality demonstration data at the segment level, while ensuring plug-and-play compatibility with existing robotic manipulation policies. The framework has three components: demonstration segmentation dividing origin data into meaningful segments, segment selection using contrastive learning to find high-quality segments, and trajectory optimization to refine suboptimal segments for better policy learning. We evaluate S2I through comprehensive experiments in simulation and real-world environments across six tasks, demonstrating that with only 3 expert demonstrations for reference, S2I can improve the performance of various downstream policies when trained with mixed-quality demonstrations. Project website: https://tonyfang.net/s2i/.	[ { "version": "v1", "created": "Mon, 30 Sep 2024 03:42:06 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 09:58:58 GMT" } ]	2025-03-18T00:00:00	[ [ "Chen", "Jingjing", "" ], [ "Fang", "Hongjie", "" ], [ "Fang", "Hao-Shu", "" ], [ "Lu", "Cewu", "" ] ]	TITLE: Towards Effective Utilization of Mixed-Quality Demonstrations in Robotic Manipulation via Segment-Level Selection and Optimization ABSTRACT: Data is crucial for robotic manipulation, as it underpins the development of robotic systems for complex tasks. While high-quality, diverse datasets enhance the performance and adaptability of robotic manipulation policies, collecting extensive expert-level data is resource-intensive. Consequently, many current datasets suffer from quality inconsistencies due to operator variability, highlighting the need for methods to utilize mixed-quality data effectively. To mitigate these issues, we propose "Select Segments to Imitate" (S2I), a framework that selects and optimizes mixed-quality demonstration data at the segment level, while ensuring plug-and-play compatibility with existing robotic manipulation policies. The framework has three components: demonstration segmentation dividing origin data into meaningful segments, segment selection using contrastive learning to find high-quality segments, and trajectory optimization to refine suboptimal segments for better policy learning. We evaluate S2I through comprehensive experiments in simulation and real-world environments across six tasks, demonstrating that with only 3 expert demonstrations for reference, S2I can improve the performance of various downstream policies when trained with mixed-quality demonstrations. Project website: https://tonyfang.net/s2i/.
2410.00871	Yunze Liu	Yunze Liu, Li Yi	MAP: Unleashing Hybrid Mamba-Transformer Vision Backbone's Potential with Masked Autoregressive Pretraining	null	Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025	null	null	cs.CV cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Hybrid Mamba-Transformer networks have recently garnered broad attention. These networks can leverage the scalability of Transformers while capitalizing on Mamba's strengths in long-context modeling and computational efficiency. However, the challenge of effectively pretraining such hybrid networks remains an open question. Existing methods, such as Masked Autoencoders (MAE) or autoregressive (AR) pretraining, primarily focus on single-type network architectures. In contrast, pretraining strategies for hybrid architectures must be effective for both Mamba and Transformer components. Based on this, we propose Masked Autoregressive Pretraining (MAP) to pretrain a hybrid Mamba-Transformer vision backbone network. This strategy combines the strengths of both MAE and Autoregressive pretraining, improving the performance of Mamba and Transformer modules within a unified paradigm. Experimental results show that the hybrid Mamba-Transformer vision backbone network pretrained with MAP significantly outperforms other pretraining strategies, achieving state-of-the-art performance. We validate the method's effectiveness on both 2D and 3D datasets and provide detailed ablation studies to support the design choices for each component. The code and checkpoints are available at https://github.com/yunzeliu/MAP	[ { "version": "v1", "created": "Tue, 1 Oct 2024 17:05:08 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 15:21:48 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Yunze", "" ], [ "Yi", "Li", "" ] ]	TITLE: MAP: Unleashing Hybrid Mamba-Transformer Vision Backbone's Potential with Masked Autoregressive Pretraining ABSTRACT: Hybrid Mamba-Transformer networks have recently garnered broad attention. These networks can leverage the scalability of Transformers while capitalizing on Mamba's strengths in long-context modeling and computational efficiency. However, the challenge of effectively pretraining such hybrid networks remains an open question. Existing methods, such as Masked Autoencoders (MAE) or autoregressive (AR) pretraining, primarily focus on single-type network architectures. In contrast, pretraining strategies for hybrid architectures must be effective for both Mamba and Transformer components. Based on this, we propose Masked Autoregressive Pretraining (MAP) to pretrain a hybrid Mamba-Transformer vision backbone network. This strategy combines the strengths of both MAE and Autoregressive pretraining, improving the performance of Mamba and Transformer modules within a unified paradigm. Experimental results show that the hybrid Mamba-Transformer vision backbone network pretrained with MAP significantly outperforms other pretraining strategies, achieving state-of-the-art performance. We validate the method's effectiveness on both 2D and 3D datasets and provide detailed ablation studies to support the design choices for each component. The code and checkpoints are available at https://github.com/yunzeliu/MAP
2410.02683	Yu Ying Chiu	Yu Ying Chiu, Liwei Jiang and Yejin Choi	DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Life	Accepted into ICLR 2025 (spotlight)	null	null	null	cs.CL cs.AI cs.LG	http://creativecommons.org/licenses/by/4.0/	As users increasingly seek guidance from LLMs for decision-making in daily life, many of these decisions are not clear-cut and depend significantly on the personal values and ethical standards of people. We present DailyDilemmas, a dataset of 1,360 moral dilemmas encountered in everyday life. Each dilemma presents two possible actions, along with affected parties and relevant human values for each action. Based on these dilemmas, we gather a repository of human values covering diverse everyday topics, such as interpersonal relationships, workplace, and environmental issues. With DailyDilemmas, we evaluate LLMs on these dilemmas to determine what action they will choose and the values represented by these action choices. Then, we analyze values through the lens of five theoretical frameworks inspired by sociology, psychology, and philosophy, including the World Values Survey, Moral Foundations Theory, Maslow's Hierarchy of Needs, Aristotle's Virtues, and Plutchik's Wheel of Emotions. For instance, we find LLMs are most aligned with self-expression over survival in World Values Survey and care over loyalty in Moral Foundations Theory. Interestingly, we find substantial preference differences in models for some core values. For example, for truthfulness, Mixtral-8x7B neglects it by 9.7% while GPT-4-turbo selects it by 9.4%. We also study the recent guidance released by OpenAI (ModelSpec), and Anthropic (Constitutional AI) to understand how their designated principles reflect their models' actual value prioritization when facing nuanced moral reasoning in daily-life settings. Finally, we find that end users cannot effectively steer such prioritization using system prompts.	[ { "version": "v1", "created": "Thu, 3 Oct 2024 17:08:52 GMT" }, { "version": "v2", "created": "Mon, 3 Mar 2025 07:20:54 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 03:54:40 GMT" } ]	2025-03-18T00:00:00	[ [ "Chiu", "Yu Ying", "" ], [ "Jiang", "Liwei", "" ], [ "Choi", "Yejin", "" ] ]	TITLE: DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Life ABSTRACT: As users increasingly seek guidance from LLMs for decision-making in daily life, many of these decisions are not clear-cut and depend significantly on the personal values and ethical standards of people. We present DailyDilemmas, a dataset of 1,360 moral dilemmas encountered in everyday life. Each dilemma presents two possible actions, along with affected parties and relevant human values for each action. Based on these dilemmas, we gather a repository of human values covering diverse everyday topics, such as interpersonal relationships, workplace, and environmental issues. With DailyDilemmas, we evaluate LLMs on these dilemmas to determine what action they will choose and the values represented by these action choices. Then, we analyze values through the lens of five theoretical frameworks inspired by sociology, psychology, and philosophy, including the World Values Survey, Moral Foundations Theory, Maslow's Hierarchy of Needs, Aristotle's Virtues, and Plutchik's Wheel of Emotions. For instance, we find LLMs are most aligned with self-expression over survival in World Values Survey and care over loyalty in Moral Foundations Theory. Interestingly, we find substantial preference differences in models for some core values. For example, for truthfulness, Mixtral-8x7B neglects it by 9.7% while GPT-4-turbo selects it by 9.4%. We also study the recent guidance released by OpenAI (ModelSpec), and Anthropic (Constitutional AI) to understand how their designated principles reflect their models' actual value prioritization when facing nuanced moral reasoning in daily-life settings. Finally, we find that end users cannot effectively steer such prioritization using system prompts.
2410.05270	Mohammad Fahes	Mohammad Fahes, Tuan-Hung Vu, Andrei Bursuc, Patrick P\'erez, Raoul de Charette	CLIP's Visual Embedding Projector is a Few-shot Cornucopia	null	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We consider the problem of adapting a contrastively pretrained vision-language model like CLIP (Radford et al., 2021) for few-shot classification. The literature addresses this problem by learning a linear classifier of the frozen visual features, optimizing word embeddings, or learning external feature adapters. We introduce an alternative way for few-shot CLIP adaptation without adding ''external'' parameters to optimize. We find that simply fine-tuning the embedding projection matrix of the vision encoder leads to better performance than all baselines. Furthermore, we show that regularizing training with the distance between the fine-tuned and pretrained matrices adds reliability for adapting CLIP, making the results stable across different learning rates in the ''validation-free'' setting. This simple approach, coined ProLIP, yields state-of-the-art performance on 11 few-shot classification benchmarks, few-shot cross-dataset transfer, domain generalization, and base-to-new class generalization. We also show that ProLIP significantly outperforms prompt tuning when extended to another task of test-time adaptation, while being one order of magnitude faster to train. Code will be made available at: https://github.com/astra-vision/ProLIP .	[ { "version": "v1", "created": "Mon, 7 Oct 2024 17:59:59 GMT" }, { "version": "v2", "created": "Fri, 6 Dec 2024 16:07:47 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 17:52:55 GMT" } ]	2025-03-18T00:00:00	[ [ "Fahes", "Mohammad", "" ], [ "Vu", "Tuan-Hung", "" ], [ "Bursuc", "Andrei", "" ], [ "Pérez", "Patrick", "" ], [ "de Charette", "Raoul", "" ] ]	TITLE: CLIP's Visual Embedding Projector is a Few-shot Cornucopia ABSTRACT: We consider the problem of adapting a contrastively pretrained vision-language model like CLIP (Radford et al., 2021) for few-shot classification. The literature addresses this problem by learning a linear classifier of the frozen visual features, optimizing word embeddings, or learning external feature adapters. We introduce an alternative way for few-shot CLIP adaptation without adding ''external'' parameters to optimize. We find that simply fine-tuning the embedding projection matrix of the vision encoder leads to better performance than all baselines. Furthermore, we show that regularizing training with the distance between the fine-tuned and pretrained matrices adds reliability for adapting CLIP, making the results stable across different learning rates in the ''validation-free'' setting. This simple approach, coined ProLIP, yields state-of-the-art performance on 11 few-shot classification benchmarks, few-shot cross-dataset transfer, domain generalization, and base-to-new class generalization. We also show that ProLIP significantly outperforms prompt tuning when extended to another task of test-time adaptation, while being one order of magnitude faster to train. Code will be made available at: https://github.com/astra-vision/ProLIP .
2410.05894	Yichen Song	Yichen Song, Jiaming Wang, Yunbo Wang, Xiaokang Yang	DimOL: Dimensional Awareness as A New 'Dimension' in Operator Learning	null	null	null	null	cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In the realm of computational physics, an enduring topic is the numerical solutions to partial differential equations (PDEs). Recently, the attention of researchers has shifted towards Neural Operator methods, renowned for their capability to approximate ``operators'' -- mappings from functions to functions. Despite the universal approximation theorem within neural operators, ensuring error bounds often requires employing numerous Fourier layers. However, what about lightweight models? In response to this question, we introduce DimOL (Dimension-aware Operator Learning), drawing insights from dimensional analysis. To implement DimOL, we propose the ProdLayer, which can be seamlessly integrated into FNO-based and Transformer-based PDE solvers, enhancing their ability to handle sum-of-products structures inherent in many physical systems. Empirically, DimOL models achieve up to 48% performance gain within the PDE datasets. Furthermore, by analyzing Fourier components' weights, we can symbolically discern the physical significance of each term. This sheds light on the opaque nature of neural networks, unveiling underlying physical principles.	[ { "version": "v1", "created": "Tue, 8 Oct 2024 10:48:50 GMT" }, { "version": "v2", "created": "Fri, 14 Feb 2025 08:27:05 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 06:54:47 GMT" } ]	2025-03-18T00:00:00	[ [ "Song", "Yichen", "" ], [ "Wang", "Jiaming", "" ], [ "Wang", "Yunbo", "" ], [ "Yang", "Xiaokang", "" ] ]	TITLE: DimOL: Dimensional Awareness as A New 'Dimension' in Operator Learning ABSTRACT: In the realm of computational physics, an enduring topic is the numerical solutions to partial differential equations (PDEs). Recently, the attention of researchers has shifted towards Neural Operator methods, renowned for their capability to approximate ``operators'' -- mappings from functions to functions. Despite the universal approximation theorem within neural operators, ensuring error bounds often requires employing numerous Fourier layers. However, what about lightweight models? In response to this question, we introduce DimOL (Dimension-aware Operator Learning), drawing insights from dimensional analysis. To implement DimOL, we propose the ProdLayer, which can be seamlessly integrated into FNO-based and Transformer-based PDE solvers, enhancing their ability to handle sum-of-products structures inherent in many physical systems. Empirically, DimOL models achieve up to 48% performance gain within the PDE datasets. Furthermore, by analyzing Fourier components' weights, we can symbolically discern the physical significance of each term. This sheds light on the opaque nature of neural networks, unveiling underlying physical principles.
2410.06418	Hossein Resani	Hossein Resani, Behrooz Nasihatkon	MIRACLE3D: Memory-efficient Integrated Robust Approach for Continual Learning on Point Clouds via Shape Model Construction	Accepted to ICLR 2025, Singapore	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	In this paper, we introduce a novel framework for memory-efficient and privacy-preserving continual learning in 3D object classification. Unlike conventional memory-based approaches in continual learning that require storing numerous exemplars, our method constructs a compact shape model for each class, retaining only the mean shape along with a few key modes of variation. This strategy not only enables the generation of diverse training samples while drastically reducing memory usage but also enhances privacy by eliminating the need to store original data. To further improve model robustness against input variations, an issue common in 3D domains due to the absence of strong backbones and limited training data, we incorporate Gradient Mode Regularization. This technique enhances model stability and broadens classification margins, resulting in accuracy improvements. We validate our approach through extensive experiments on the ModelNet40, ShapeNet, and ScanNet datasets, where we achieve state-of-the-art performance. Notably, our method consumes only 15% of the memory required by competing methods on the ModelNet40 and ShapeNet, while achieving comparable performance on the challenging ScanNet dataset with just 8.5% of the memory. These results underscore the scalability, effectiveness, and privacy-preserving strengths of our framework for 3D object classification.	[ { "version": "v1", "created": "Tue, 8 Oct 2024 23:12:33 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 01:55:58 GMT" } ]	2025-03-18T00:00:00	[ [ "Resani", "Hossein", "" ], [ "Nasihatkon", "Behrooz", "" ] ]	TITLE: MIRACLE3D: Memory-efficient Integrated Robust Approach for Continual Learning on Point Clouds via Shape Model Construction ABSTRACT: In this paper, we introduce a novel framework for memory-efficient and privacy-preserving continual learning in 3D object classification. Unlike conventional memory-based approaches in continual learning that require storing numerous exemplars, our method constructs a compact shape model for each class, retaining only the mean shape along with a few key modes of variation. This strategy not only enables the generation of diverse training samples while drastically reducing memory usage but also enhances privacy by eliminating the need to store original data. To further improve model robustness against input variations, an issue common in 3D domains due to the absence of strong backbones and limited training data, we incorporate Gradient Mode Regularization. This technique enhances model stability and broadens classification margins, resulting in accuracy improvements. We validate our approach through extensive experiments on the ModelNet40, ShapeNet, and ScanNet datasets, where we achieve state-of-the-art performance. Notably, our method consumes only 15% of the memory required by competing methods on the ModelNet40 and ShapeNet, while achieving comparable performance on the challenging ScanNet dataset with just 8.5% of the memory. These results underscore the scalability, effectiveness, and privacy-preserving strengths of our framework for 3D object classification.
2410.06757	Peng Zhang	Peng Zhang, Qianqian Xue, Xingyu Liu, Guanglei Zhang, Wenjian Wang, Jiye Liang	MDiff-FMT: Morphology-aware Diffusion Model for Fluorescence Molecular Tomography with Small-scale Datasets	null	null	null	null	eess.IV cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Fluorescence molecular tomography (FMT) is a sensitive optical imaging technology widely used in biomedical research. However, the ill-posedness of the inverse problem poses a huge challenge to FMT reconstruction. Although end-to-end deep learning algorithms have been widely used to address this critical issue, they still suffer from high data dependency and poor morphological restoration. In this paper, we report for the first time a morphology-aware diffusion model, MDiff-FMT, based on denoising diffusion probabilistic model (DDPM) to achieve high-fidelity morphological reconstruction for FMT. First, we use the noise addition of DDPM to simulate the process of the gradual degradation of morphological features, and achieve fine-grained reconstruction of morphological features through a stepwise probabilistic sampling mechanism, avoiding problems such as loss of structure details that may occur in end-to-end deep learning methods. Additionally, we introduce the conditional fluorescence image as structural prior information to sample a high-fidelity reconstructed image from the noisy images. Numerous numerical and real phantom experimental results show that the proposed MDiff-FMT achieves SOTA results in morphological reconstruction of FMT without relying on large-scale datasets.	[ { "version": "v1", "created": "Wed, 9 Oct 2024 10:41:31 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 04:47:18 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Peng", "" ], [ "Xue", "Qianqian", "" ], [ "Liu", "Xingyu", "" ], [ "Zhang", "Guanglei", "" ], [ "Wang", "Wenjian", "" ], [ "Liang", "Jiye", "" ] ]	TITLE: MDiff-FMT: Morphology-aware Diffusion Model for Fluorescence Molecular Tomography with Small-scale Datasets ABSTRACT: Fluorescence molecular tomography (FMT) is a sensitive optical imaging technology widely used in biomedical research. However, the ill-posedness of the inverse problem poses a huge challenge to FMT reconstruction. Although end-to-end deep learning algorithms have been widely used to address this critical issue, they still suffer from high data dependency and poor morphological restoration. In this paper, we report for the first time a morphology-aware diffusion model, MDiff-FMT, based on denoising diffusion probabilistic model (DDPM) to achieve high-fidelity morphological reconstruction for FMT. First, we use the noise addition of DDPM to simulate the process of the gradual degradation of morphological features, and achieve fine-grained reconstruction of morphological features through a stepwise probabilistic sampling mechanism, avoiding problems such as loss of structure details that may occur in end-to-end deep learning methods. Additionally, we introduce the conditional fluorescence image as structural prior information to sample a high-fidelity reconstructed image from the noisy images. Numerous numerical and real phantom experimental results show that the proposed MDiff-FMT achieves SOTA results in morphological reconstruction of FMT without relying on large-scale datasets.
2410.08392	Howon Lee	Howon Lee, Aanchal Save, Pranay Seshadri, Juergen Rauleder	Large Airfoil Models	null	null	null	null	physics.flu-dyn	http://creativecommons.org/licenses/by/4.0/	The development of a Large Airfoil Model (LAM), a transformative approach for answering technical questions on airfoil aerodynamics, requires a vast dataset and a model to leverage it. To build this foundation, a novel probabilistic machine learning approach, A Deep Airfoil Prediction Tool (ADAPT), has been developed. ADAPT makes uncertainty-aware predictions of airfoil pressure coefficient ($C_p$) distributions by harnessing experimental data and incorporating measurement uncertainties. By employing deep kernel learning, performing Gaussian Process Regression in a ten-dimensional latent space learned by a neural network, ADAPT effectively handles unstructured experimental datasets. In tandem, Airfoil Surface Pressure Information Repository of Experiments (ASPIRE), the first large-scale, open-source repository of airfoil experimental data has been developed. ASPIRE integrates century-old historical data with modern reports, forming an unparalleled resource of real-world pressure measurements. This addresses a critical gap left by prior repositories, which relied primarily on numerical simulations. Demonstrative results for three airfoils show that ADAPT accurately predicts $C_p$ distributions and aerodynamic coefficients across varied flow conditions, achieving a mean absolute error in enclosed area ($\text{MAE}_\text{enclosed}$) of 0.029. ASPIRE and ADAPT lay the foundation for an interactive airfoil analysis tool driven by a large language model, enabling users to perform design tasks based on natural language questions rather than explicit technical input.	[ { "version": "v1", "created": "Thu, 10 Oct 2024 21:59:29 GMT" }, { "version": "v2", "created": "Thu, 7 Nov 2024 16:18:31 GMT" }, { "version": "v3", "created": "Thu, 12 Dec 2024 00:31:52 GMT" }, { "version": "v4", "created": "Sun, 16 Mar 2025 19:37:44 GMT" } ]	2025-03-18T00:00:00	[ [ "Lee", "Howon", "" ], [ "Save", "Aanchal", "" ], [ "Seshadri", "Pranay", "" ], [ "Rauleder", "Juergen", "" ] ]	TITLE: Large Airfoil Models ABSTRACT: The development of a Large Airfoil Model (LAM), a transformative approach for answering technical questions on airfoil aerodynamics, requires a vast dataset and a model to leverage it. To build this foundation, a novel probabilistic machine learning approach, A Deep Airfoil Prediction Tool (ADAPT), has been developed. ADAPT makes uncertainty-aware predictions of airfoil pressure coefficient ($C_p$) distributions by harnessing experimental data and incorporating measurement uncertainties. By employing deep kernel learning, performing Gaussian Process Regression in a ten-dimensional latent space learned by a neural network, ADAPT effectively handles unstructured experimental datasets. In tandem, Airfoil Surface Pressure Information Repository of Experiments (ASPIRE), the first large-scale, open-source repository of airfoil experimental data has been developed. ASPIRE integrates century-old historical data with modern reports, forming an unparalleled resource of real-world pressure measurements. This addresses a critical gap left by prior repositories, which relied primarily on numerical simulations. Demonstrative results for three airfoils show that ADAPT accurately predicts $C_p$ distributions and aerodynamic coefficients across varied flow conditions, achieving a mean absolute error in enclosed area ($\text{MAE}_\text{enclosed}$) of 0.029. ASPIRE and ADAPT lay the foundation for an interactive airfoil analysis tool driven by a large language model, enabling users to perform design tasks based on natural language questions rather than explicit technical input.
2410.09374	Junkai Niu	Junkai Niu, Sheng Zhong, Xiuyuan Lu, Shaojie Shen, Guillermo Gallego, Yi Zhou	ESVO2: Direct Visual-Inertial Odometry with Stereo Event Cameras	null	IEEE Transactions on Robotics, 2025	10.1109/TRO.2025.3548523	null	cs.CV cs.RO	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Event-based visual odometry is a specific branch of visual Simultaneous Localization and Mapping (SLAM) techniques, which aims at solving tracking and mapping subproblems (typically in parallel), by exploiting the special working principles of neuromorphic (i.e., event-based) cameras. Due to the motion-dependent nature of event data, explicit data association (i.e., feature matching) under large-baseline view-point changes is difficult to establish, making direct methods a more rational choice. However, state-of-the-art direct methods are limited by the high computational complexity of the mapping sub-problem and the degeneracy of camera pose tracking in certain degrees of freedom (DoF) in rotation. In this paper, we tackle these issues by building an event-based stereo visual-inertial odometry system on top of a direct pipeline. Specifically, to speed up the mapping operation, we propose an efficient strategy for sampling contour points according to the local dynamics of events. The mapping performance is also improved in terms of structure completeness and local smoothness by merging the temporal stereo and static stereo results. To circumvent the degeneracy of camera pose tracking in recovering the pitch and yaw components of general 6-DoF motion, we introduce IMU measurements as motion priors via pre-integration. To this end, a compact back-end is proposed for continuously updating the IMU bias and predicting the linear velocity, enabling an accurate motion prediction for camera pose tracking. The resulting system scales well with modern high-resolution event cameras and leads to better global positioning accuracy in large-scale outdoor environments. Extensive evaluations on five publicly available datasets featuring different resolutions and scenarios justify the superior performance of the proposed system against five state-of-the-art methods.	[ { "version": "v1", "created": "Sat, 12 Oct 2024 05:35:27 GMT" }, { "version": "v2", "created": "Fri, 17 Jan 2025 15:52:06 GMT" }, { "version": "v3", "created": "Mon, 3 Mar 2025 05:31:05 GMT" } ]	2025-03-18T00:00:00	[ [ "Niu", "Junkai", "" ], [ "Zhong", "Sheng", "" ], [ "Lu", "Xiuyuan", "" ], [ "Shen", "Shaojie", "" ], [ "Gallego", "Guillermo", "" ], [ "Zhou", "Yi", "" ] ]	TITLE: ESVO2: Direct Visual-Inertial Odometry with Stereo Event Cameras ABSTRACT: Event-based visual odometry is a specific branch of visual Simultaneous Localization and Mapping (SLAM) techniques, which aims at solving tracking and mapping subproblems (typically in parallel), by exploiting the special working principles of neuromorphic (i.e., event-based) cameras. Due to the motion-dependent nature of event data, explicit data association (i.e., feature matching) under large-baseline view-point changes is difficult to establish, making direct methods a more rational choice. However, state-of-the-art direct methods are limited by the high computational complexity of the mapping sub-problem and the degeneracy of camera pose tracking in certain degrees of freedom (DoF) in rotation. In this paper, we tackle these issues by building an event-based stereo visual-inertial odometry system on top of a direct pipeline. Specifically, to speed up the mapping operation, we propose an efficient strategy for sampling contour points according to the local dynamics of events. The mapping performance is also improved in terms of structure completeness and local smoothness by merging the temporal stereo and static stereo results. To circumvent the degeneracy of camera pose tracking in recovering the pitch and yaw components of general 6-DoF motion, we introduce IMU measurements as motion priors via pre-integration. To this end, a compact back-end is proposed for continuously updating the IMU bias and predicting the linear velocity, enabling an accurate motion prediction for camera pose tracking. The resulting system scales well with modern high-resolution event cameras and leads to better global positioning accuracy in large-scale outdoor environments. Extensive evaluations on five publicly available datasets featuring different resolutions and scenarios justify the superior performance of the proposed system against five state-of-the-art methods.
2410.10624	Zechen Li	Zechen Li, Shohreh Deldari, Linyao Chen, Hao Xue and Flora D. Salim	SensorLLM: Aligning Large Language Models with Motion Sensors for Human Activity Recognition	null	null	null	null	cs.CL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We introduce SensorLLM, a two-stage framework that enables Large Language Models (LLMs) to perform human activity recognition (HAR) from sensor data. Despite their strong reasoning and generalization capabilities, LLMs remain underutilized for motion sensor data due to the lack of semantic context in time-series, computational constraints, and challenges in processing numerical inputs. SensorLLM addresses these limitations through a Sensor-Language Alignment stage, where we introduce special tokens for each sensor channel and automatically generate textual trend descriptions. This alignment enables LLMs to capture numerical variations, channel-specific features, and data of varying duration--without requiring human annotations. In the subsequent Task-Aware Tuning stage, we refine the model for HAR classification, achieving performance that matches or surpasses state-of-the-art methods. Our results demonstrate that SensorLLM evolves into an effective sensor learner, reasoner, and classifier through Sensor-Language Alignment, generalizing across diverse HAR datasets. We believe this work establishes a foundation for future research on time-series and text alignment, paving the way for foundation models in sensor data analysis.	[ { "version": "v1", "created": "Mon, 14 Oct 2024 15:30:41 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 09:28:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Zechen", "" ], [ "Deldari", "Shohreh", "" ], [ "Chen", "Linyao", "" ], [ "Xue", "Hao", "" ], [ "Salim", "Flora D.", "" ] ]	TITLE: SensorLLM: Aligning Large Language Models with Motion Sensors for Human Activity Recognition ABSTRACT: We introduce SensorLLM, a two-stage framework that enables Large Language Models (LLMs) to perform human activity recognition (HAR) from sensor data. Despite their strong reasoning and generalization capabilities, LLMs remain underutilized for motion sensor data due to the lack of semantic context in time-series, computational constraints, and challenges in processing numerical inputs. SensorLLM addresses these limitations through a Sensor-Language Alignment stage, where we introduce special tokens for each sensor channel and automatically generate textual trend descriptions. This alignment enables LLMs to capture numerical variations, channel-specific features, and data of varying duration--without requiring human annotations. In the subsequent Task-Aware Tuning stage, we refine the model for HAR classification, achieving performance that matches or surpasses state-of-the-art methods. Our results demonstrate that SensorLLM evolves into an effective sensor learner, reasoner, and classifier through Sensor-Language Alignment, generalizing across diverse HAR datasets. We believe this work establishes a foundation for future research on time-series and text alignment, paving the way for foundation models in sensor data analysis.
2410.10880	Hengxiang Zhang	Hengxiang Zhang, Songxin Zhang, Bingyi Jing, Hongxin Wei	Fine-tuning can Help Detect Pretraining Data from Large Language Models	null	null	null	null	cs.CL cs.AI cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In the era of large language models (LLMs), detecting pretraining data has been increasingly important due to concerns about fair evaluation and ethical risks. Current methods differentiate members and non-members by designing scoring functions, like Perplexity and Min-k%. However, the diversity and complexity of training data magnifies the difficulty of distinguishing, leading to suboptimal performance in detecting pretraining data. In this paper, we first explore the benefits of unseen data, which can be easily collected after the release of the LLM. We find that the perplexities of LLMs shift differently for members and non-members, after fine-tuning with a small amount of previously unseen data. In light of this, we introduce a novel and effective method termed Fine-tuned Score Deviation(FSD), which improves the performance of current scoring functions for pretraining data detection. In particular, we propose to measure the deviation distance of current scores after fine-tuning on a small amount of unseen data within the same domain. In effect, using a few unseen data can largely decrease the scores of all non-members, leading to a larger deviation distance than members. Extensive experiments demonstrate the effectiveness of our method, significantly improving the AUC score on common benchmark datasets across various models.	[ { "version": "v1", "created": "Wed, 9 Oct 2024 15:36:42 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 12:29:05 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Hengxiang", "" ], [ "Zhang", "Songxin", "" ], [ "Jing", "Bingyi", "" ], [ "Wei", "Hongxin", "" ] ]	TITLE: Fine-tuning can Help Detect Pretraining Data from Large Language Models ABSTRACT: In the era of large language models (LLMs), detecting pretraining data has been increasingly important due to concerns about fair evaluation and ethical risks. Current methods differentiate members and non-members by designing scoring functions, like Perplexity and Min-k%. However, the diversity and complexity of training data magnifies the difficulty of distinguishing, leading to suboptimal performance in detecting pretraining data. In this paper, we first explore the benefits of unseen data, which can be easily collected after the release of the LLM. We find that the perplexities of LLMs shift differently for members and non-members, after fine-tuning with a small amount of previously unseen data. In light of this, we introduce a novel and effective method termed Fine-tuned Score Deviation(FSD), which improves the performance of current scoring functions for pretraining data detection. In particular, we propose to measure the deviation distance of current scores after fine-tuning on a small amount of unseen data within the same domain. In effect, using a few unseen data can largely decrease the scores of all non-members, leading to a larger deviation distance than members. Extensive experiments demonstrate the effectiveness of our method, significantly improving the AUC score on common benchmark datasets across various models.
2410.11506	Hongyu An	Hongyu An, Xinfeng Zhang, Shijie Zhao, Li Zhang, Ruiqin Xiong	Spatio-Temporal Distortion Aware Omnidirectional Video Super-Resolution	null	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Omnidirectional video (ODV) provides an immersive visual experience and is widely utilized in virtual reality and augmented reality. However, restricted capturing devices and transmission bandwidth lead to low-resolution ODVs. Video super-resolution (SR) is proposed to enhance resolution, but practical ODV spatial projection distortions and temporal flickering are not well addressed directly applying existing methods. To achieve better ODV-SR reconstruction, we propose a Spatio-Temporal Distortion Aware Network (STDAN) oriented to ODV characteristics. Specifically, a spatially continuous distortion modulation module is introduced to improve discrete projection distortions. Next, we design an interlaced multi-frame reconstruction mechanism to refine temporal consistency across frames. Furthermore, we incorporate latitude-saliency adaptive weights during training to concentrate on regions with higher texture complexity and human-watching interest. In general, we explore inference-free and real-world viewing matched strategies to provide an application-friendly method on a novel ODV-SR dataset with practical scenarios. Extensive experimental results demonstrate the superior performance of the proposed STDAN over state-of-the-art methods.	[ { "version": "v1", "created": "Tue, 15 Oct 2024 11:17:19 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 16:22:15 GMT" } ]	2025-03-18T00:00:00	[ [ "An", "Hongyu", "" ], [ "Zhang", "Xinfeng", "" ], [ "Zhao", "Shijie", "" ], [ "Zhang", "Li", "" ], [ "Xiong", "Ruiqin", "" ] ]	TITLE: Spatio-Temporal Distortion Aware Omnidirectional Video Super-Resolution ABSTRACT: Omnidirectional video (ODV) provides an immersive visual experience and is widely utilized in virtual reality and augmented reality. However, restricted capturing devices and transmission bandwidth lead to low-resolution ODVs. Video super-resolution (SR) is proposed to enhance resolution, but practical ODV spatial projection distortions and temporal flickering are not well addressed directly applying existing methods. To achieve better ODV-SR reconstruction, we propose a Spatio-Temporal Distortion Aware Network (STDAN) oriented to ODV characteristics. Specifically, a spatially continuous distortion modulation module is introduced to improve discrete projection distortions. Next, we design an interlaced multi-frame reconstruction mechanism to refine temporal consistency across frames. Furthermore, we incorporate latitude-saliency adaptive weights during training to concentrate on regions with higher texture complexity and human-watching interest. In general, we explore inference-free and real-world viewing matched strategies to provide an application-friendly method on a novel ODV-SR dataset with practical scenarios. Extensive experimental results demonstrate the superior performance of the proposed STDAN over state-of-the-art methods.
2410.11698	Travis Lloyd	Travis Lloyd, Jennah Gosciak, Tung Nguyen, Mor Naaman	AI Rules? Characterizing Reddit Community Policies Towards AI-Generated Content	Forthcoming at ACM CHI 2025	null	null	null	cs.CY cs.SI	http://creativecommons.org/licenses/by/4.0/	How are Reddit communities responding to AI-generated content? We explored this question through a large-scale analysis of subreddit community rules and their change over time. We collected the metadata and community rules for over $300,000$ public subreddits and measured the prevalence of rules governing AI. We labeled subreddits and AI rules according to existing taxonomies from the HCI literature and a new taxonomy we developed specific to AI rules. While rules about AI are still relatively uncommon, the number of subreddits with these rules more than doubled over the course of a year. AI rules are more common in larger subreddits and communities focused on art or celebrity topics, and less common in those focused on social support. These rules often focus on AI images and evoke, as justification, concerns about quality and authenticity. Overall, our findings illustrate the emergence of varied concerns about AI, in different community contexts. Platform designers and HCI researchers should heed these concerns if they hope to encourage community self-determination in the age of generative AI. We make our datasets public to enable future large-scale studies of community self-governance.	[ { "version": "v1", "created": "Tue, 15 Oct 2024 15:31:41 GMT" }, { "version": "v2", "created": "Fri, 20 Dec 2024 19:57:34 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 19:30:03 GMT" } ]	2025-03-18T00:00:00	[ [ "Lloyd", "Travis", "" ], [ "Gosciak", "Jennah", "" ], [ "Nguyen", "Tung", "" ], [ "Naaman", "Mor", "" ] ]	TITLE: AI Rules? Characterizing Reddit Community Policies Towards AI-Generated Content ABSTRACT: How are Reddit communities responding to AI-generated content? We explored this question through a large-scale analysis of subreddit community rules and their change over time. We collected the metadata and community rules for over $300,000$ public subreddits and measured the prevalence of rules governing AI. We labeled subreddits and AI rules according to existing taxonomies from the HCI literature and a new taxonomy we developed specific to AI rules. While rules about AI are still relatively uncommon, the number of subreddits with these rules more than doubled over the course of a year. AI rules are more common in larger subreddits and communities focused on art or celebrity topics, and less common in those focused on social support. These rules often focus on AI images and evoke, as justification, concerns about quality and authenticity. Overall, our findings illustrate the emergence of varied concerns about AI, in different community contexts. Platform designers and HCI researchers should heed these concerns if they hope to encourage community self-determination in the age of generative AI. We make our datasets public to enable future large-scale studies of community self-governance.
2410.14675	Yukun Huang	Yukun Huang, Sanxing Chen, Hongyi Cai, Bhuwan Dhingra	To Trust or Not to Trust? Enhancing Large Language Models' Situated Faithfulness to External Contexts	null	null	null	null	cs.CL cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Large Language Models (LLMs) are often augmented with external contexts, such as those used in retrieval-augmented generation (RAG). However, these contexts can be inaccurate or intentionally misleading, leading to conflicts with the model's internal knowledge. We argue that robust LLMs should demonstrate situated faithfulness, dynamically calibrating their trust in external information based on their confidence in the internal knowledge and the external context to resolve knowledge conflicts. To benchmark this capability, we evaluate LLMs across several QA datasets, including a newly created dataset featuring in-the-wild incorrect contexts sourced from Reddit posts. We show that when provided with both correct and incorrect contexts, both open-source and proprietary models tend to overly rely on external information, regardless of its factual accuracy. To enhance situated faithfulness, we propose two approaches: Self-Guided Confidence Reasoning (SCR) and Rule-Based Confidence Reasoning (RCR). SCR enables models to self-assess the confidence of external information relative to their own internal knowledge to produce the most accurate answer. RCR, in contrast, extracts explicit confidence signals from the LLM and determines the final answer using predefined rules. Our results show that for LLMs with strong reasoning capabilities, such as GPT-4o and GPT-4o mini, SCR outperforms RCR, achieving improvements of up to 24.2% over a direct input augmentation baseline. Conversely, for a smaller model like Llama-3-8B, RCR outperforms SCR. Fine-tuning SCR with our proposed Confidence Reasoning Direct Preference Optimization (CR-DPO) method improves performance on both seen and unseen datasets, yielding an average improvement of 8.9% on Llama-3-8B. In addition to quantitative results, we offer insights into the relative strengths of SCR and RCR.	[ { "version": "v1", "created": "Fri, 18 Oct 2024 17:59:47 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 04:47:58 GMT" } ]	2025-03-18T00:00:00	[ [ "Huang", "Yukun", "" ], [ "Chen", "Sanxing", "" ], [ "Cai", "Hongyi", "" ], [ "Dhingra", "Bhuwan", "" ] ]	TITLE: To Trust or Not to Trust? Enhancing Large Language Models' Situated Faithfulness to External Contexts ABSTRACT: Large Language Models (LLMs) are often augmented with external contexts, such as those used in retrieval-augmented generation (RAG). However, these contexts can be inaccurate or intentionally misleading, leading to conflicts with the model's internal knowledge. We argue that robust LLMs should demonstrate situated faithfulness, dynamically calibrating their trust in external information based on their confidence in the internal knowledge and the external context to resolve knowledge conflicts. To benchmark this capability, we evaluate LLMs across several QA datasets, including a newly created dataset featuring in-the-wild incorrect contexts sourced from Reddit posts. We show that when provided with both correct and incorrect contexts, both open-source and proprietary models tend to overly rely on external information, regardless of its factual accuracy. To enhance situated faithfulness, we propose two approaches: Self-Guided Confidence Reasoning (SCR) and Rule-Based Confidence Reasoning (RCR). SCR enables models to self-assess the confidence of external information relative to their own internal knowledge to produce the most accurate answer. RCR, in contrast, extracts explicit confidence signals from the LLM and determines the final answer using predefined rules. Our results show that for LLMs with strong reasoning capabilities, such as GPT-4o and GPT-4o mini, SCR outperforms RCR, achieving improvements of up to 24.2% over a direct input augmentation baseline. Conversely, for a smaller model like Llama-3-8B, RCR outperforms SCR. Fine-tuning SCR with our proposed Confidence Reasoning Direct Preference Optimization (CR-DPO) method improves performance on both seen and unseen datasets, yielding an average improvement of 8.9% on Llama-3-8B. In addition to quantitative results, we offer insights into the relative strengths of SCR and RCR.
2410.14729	Zixin Wang	Zixin Wang, Dong Gong, Sen Wang, Zi Huang, Yadan Luo	Is Less More? Exploring Token Condensation as Training-free Test-time Adaptation	16 pages, 8 figures	null	null	null	cs.CV cs.AI cs.CL cs.LG	http://creativecommons.org/licenses/by-nc-nd/4.0/	Contrastive Language-Image Pretraining (CLIP) excels at learning generalizable image representations but often falls short in zero-shot inference on certain downstream datasets. Test-time adaptation (TTA) mitigates this issue by adjusting components like normalization layers or context prompts, yet it typically requires large batch sizes and extensive augmentations, leading to high computational costs. This raises a key question: Can VLMs' performance drop in specific test cases be mitigated through efficient, training-free approaches? To explore the solution, we investigate token condensation (TC) techniques, originally designed to enhance vision transformer efficiency by refining token usage during inference. We observe that informative tokens improve visual-text alignment in VLMs like CLIP on unseen datasets. However, existing TC methods often fail to maintain in-distribution performance when reducing tokens, prompting us to ask: How can we transform TC into an effective ``free-lunch'' adaptation strategy for VLMs? To address this, we propose Token Condensation as Adaptation (TCA), a training-free adaptation method that takes a step beyond standard TC. Rather than passively discarding tokens, TCA condenses token representation by introducing reservoir-based domain anchor tokens for information-preserving token reduction and logits correction. TCA achieves up to a 21.4% performance improvement over the strongest baseline on cross-dataset benchmark and the CIFAR-100-Corrupted dataset while reducing GFLOPs by 12.2% to 48.9%, with minimal hyperparameter dependency on both CLIP and SigLIP series.	[ { "version": "v1", "created": "Wed, 16 Oct 2024 07:13:35 GMT" }, { "version": "v2", "created": "Thu, 21 Nov 2024 12:17:29 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 09:01:31 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Zixin", "" ], [ "Gong", "Dong", "" ], [ "Wang", "Sen", "" ], [ "Huang", "Zi", "" ], [ "Luo", "Yadan", "" ] ]	TITLE: Is Less More? Exploring Token Condensation as Training-free Test-time Adaptation ABSTRACT: Contrastive Language-Image Pretraining (CLIP) excels at learning generalizable image representations but often falls short in zero-shot inference on certain downstream datasets. Test-time adaptation (TTA) mitigates this issue by adjusting components like normalization layers or context prompts, yet it typically requires large batch sizes and extensive augmentations, leading to high computational costs. This raises a key question: Can VLMs' performance drop in specific test cases be mitigated through efficient, training-free approaches? To explore the solution, we investigate token condensation (TC) techniques, originally designed to enhance vision transformer efficiency by refining token usage during inference. We observe that informative tokens improve visual-text alignment in VLMs like CLIP on unseen datasets. However, existing TC methods often fail to maintain in-distribution performance when reducing tokens, prompting us to ask: How can we transform TC into an effective ``free-lunch'' adaptation strategy for VLMs? To address this, we propose Token Condensation as Adaptation (TCA), a training-free adaptation method that takes a step beyond standard TC. Rather than passively discarding tokens, TCA condenses token representation by introducing reservoir-based domain anchor tokens for information-preserving token reduction and logits correction. TCA achieves up to a 21.4% performance improvement over the strongest baseline on cross-dataset benchmark and the CIFAR-100-Corrupted dataset while reducing GFLOPs by 12.2% to 48.9%, with minimal hyperparameter dependency on both CLIP and SigLIP series.
2410.15143	Minhyuk Seo	Minhyuk Seo, Hyunseo Koh, Jonghyun Choi	Budgeted Online Continual Learning by Adaptive Layer Freezing and Frequency-based Sampling	ICLR 2025 Spotlight	null	null	null	cs.LG cs.AI cs.CV	http://creativecommons.org/licenses/by-nc-sa/4.0/	The majority of online continual learning (CL) advocates single-epoch training and imposes restrictions on the size of replay memory. However, single-epoch training would incur a different amount of computations per CL algorithm, and the additional storage cost to store logit or model in addition to replay memory is largely ignored in calculating the storage budget. Arguing different computational and storage budgets hinder fair comparison among CL algorithms in practice, we propose to use floating point operations (FLOPs) and total memory size in Byte as a metric for computational and memory budgets, respectively, to compare and develop CL algorithms in the same 'total resource budget.' To improve a CL method in a limited total budget, we propose adaptive layer freezing that does not update the layers for less informative batches to reduce computational costs with a negligible loss of accuracy. In addition, we propose a memory retrieval method that allows the model to learn the same amount of knowledge as using random retrieval in fewer iterations. Empirical validations on the CIFAR-10/100, CLEAR-10/100, and ImageNet-1K datasets demonstrate that the proposed approach outperforms the state-of-the-art methods within the same total budget	[ { "version": "v1", "created": "Sat, 19 Oct 2024 16:00:00 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 20:18:42 GMT" } ]	2025-03-18T00:00:00	[ [ "Seo", "Minhyuk", "" ], [ "Koh", "Hyunseo", "" ], [ "Choi", "Jonghyun", "" ] ]	TITLE: Budgeted Online Continual Learning by Adaptive Layer Freezing and Frequency-based Sampling ABSTRACT: The majority of online continual learning (CL) advocates single-epoch training and imposes restrictions on the size of replay memory. However, single-epoch training would incur a different amount of computations per CL algorithm, and the additional storage cost to store logit or model in addition to replay memory is largely ignored in calculating the storage budget. Arguing different computational and storage budgets hinder fair comparison among CL algorithms in practice, we propose to use floating point operations (FLOPs) and total memory size in Byte as a metric for computational and memory budgets, respectively, to compare and develop CL algorithms in the same 'total resource budget.' To improve a CL method in a limited total budget, we propose adaptive layer freezing that does not update the layers for less informative batches to reduce computational costs with a negligible loss of accuracy. In addition, we propose a memory retrieval method that allows the model to learn the same amount of knowledge as using random retrieval in fewer iterations. Empirical validations on the CIFAR-10/100, CLEAR-10/100, and ImageNet-1K datasets demonstrate that the proposed approach outperforms the state-of-the-art methods within the same total budget
2410.15154	Yin Li	Yin Li, Liangwei Wang, Shiyuan Piao, Boo-Ho Yang, Ziyue Li, Wei Zeng, and Fugee Tsung	MCCoder: Streamlining Motion Control with LLM-Assisted Code Generation and Rigorous Verification	null	null	null	null	cs.AI	http://creativecommons.org/licenses/by-nc-nd/4.0/	Large Language Models (LLMs) have demonstrated significant potential in code generation. However, in the factory automation sector, particularly motion control, manual programming, alongside inefficient and unsafe debugging practices, remains prevalent. This stems from the complex interplay of mechanical and electrical systems and stringent safety requirements. Moreover, most current AI-assisted motion control programming efforts focus on PLCs, with little attention given to high-level languages and function libraries. To address these challenges, we introduce MCCoder, an LLM-powered system tailored for generating motion control code, integrated with a soft-motion controller. MCCoder improves code generation through a structured workflow that combines multitask decomposition, hybrid retrieval-augmented generation (RAG), and iterative self-correction, utilizing a well-established motion library. Additionally, it integrates a 3D simulator for intuitive motion validation and logs of full motion trajectories for data verification, significantly enhancing accuracy and safety. In the absence of benchmark datasets and metrics tailored for evaluating motion control code generation, we propose MCEVAL, a dataset spanning motion tasks of varying complexity. Experiments show that MCCoder outperforms baseline models using Advanced RAG, achieving an overall performance gain of 33.09% and a 131.77% improvement on complex tasks in the MCEVAL dataset.	[ { "version": "v1", "created": "Sat, 19 Oct 2024 16:46:21 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 06:03:20 GMT" } ]	2025-03-18T00:00:00	[ [ "Li", "Yin", "" ], [ "Wang", "Liangwei", "" ], [ "Piao", "Shiyuan", "" ], [ "Yang", "Boo-Ho", "" ], [ "Li", "Ziyue", "" ], [ "Zeng", "Wei", "" ], [ "Tsung", "Fugee", "" ] ]	TITLE: MCCoder: Streamlining Motion Control with LLM-Assisted Code Generation and Rigorous Verification ABSTRACT: Large Language Models (LLMs) have demonstrated significant potential in code generation. However, in the factory automation sector, particularly motion control, manual programming, alongside inefficient and unsafe debugging practices, remains prevalent. This stems from the complex interplay of mechanical and electrical systems and stringent safety requirements. Moreover, most current AI-assisted motion control programming efforts focus on PLCs, with little attention given to high-level languages and function libraries. To address these challenges, we introduce MCCoder, an LLM-powered system tailored for generating motion control code, integrated with a soft-motion controller. MCCoder improves code generation through a structured workflow that combines multitask decomposition, hybrid retrieval-augmented generation (RAG), and iterative self-correction, utilizing a well-established motion library. Additionally, it integrates a 3D simulator for intuitive motion validation and logs of full motion trajectories for data verification, significantly enhancing accuracy and safety. In the absence of benchmark datasets and metrics tailored for evaluating motion control code generation, we propose MCEVAL, a dataset spanning motion tasks of varying complexity. Experiments show that MCCoder outperforms baseline models using Advanced RAG, achieving an overall performance gain of 33.09% and a 131.77% improvement on complex tasks in the MCEVAL dataset.
2410.18325	Kim Sung-Bin	Kim Sung-Bin, Oh Hyun-Bin, JungMok Lee, Arda Senocak, Joon Son Chung, Tae-Hyun Oh	AVHBench: A Cross-Modal Hallucination Benchmark for Audio-Visual Large Language Models	ICLR 2025	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Following the success of Large Language Models (LLMs), expanding their boundaries to new modalities represents a significant paradigm shift in multimodal understanding. Human perception is inherently multimodal, relying not only on text but also on auditory and visual cues for a complete understanding of the world. In recognition of this fact, audio-visual LLMs have recently emerged. Despite promising developments, the lack of dedicated benchmarks poses challenges for understanding and evaluating models. In this work, we show that audio-visual LLMs struggle to discern subtle relationships between audio and visual signals, leading to hallucinations and highlighting the need for reliable benchmarks. To address this, we introduce AVHBench, the first comprehensive benchmark specifically designed to evaluate the perception and comprehension capabilities of audio-visual LLMs. Our benchmark includes tests for assessing hallucinations, as well as the cross-modal matching and reasoning abilities of these models. Our results reveal that most existing audio-visual LLMs struggle with hallucinations caused by cross-interactions between modalities, due to their limited capacity to perceive complex multimodal signals and their relationships. Additionally, we demonstrate that simple training with our AVHBench improves robustness of audio-visual LLMs against hallucinations. Dataset: https://github.com/kaist-ami/AVHBench	[ { "version": "v1", "created": "Wed, 23 Oct 2024 23:36:06 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 08:14:35 GMT" } ]	2025-03-18T00:00:00	[ [ "Sung-Bin", "Kim", "" ], [ "Hyun-Bin", "Oh", "" ], [ "Lee", "JungMok", "" ], [ "Senocak", "Arda", "" ], [ "Chung", "Joon Son", "" ], [ "Oh", "Tae-Hyun", "" ] ]	TITLE: AVHBench: A Cross-Modal Hallucination Benchmark for Audio-Visual Large Language Models ABSTRACT: Following the success of Large Language Models (LLMs), expanding their boundaries to new modalities represents a significant paradigm shift in multimodal understanding. Human perception is inherently multimodal, relying not only on text but also on auditory and visual cues for a complete understanding of the world. In recognition of this fact, audio-visual LLMs have recently emerged. Despite promising developments, the lack of dedicated benchmarks poses challenges for understanding and evaluating models. In this work, we show that audio-visual LLMs struggle to discern subtle relationships between audio and visual signals, leading to hallucinations and highlighting the need for reliable benchmarks. To address this, we introduce AVHBench, the first comprehensive benchmark specifically designed to evaluate the perception and comprehension capabilities of audio-visual LLMs. Our benchmark includes tests for assessing hallucinations, as well as the cross-modal matching and reasoning abilities of these models. Our results reveal that most existing audio-visual LLMs struggle with hallucinations caused by cross-interactions between modalities, due to their limited capacity to perceive complex multimodal signals and their relationships. Additionally, we demonstrate that simple training with our AVHBench improves robustness of audio-visual LLMs against hallucinations. Dataset: https://github.com/kaist-ami/AVHBench
2410.18656	Torbj{\o}rn Smith	Torbj{\o}rn Smith and Olav Egeland	Learning dissipative Hamiltonian dynamics with reproducing kernel Hilbert spaces and random Fourier features	null	IFAC-PapersOnLine: The 4th Modeling, Estimation, and Control Conference - 2024	10.1016/j.ifacol.2025.01.146	null	cs.LG cs.RO	http://creativecommons.org/licenses/by/4.0/	This paper presents a new method for learning dissipative Hamiltonian dynamics from a limited and noisy dataset. The method uses the Helmholtz decomposition to learn a vector field as the sum of a symplectic and a dissipative vector field. The two vector fields are learned using two reproducing kernel Hilbert spaces, defined by a symplectic and a curl-free kernel, where the kernels are specialized to enforce odd symmetry. Random Fourier features are used to approximate the kernels to reduce the dimension of the optimization problem. The performance of the method is validated in simulations for two dissipative Hamiltonian systems, and it is shown that the method improves predictive accuracy significantly compared to a method where a Gaussian separable kernel is used.	[ { "version": "v1", "created": "Thu, 24 Oct 2024 11:35:39 GMT" } ]	2025-03-18T00:00:00	[ [ "Smith", "Torbjørn", "" ], [ "Egeland", "Olav", "" ] ]	TITLE: Learning dissipative Hamiltonian dynamics with reproducing kernel Hilbert spaces and random Fourier features ABSTRACT: This paper presents a new method for learning dissipative Hamiltonian dynamics from a limited and noisy dataset. The method uses the Helmholtz decomposition to learn a vector field as the sum of a symplectic and a dissipative vector field. The two vector fields are learned using two reproducing kernel Hilbert spaces, defined by a symplectic and a curl-free kernel, where the kernels are specialized to enforce odd symmetry. Random Fourier features are used to approximate the kernels to reduce the dimension of the optimization problem. The performance of the method is validated in simulations for two dissipative Hamiltonian systems, and it is shown that the method improves predictive accuracy significantly compared to a method where a Gaussian separable kernel is used.
2410.19371	Talal Alrawajfeh	Talal Alrawajfeh, Joonas J\"alk\"o, Antti Honkela	Noise-Aware Differentially Private Variational Inference	null	null	null	null	stat.ML cs.CR cs.LG	http://creativecommons.org/licenses/by/4.0/	Differential privacy (DP) provides robust privacy guarantees for statistical inference, but this can lead to unreliable results and biases in downstream applications. While several noise-aware approaches have been proposed which integrate DP perturbation into the inference, they are limited to specific types of simple probabilistic models. In this work, we propose a novel method for noise-aware approximate Bayesian inference based on stochastic gradient variational inference which can also be applied to high-dimensional and non-conjugate models. We also propose a more accurate evaluation method for noise-aware posteriors. Empirically, our inference method has similar performance to existing methods in the domain where they are applicable. Outside this domain, we obtain accurate coverages on high-dimensional Bayesian linear regression and well-calibrated predictive probabilities on Bayesian logistic regression with the UCI Adult dataset.	[ { "version": "v1", "created": "Fri, 25 Oct 2024 08:18:49 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 13:23:33 GMT" } ]	2025-03-18T00:00:00	[ [ "Alrawajfeh", "Talal", "" ], [ "Jälkö", "Joonas", "" ], [ "Honkela", "Antti", "" ] ]	TITLE: Noise-Aware Differentially Private Variational Inference ABSTRACT: Differential privacy (DP) provides robust privacy guarantees for statistical inference, but this can lead to unreliable results and biases in downstream applications. While several noise-aware approaches have been proposed which integrate DP perturbation into the inference, they are limited to specific types of simple probabilistic models. In this work, we propose a novel method for noise-aware approximate Bayesian inference based on stochastic gradient variational inference which can also be applied to high-dimensional and non-conjugate models. We also propose a more accurate evaluation method for noise-aware posteriors. Empirically, our inference method has similar performance to existing methods in the domain where they are applicable. Outside this domain, we obtain accurate coverages on high-dimensional Bayesian linear regression and well-calibrated predictive probabilities on Bayesian logistic regression with the UCI Adult dataset.
2410.23642	Ramin Nateghi	Ramin Nateghi, Ruoji Zhou, Madeline Saft, Marina Schnauss, Clayton Neill, Ridwan Alam, Nicole Handa, Mitchell Huang, Eric V Li, Jeffery A Goldstein, Edward M Schaeffer, Menatalla Nadim, Fattaneh Pourakpour, Bogdan Isaila, Christopher Felicelli, Vikas Mehta, Behtash G Nezami, Ashley Ross, Ximing Yang, Lee AD Cooper	Development and prospective validation of a prostate cancer detection, grading, and workflow optimization system at an academic medical center	null	null	null	null	eess.IV cs.CV	http://creativecommons.org/licenses/by-nc-nd/4.0/	Artificial intelligence may assist healthcare systems in meeting increasing demand for pathology services while maintaining diagnostic quality and reducing turnaround time and costs. We aimed to investigate the performance of an institutionally developed system for prostate cancer detection, grading, and workflow optimization and to contrast this with commercial alternatives. From August 2021 to March 2023, we scanned 21,396 slides from 1,147 patients receiving prostate biopsy. We developed models for cancer detection, grading, and screening of equivocal cases for IHC ordering. We compared the performance of task-specific prostate models with general-purpose foundation models in a prospectively collected dataset that reflects our patient population. We also evaluated the contributions of a bespoke model designed to improve sensitivity to small cancer foci and perception of low-resolution patterns. We found high concordance with pathologist ground-truth in detection (area under curve 98.5%, sensitivity 95.0%, and specificity 97.8%), ISUP grading (Cohen's kappa 0.869), grade group 3 or higher classification (area under curve 97.5%, sensitivity 94.9%, specificity 96.6%). Screening models could correctly classify 55% of biopsy blocks where immunohistochemistry was ordered with a 1.4% error rate. No statistically significant differences were observed between task-specific and foundation models in cancer detection, although the task-specific model is significantly smaller and faster. Institutions like academic medical centers that have high scanning volumes and report abstraction capabilities can develop highly accurate computational pathology models for internal use. These models have the potential to aid in quality control role and to improve resource allocation and workflow in the pathology lab to help meet future challenges in prostate cancer diagnosis.	[ { "version": "v1", "created": "Thu, 31 Oct 2024 05:29:18 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 22:39:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Nateghi", "Ramin", "" ], [ "Zhou", "Ruoji", "" ], [ "Saft", "Madeline", "" ], [ "Schnauss", "Marina", "" ], [ "Neill", "Clayton", "" ], [ "Alam", "Ridwan", "" ], [ "Handa", "Nicole", "" ], [ "Huang", "Mitchell", "" ], [ "Li", "Eric V", "" ], [ "Goldstein", "Jeffery A", "" ], [ "Schaeffer", "Edward M", "" ], [ "Nadim", "Menatalla", "" ], [ "Pourakpour", "Fattaneh", "" ], [ "Isaila", "Bogdan", "" ], [ "Felicelli", "Christopher", "" ], [ "Mehta", "Vikas", "" ], [ "Nezami", "Behtash G", "" ], [ "Ross", "Ashley", "" ], [ "Yang", "Ximing", "" ], [ "Cooper", "Lee AD", "" ] ]	TITLE: Development and prospective validation of a prostate cancer detection, grading, and workflow optimization system at an academic medical center ABSTRACT: Artificial intelligence may assist healthcare systems in meeting increasing demand for pathology services while maintaining diagnostic quality and reducing turnaround time and costs. We aimed to investigate the performance of an institutionally developed system for prostate cancer detection, grading, and workflow optimization and to contrast this with commercial alternatives. From August 2021 to March 2023, we scanned 21,396 slides from 1,147 patients receiving prostate biopsy. We developed models for cancer detection, grading, and screening of equivocal cases for IHC ordering. We compared the performance of task-specific prostate models with general-purpose foundation models in a prospectively collected dataset that reflects our patient population. We also evaluated the contributions of a bespoke model designed to improve sensitivity to small cancer foci and perception of low-resolution patterns. We found high concordance with pathologist ground-truth in detection (area under curve 98.5%, sensitivity 95.0%, and specificity 97.8%), ISUP grading (Cohen's kappa 0.869), grade group 3 or higher classification (area under curve 97.5%, sensitivity 94.9%, specificity 96.6%). Screening models could correctly classify 55% of biopsy blocks where immunohistochemistry was ordered with a 1.4% error rate. No statistically significant differences were observed between task-specific and foundation models in cancer detection, although the task-specific model is significantly smaller and faster. Institutions like academic medical centers that have high scanning volumes and report abstraction capabilities can develop highly accurate computational pathology models for internal use. These models have the potential to aid in quality control role and to improve resource allocation and workflow in the pathology lab to help meet future challenges in prostate cancer diagnosis.
2410.23996	Chenyu Wang	Chenyu Wang, Sharut Gupta, Xinyi Zhang, Sana Tonekaboni, Stefanie Jegelka, Tommi Jaakkola, Caroline Uhler	An Information Criterion for Controlled Disentanglement of Multimodal Data	ICLR 2025	null	null	null	cs.LG cs.AI cs.IT math.IT	http://creativecommons.org/licenses/by/4.0/	Multimodal representation learning seeks to relate and decompose information inherent in multiple modalities. By disentangling modality-specific information from information that is shared across modalities, we can improve interpretability and robustness and enable downstream tasks such as the generation of counterfactual outcomes. Separating the two types of information is challenging since they are often deeply entangled in many real-world applications. We propose Disentangled Self-Supervised Learning (DisentangledSSL), a novel self-supervised approach for learning disentangled representations. We present a comprehensive analysis of the optimality of each disentangled representation, particularly focusing on the scenario not covered in prior work where the so-called Minimum Necessary Information (MNI) point is not attainable. We demonstrate that DisentangledSSL successfully learns shared and modality-specific features on multiple synthetic and real-world datasets and consistently outperforms baselines on various downstream tasks, including prediction tasks for vision-language data, as well as molecule-phenotype retrieval tasks for biological data. The code is available at https://github.com/uhlerlab/DisentangledSSL.	[ { "version": "v1", "created": "Thu, 31 Oct 2024 14:57:31 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 16:27:27 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Chenyu", "" ], [ "Gupta", "Sharut", "" ], [ "Zhang", "Xinyi", "" ], [ "Tonekaboni", "Sana", "" ], [ "Jegelka", "Stefanie", "" ], [ "Jaakkola", "Tommi", "" ], [ "Uhler", "Caroline", "" ] ]	TITLE: An Information Criterion for Controlled Disentanglement of Multimodal Data ABSTRACT: Multimodal representation learning seeks to relate and decompose information inherent in multiple modalities. By disentangling modality-specific information from information that is shared across modalities, we can improve interpretability and robustness and enable downstream tasks such as the generation of counterfactual outcomes. Separating the two types of information is challenging since they are often deeply entangled in many real-world applications. We propose Disentangled Self-Supervised Learning (DisentangledSSL), a novel self-supervised approach for learning disentangled representations. We present a comprehensive analysis of the optimality of each disentangled representation, particularly focusing on the scenario not covered in prior work where the so-called Minimum Necessary Information (MNI) point is not attainable. We demonstrate that DisentangledSSL successfully learns shared and modality-specific features on multiple synthetic and real-world datasets and consistently outperforms baselines on various downstream tasks, including prediction tasks for vision-language data, as well as molecule-phenotype retrieval tasks for biological data. The code is available at https://github.com/uhlerlab/DisentangledSSL.
2410.24160	Fu Feng	Fu Feng, Yucheng Xie, Xu Yang, Jing Wang, Xin Geng	Redefining <Creative> in Dictionary: Towards an Enhanced Semantic Understanding of Creative Generation	null	null	null	null	cs.CV cs.CL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	``Creative'' remains an inherently abstract concept for both humans and diffusion models. While text-to-image (T2I) diffusion models can easily generate out-of-distribution concepts like ``a blue banana'', they struggle with generating combinatorial objects such as ``a creative mixture that resembles a lettuce and a mantis'', due to difficulties in understanding the semantic depth of ``creative''. Current methods rely heavily on synthesizing reference prompts or images to achieve a creative effect, typically requiring retraining for each unique creative output-a process that is computationally intensive and limits practical applications. To address this, we introduce CreTok, which brings meta-creativity to diffusion models by redefining ``creative'' as a new token, \texttt{<CreTok>}, thus enhancing models' semantic understanding for combinatorial creativity. CreTok achieves such redefinition by iteratively sampling diverse text pairs from our proposed CangJie dataset to form adaptive prompts and restrictive prompts, and then optimizing the similarity between their respective text embeddings. Extensive experiments demonstrate that <CreTok> enables the universal and direct generation of combinatorial creativity across diverse concepts without additional training, achieving state-of-the-art performance with improved text-image alignment and higher human preference ratings. Code will be made available at https://github.com/fu-feng/CreTok.	[ { "version": "v1", "created": "Thu, 31 Oct 2024 17:19:03 GMT" }, { "version": "v2", "created": "Wed, 20 Nov 2024 10:22:59 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 06:33:07 GMT" } ]	2025-03-18T00:00:00	[ [ "Feng", "Fu", "" ], [ "Xie", "Yucheng", "" ], [ "Yang", "Xu", "" ], [ "Wang", "Jing", "" ], [ "Geng", "Xin", "" ] ]	TITLE: Redefining <Creative> in Dictionary: Towards an Enhanced Semantic Understanding of Creative Generation ABSTRACT: ``Creative'' remains an inherently abstract concept for both humans and diffusion models. While text-to-image (T2I) diffusion models can easily generate out-of-distribution concepts like ``a blue banana'', they struggle with generating combinatorial objects such as ``a creative mixture that resembles a lettuce and a mantis'', due to difficulties in understanding the semantic depth of ``creative''. Current methods rely heavily on synthesizing reference prompts or images to achieve a creative effect, typically requiring retraining for each unique creative output-a process that is computationally intensive and limits practical applications. To address this, we introduce CreTok, which brings meta-creativity to diffusion models by redefining ``creative'' as a new token, \texttt{<CreTok>}, thus enhancing models' semantic understanding for combinatorial creativity. CreTok achieves such redefinition by iteratively sampling diverse text pairs from our proposed CangJie dataset to form adaptive prompts and restrictive prompts, and then optimizing the similarity between their respective text embeddings. Extensive experiments demonstrate that <CreTok> enables the universal and direct generation of combinatorial creativity across diverse concepts without additional training, achieving state-of-the-art performance with improved text-image alignment and higher human preference ratings. Code will be made available at https://github.com/fu-feng/CreTok.
2411.01841	XiaoBei Niu	Shi Dong and Xiaobei Niu and Rui Zhong and Zhifeng Wang and Mingzhang Zuo	Leveraging Label Semantics and Meta-Label Refinement for Multi-Label Question Classification	null	null	null	null	cs.LG cs.CL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Accurate annotation of educational resources is crucial for effective personalized learning and resource recommendation in online education. However, fine-grained knowledge labels often overlap or share similarities, making it difficult for existing multi-label classification methods to differentiate them. The label distribution imbalance due to sparsity of human annotations further intensifies these challenges. To address these issues, this paper introduces RR2QC, a novel Retrieval Reranking method to multi-label Question Classification by leveraging label semantics and meta-label refinement. First, RR2QC improves the pre-training strategy by utilizing semantic relationships within and across label groups. Second, it introduces a class center learning task to align questions with label semantics during downstream training. Finally, this method decomposes labels into meta-labels and uses a meta-label classifier to rerank the retrieved label sequences. In doing so, RR2QC enhances the understanding and prediction capability of long-tail labels by learning from meta-labels that frequently appear in other labels. Additionally, a mathematical LLM is used to generate solutions for questions, extracting latent information to further refine the model's insights. Experimental results show that RR2QC outperforms existing methods in Precision@K and F1 scores across multiple educational datasets, demonstrating its effectiveness for online education applications. The code and datasets are available at https://github.com/78Erii/RR2QC.	[ { "version": "v1", "created": "Mon, 4 Nov 2024 06:27:14 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 17:31:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Dong", "Shi", "" ], [ "Niu", "Xiaobei", "" ], [ "Zhong", "Rui", "" ], [ "Wang", "Zhifeng", "" ], [ "Zuo", "Mingzhang", "" ] ]	TITLE: Leveraging Label Semantics and Meta-Label Refinement for Multi-Label Question Classification ABSTRACT: Accurate annotation of educational resources is crucial for effective personalized learning and resource recommendation in online education. However, fine-grained knowledge labels often overlap or share similarities, making it difficult for existing multi-label classification methods to differentiate them. The label distribution imbalance due to sparsity of human annotations further intensifies these challenges. To address these issues, this paper introduces RR2QC, a novel Retrieval Reranking method to multi-label Question Classification by leveraging label semantics and meta-label refinement. First, RR2QC improves the pre-training strategy by utilizing semantic relationships within and across label groups. Second, it introduces a class center learning task to align questions with label semantics during downstream training. Finally, this method decomposes labels into meta-labels and uses a meta-label classifier to rerank the retrieved label sequences. In doing so, RR2QC enhances the understanding and prediction capability of long-tail labels by learning from meta-labels that frequently appear in other labels. Additionally, a mathematical LLM is used to generate solutions for questions, extracting latent information to further refine the model's insights. Experimental results show that RR2QC outperforms existing methods in Precision@K and F1 scores across multiple educational datasets, demonstrating its effectiveness for online education applications. The code and datasets are available at https://github.com/78Erii/RR2QC.
2411.02136	Robert Fonod	Robert Fonod and Haechan Cho and Hwasoo Yeo and Nikolas Geroliminis	Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery	null	null	null	null	cs.CV cs.AI cs.LG	http://creativecommons.org/licenses/by/4.0/	This paper presents a framework for extracting georeferenced vehicle trajectories from high-altitude drone imagery, addressing key challenges in urban traffic monitoring and the limitations of traditional ground-based systems. Our approach integrates several novel contributions, including a tailored object detector optimized for high-altitude bird's-eye view perspectives, a unique track stabilization method that uses detected vehicle bounding boxes as exclusion masks during image registration, and an orthophoto and master frame-based georeferencing strategy that enhances consistent alignment across multiple drone viewpoints. Additionally, our framework features robust vehicle dimension estimation and detailed road segmentation, enabling comprehensive traffic analysis. Conducted in the Songdo International Business District, South Korea, the study utilized a multi-drone experiment covering 20 intersections, capturing approximately 12TB of 4K video data over four days. The framework produced two high-quality datasets: the Songdo Traffic dataset, comprising approximately 700,000 unique vehicle trajectories, and the Songdo Vision dataset, containing over 5,000 human-annotated images with about 300,000 vehicle instances in four classes. Comparisons with high-precision sensor data from an instrumented probe vehicle highlight the accuracy and consistency of our extraction pipeline in dense urban environments. The public release of Songdo Traffic and Songdo Vision, and the complete source code for the extraction pipeline, establishes new benchmarks in data quality, reproducibility, and scalability in traffic research. Results demonstrate the potential of integrating drone technology with advanced computer vision for precise and cost-effective urban traffic monitoring, providing valuable resources for developing intelligent transportation systems and enhancing traffic management strategies.	[ { "version": "v1", "created": "Mon, 4 Nov 2024 14:49:01 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 09:25:50 GMT" } ]	2025-03-18T00:00:00	[ [ "Fonod", "Robert", "" ], [ "Cho", "Haechan", "" ], [ "Yeo", "Hwasoo", "" ], [ "Geroliminis", "Nikolas", "" ] ]	TITLE: Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery ABSTRACT: This paper presents a framework for extracting georeferenced vehicle trajectories from high-altitude drone imagery, addressing key challenges in urban traffic monitoring and the limitations of traditional ground-based systems. Our approach integrates several novel contributions, including a tailored object detector optimized for high-altitude bird's-eye view perspectives, a unique track stabilization method that uses detected vehicle bounding boxes as exclusion masks during image registration, and an orthophoto and master frame-based georeferencing strategy that enhances consistent alignment across multiple drone viewpoints. Additionally, our framework features robust vehicle dimension estimation and detailed road segmentation, enabling comprehensive traffic analysis. Conducted in the Songdo International Business District, South Korea, the study utilized a multi-drone experiment covering 20 intersections, capturing approximately 12TB of 4K video data over four days. The framework produced two high-quality datasets: the Songdo Traffic dataset, comprising approximately 700,000 unique vehicle trajectories, and the Songdo Vision dataset, containing over 5,000 human-annotated images with about 300,000 vehicle instances in four classes. Comparisons with high-precision sensor data from an instrumented probe vehicle highlight the accuracy and consistency of our extraction pipeline in dense urban environments. The public release of Songdo Traffic and Songdo Vision, and the complete source code for the extraction pipeline, establishes new benchmarks in data quality, reproducibility, and scalability in traffic research. Results demonstrate the potential of integrating drone technology with advanced computer vision for precise and cost-effective urban traffic monitoring, providing valuable resources for developing intelligent transportation systems and enhancing traffic management strategies.
2411.06518	Yuewen Sun	Yuewen Sun, Lingjing Kong, Guangyi Chen, Loka Li, Gongxu Luo, Zijian Li, Yixuan Zhang, Yujia Zheng, Mengyue Yang, Petar Stojanov, Eran Segal, Eric P. Xing, Kun Zhang	Causal Representation Learning from Multimodal Biomedical Observations	null	null	null	null	cs.LG q-bio.QM stat.ME	http://creativecommons.org/licenses/by/4.0/	Prevalent in biomedical applications (e.g., human phenotype research), multimodal datasets can provide valuable insights into the underlying physiological mechanisms. However, current machine learning (ML) models designed to analyze these datasets often lack interpretability and identifiability guarantees, which are essential for biomedical research. Recent advances in causal representation learning have shown promise in identifying interpretable latent causal variables with formal theoretical guarantees. Unfortunately, most current work on multimodal distributions either relies on restrictive parametric assumptions or yields only coarse identification results, limiting their applicability to biomedical research that favors a detailed understanding of the mechanisms. In this work, we aim to develop flexible identification conditions for multimodal data and principled methods to facilitate the understanding of biomedical datasets. Theoretically, we consider a nonparametric latent distribution (c.f., parametric assumptions in previous work) that allows for causal relationships across potentially different modalities. We establish identifiability guarantees for each latent component, extending the subspace identification results from previous work. Our key theoretical contribution is the structural sparsity of causal connections between modalities, which, as we will discuss, is natural for a large collection of biomedical systems. Empirically, we present a practical framework to instantiate our theoretical insights. We demonstrate the effectiveness of our approach through extensive experiments on both numerical and synthetic datasets. Results on a real-world human phenotype dataset are consistent with established biomedical research, validating our theoretical and methodological framework.	[ { "version": "v1", "created": "Sun, 10 Nov 2024 16:40:27 GMT" }, { "version": "v2", "created": "Thu, 13 Mar 2025 08:56:49 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 13:07:14 GMT" } ]	2025-03-18T00:00:00	[ [ "Sun", "Yuewen", "" ], [ "Kong", "Lingjing", "" ], [ "Chen", "Guangyi", "" ], [ "Li", "Loka", "" ], [ "Luo", "Gongxu", "" ], [ "Li", "Zijian", "" ], [ "Zhang", "Yixuan", "" ], [ "Zheng", "Yujia", "" ], [ "Yang", "Mengyue", "" ], [ "Stojanov", "Petar", "" ], [ "Segal", "Eran", "" ], [ "Xing", "Eric P.", "" ], [ "Zhang", "Kun", "" ] ]	TITLE: Causal Representation Learning from Multimodal Biomedical Observations ABSTRACT: Prevalent in biomedical applications (e.g., human phenotype research), multimodal datasets can provide valuable insights into the underlying physiological mechanisms. However, current machine learning (ML) models designed to analyze these datasets often lack interpretability and identifiability guarantees, which are essential for biomedical research. Recent advances in causal representation learning have shown promise in identifying interpretable latent causal variables with formal theoretical guarantees. Unfortunately, most current work on multimodal distributions either relies on restrictive parametric assumptions or yields only coarse identification results, limiting their applicability to biomedical research that favors a detailed understanding of the mechanisms. In this work, we aim to develop flexible identification conditions for multimodal data and principled methods to facilitate the understanding of biomedical datasets. Theoretically, we consider a nonparametric latent distribution (c.f., parametric assumptions in previous work) that allows for causal relationships across potentially different modalities. We establish identifiability guarantees for each latent component, extending the subspace identification results from previous work. Our key theoretical contribution is the structural sparsity of causal connections between modalities, which, as we will discuss, is natural for a large collection of biomedical systems. Empirically, we present a practical framework to instantiate our theoretical insights. We demonstrate the effectiveness of our approach through extensive experiments on both numerical and synthetic datasets. Results on a real-world human phenotype dataset are consistent with established biomedical research, validating our theoretical and methodological framework.
2411.06802	Yuxiu Shao	Yuxiu Shao (1 and 2), David Dahmen (3), Stefano Recanatesi (4), Eric Shea-Brown (5 and 6), Srdjan Ostojic (2) ((1) School of Systems Science, Beijing Normal University, China, (2) Laboratoire de Neurosciences Cognitives et Computationnelles, INSERM U960, Ecole Normale Superieure - PSL Research University, France, (3) Institute for Advanced Simulation (IAS-6) Computational and Systems Neuroscience, J\"ulich Research Center, Germany, (4) Technion, Israel Institute of Technology, Israel, (5) Department of Applied Mathematics and Computational Neuroscience Center, University of Washington, USA, (6) Allen Institute for Brain Science, USA)	Identifying the impact of local connectivity patterns on dynamics in excitatory-inhibitory networks	30 pages, 17 figures	null	null	null	q-bio.NC cond-mat.dis-nn cs.NE	http://creativecommons.org/licenses/by-nc-nd/4.0/	Networks of excitatory and inhibitory (EI) neurons form a canonical circuit in the brain. Seminal theoretical results on dynamics of such networks are based on the assumption that synaptic strengths depend on the type of neurons they connect, but are otherwise statistically independent. Recent synaptic physiology datasets however highlight the prominence of specific connectivity patterns that go well beyond what is expected from independent connections. While decades of influential research have demonstrated the strong role of the basic EI cell type structure, to which extent additional connectivity features influence dynamics remains to be fully determined. Here we examine the effects of pairwise connectivity motifs on the linear dynamics in EI networks using an analytical framework that approximates the connectivity in terms of low-rank structures. This low-rank approximation is based on a mathematical derivation of the dominant eigenvalues of the connectivity matrix and predicts the impact on responses to external inputs of connectivity motifs and their interactions with cell-type structure. Our results reveal that a particular pattern of connectivity, chain motifs, have a much stronger impact on dominant eigenmodes than other pairwise motifs. An overrepresentation of chain motifs induces a strong positive eigenvalue in inhibition-dominated networks and generates a potential instability that requires revisiting the classical excitation-inhibition balance criteria. Examining effects of external inputs, we show that chain motifs can on their own induce paradoxical responses where an increased input to inhibitory neurons leads to a decrease in their activity due to the recurrent feedback. These findings have direct implications for the interpretation of experiments in which responses to optogenetic perturbations are measured and used to infer the dynamical regime of cortical circuits.	[ { "version": "v1", "created": "Mon, 11 Nov 2024 08:57:44 GMT" }, { "version": "v2", "created": "Wed, 27 Nov 2024 13:27:46 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 13:59:15 GMT" } ]	2025-03-18T00:00:00	[ [ "Shao", "Yuxiu", "", "1 and 2" ], [ "Dahmen", "David", "", "5 and 6" ], [ "Recanatesi", "Stefano", "", "5 and 6" ], [ "Shea-Brown", "Eric", "", "5 and 6" ], [ "Ostojic", "Srdjan", "" ] ]	TITLE: Identifying the impact of local connectivity patterns on dynamics in excitatory-inhibitory networks ABSTRACT: Networks of excitatory and inhibitory (EI) neurons form a canonical circuit in the brain. Seminal theoretical results on dynamics of such networks are based on the assumption that synaptic strengths depend on the type of neurons they connect, but are otherwise statistically independent. Recent synaptic physiology datasets however highlight the prominence of specific connectivity patterns that go well beyond what is expected from independent connections. While decades of influential research have demonstrated the strong role of the basic EI cell type structure, to which extent additional connectivity features influence dynamics remains to be fully determined. Here we examine the effects of pairwise connectivity motifs on the linear dynamics in EI networks using an analytical framework that approximates the connectivity in terms of low-rank structures. This low-rank approximation is based on a mathematical derivation of the dominant eigenvalues of the connectivity matrix and predicts the impact on responses to external inputs of connectivity motifs and their interactions with cell-type structure. Our results reveal that a particular pattern of connectivity, chain motifs, have a much stronger impact on dominant eigenmodes than other pairwise motifs. An overrepresentation of chain motifs induces a strong positive eigenvalue in inhibition-dominated networks and generates a potential instability that requires revisiting the classical excitation-inhibition balance criteria. Examining effects of external inputs, we show that chain motifs can on their own induce paradoxical responses where an increased input to inhibitory neurons leads to a decrease in their activity due to the recurrent feedback. These findings have direct implications for the interpretation of experiments in which responses to optogenetic perturbations are measured and used to infer the dynamical regime of cortical circuits.
2411.06976	He Huang	He Huang and Wenjie Huang and Qi Yang and Yiling Xu and Zhu li	A Hierarchical Compression Technique for 3D Gaussian Splatting Compression	null	null	null	null	cs.CV cs.MM	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	3D Gaussian Splatting (GS) demonstrates excellent rendering quality and generation speed in novel view synthesis. However, substantial data size poses challenges for storage and transmission, making 3D GS compression an essential technology. Current 3D GS compression research primarily focuses on developing more compact scene representations, such as converting explicit 3D GS data into implicit forms. In contrast, compression of the GS data itself has hardly been explored. To address this gap, we propose a Hierarchical GS Compression (HGSC) technique. Initially, we prune unimportant Gaussians based on importance scores derived from both global and local significance, effectively reducing redundancy while maintaining visual quality. An Octree structure is used to compress 3D positions. Based on the 3D GS Octree, we implement a hierarchical attribute compression strategy by employing a KD-tree to partition the 3D GS into multiple blocks. We apply farthest point sampling to select anchor primitives within each block and others as non-anchor primitives with varying Levels of Details (LoDs). Anchor primitives serve as reference points for predicting non-anchor primitives across different LoDs to reduce spatial redundancy. For anchor primitives, we use the region adaptive hierarchical transform to achieve near-lossless compression of various attributes. For non-anchor primitives, each is predicted based on the k-nearest anchor primitives. To further minimize prediction errors, the reconstructed LoD and anchor primitives are combined to form new anchor primitives to predict the next LoD. Our method notably achieves superior compression quality and a significant data size reduction of over 4.5 times compared to the state-of-the-art compression method on small scenes datasets.	[ { "version": "v1", "created": "Mon, 11 Nov 2024 13:34:24 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 12:12:03 GMT" } ]	2025-03-18T00:00:00	[ [ "Huang", "He", "" ], [ "Huang", "Wenjie", "" ], [ "Yang", "Qi", "" ], [ "Xu", "Yiling", "" ], [ "li", "Zhu", "" ] ]	TITLE: A Hierarchical Compression Technique for 3D Gaussian Splatting Compression ABSTRACT: 3D Gaussian Splatting (GS) demonstrates excellent rendering quality and generation speed in novel view synthesis. However, substantial data size poses challenges for storage and transmission, making 3D GS compression an essential technology. Current 3D GS compression research primarily focuses on developing more compact scene representations, such as converting explicit 3D GS data into implicit forms. In contrast, compression of the GS data itself has hardly been explored. To address this gap, we propose a Hierarchical GS Compression (HGSC) technique. Initially, we prune unimportant Gaussians based on importance scores derived from both global and local significance, effectively reducing redundancy while maintaining visual quality. An Octree structure is used to compress 3D positions. Based on the 3D GS Octree, we implement a hierarchical attribute compression strategy by employing a KD-tree to partition the 3D GS into multiple blocks. We apply farthest point sampling to select anchor primitives within each block and others as non-anchor primitives with varying Levels of Details (LoDs). Anchor primitives serve as reference points for predicting non-anchor primitives across different LoDs to reduce spatial redundancy. For anchor primitives, we use the region adaptive hierarchical transform to achieve near-lossless compression of various attributes. For non-anchor primitives, each is predicted based on the k-nearest anchor primitives. To further minimize prediction errors, the reconstructed LoD and anchor primitives are combined to form new anchor primitives to predict the next LoD. Our method notably achieves superior compression quality and a significant data size reduction of over 4.5 times compared to the state-of-the-art compression method on small scenes datasets.
2411.07107	Brian DuSell	Alexandra Butoi and Ghazal Khalighinejad and Anej Svete and Josef Valvoda and Ryan Cotterell and Brian DuSell	Training Neural Networks as Recognizers of Formal Languages	44 pages, 3 figures. ICLR 2025	null	null	null	cs.CL cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Characterizing the computational power of neural network architectures in terms of formal language theory remains a crucial line of research, as it describes lower and upper bounds on the reasoning capabilities of modern AI. However, when empirically testing these bounds, existing work often leaves a discrepancy between experiments and the formal claims they are meant to support. The problem is that formal language theory pertains specifically to recognizers: machines that receive a string as input and classify whether it belongs to a language. On the other hand, it is common instead to evaluate language models on proxy tasks, e.g., language modeling or sequence-to-sequence transduction, that are similar in only an informal sense to the underlying theory. We correct this mismatch by training and evaluating neural networks directly as binary classifiers of strings, using a general method that can be applied to a wide variety of languages. As part of this, we extend an algorithm recently proposed by Sn{\ae}bjarnarson et al. (2025) for efficient length-controlled sampling of strings from regular languages. We provide results on a variety of languages across the Chomsky hierarchy for three neural architectures: a simple RNN, an LSTM, and a causally-masked transformer. We find that the RNN and LSTM often outperform the transformer, and that auxiliary training objectives such as language modeling can help, although no single objective uniformly improves performance across languages and architectures. Our contributions will facilitate theoretically sound empirical testing of language recognition claims in future work. We have released our datasets as a benchmark called FLaRe (Formal Language Recognition), along with our code.	[ { "version": "v1", "created": "Mon, 11 Nov 2024 16:33:25 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 14:51:27 GMT" } ]	2025-03-18T00:00:00	[ [ "Butoi", "Alexandra", "" ], [ "Khalighinejad", "Ghazal", "" ], [ "Svete", "Anej", "" ], [ "Valvoda", "Josef", "" ], [ "Cotterell", "Ryan", "" ], [ "DuSell", "Brian", "" ] ]	TITLE: Training Neural Networks as Recognizers of Formal Languages ABSTRACT: Characterizing the computational power of neural network architectures in terms of formal language theory remains a crucial line of research, as it describes lower and upper bounds on the reasoning capabilities of modern AI. However, when empirically testing these bounds, existing work often leaves a discrepancy between experiments and the formal claims they are meant to support. The problem is that formal language theory pertains specifically to recognizers: machines that receive a string as input and classify whether it belongs to a language. On the other hand, it is common instead to evaluate language models on proxy tasks, e.g., language modeling or sequence-to-sequence transduction, that are similar in only an informal sense to the underlying theory. We correct this mismatch by training and evaluating neural networks directly as binary classifiers of strings, using a general method that can be applied to a wide variety of languages. As part of this, we extend an algorithm recently proposed by Sn{\ae}bjarnarson et al. (2025) for efficient length-controlled sampling of strings from regular languages. We provide results on a variety of languages across the Chomsky hierarchy for three neural architectures: a simple RNN, an LSTM, and a causally-masked transformer. We find that the RNN and LSTM often outperform the transformer, and that auxiliary training objectives such as language modeling can help, although no single objective uniformly improves performance across languages and architectures. Our contributions will facilitate theoretically sound empirical testing of language recognition claims in future work. We have released our datasets as a benchmark called FLaRe (Formal Language Recognition), along with our code.
2411.07758	Wen Dongcheng	Ran Lingyan, Wen Dongcheng, Zhuo Tao, Zhang Shizhou, Zhang Xiuwei and Zhang Yanning	AdaSemiCD: An Adaptive Semi-Supervised Change Detection Method Based on Pseudo-Label Evaluation	Accepted by IEEE Transactions on Geoscience and Remote Sensing(TGRS)	null	10.1109/TGRS.2025.3551504	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Change Detection (CD) is an essential field in remote sensing, with a primary focus on identifying areas of change in bi-temporal image pairs captured at varying intervals of the same region by a satellite. The data annotation process for the CD task is both time-consuming and labor-intensive. To make better use of the scarce labeled data and abundant unlabeled data, we present an adaptive dynamic semi-supervised learning method, AdaSemiCD, to improve the use of pseudo-labels and optimize the training process. Initially, due to the extreme class imbalance inherent in CD, the model is more inclined to focus on the background class, and it is easy to confuse the boundary of the target object. Considering these two points, we develop a measurable evaluation metric for pseudo-labels that enhances the representation of information entropy by class rebalancing and amplification of confusing areas to give a larger weight to prospects change objects. Subsequently, to enhance the reliability of sample-wise pseudo-labels, we introduce the AdaFusion module, which is capable of dynamically identifying the most uncertain region and substituting it with more trustworthy content. Lastly, to ensure better training stability, we introduce the AdaEMA module, which updates the teacher model using only batches of trusted samples. Experimental results from LEVIR-CD, WHU-CD, and CDD datasets validate the efficacy and universality of our proposed adaptive training framework.	[ { "version": "v1", "created": "Tue, 12 Nov 2024 12:35:34 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 07:28:26 GMT" } ]	2025-03-18T00:00:00	[ [ "Lingyan", "Ran", "" ], [ "Dongcheng", "Wen", "" ], [ "Tao", "Zhuo", "" ], [ "Shizhou", "Zhang", "" ], [ "Xiuwei", "Zhang", "" ], [ "Yanning", "Zhang", "" ] ]	TITLE: AdaSemiCD: An Adaptive Semi-Supervised Change Detection Method Based on Pseudo-Label Evaluation ABSTRACT: Change Detection (CD) is an essential field in remote sensing, with a primary focus on identifying areas of change in bi-temporal image pairs captured at varying intervals of the same region by a satellite. The data annotation process for the CD task is both time-consuming and labor-intensive. To make better use of the scarce labeled data and abundant unlabeled data, we present an adaptive dynamic semi-supervised learning method, AdaSemiCD, to improve the use of pseudo-labels and optimize the training process. Initially, due to the extreme class imbalance inherent in CD, the model is more inclined to focus on the background class, and it is easy to confuse the boundary of the target object. Considering these two points, we develop a measurable evaluation metric for pseudo-labels that enhances the representation of information entropy by class rebalancing and amplification of confusing areas to give a larger weight to prospects change objects. Subsequently, to enhance the reliability of sample-wise pseudo-labels, we introduce the AdaFusion module, which is capable of dynamically identifying the most uncertain region and substituting it with more trustworthy content. Lastly, to ensure better training stability, we introduce the AdaEMA module, which updates the teacher model using only batches of trusted samples. Experimental results from LEVIR-CD, WHU-CD, and CDD datasets validate the efficacy and universality of our proposed adaptive training framework.
2411.09145	ChengBo Yuan	Chengbo Yuan, Geng Chen, Li Yi, Yang Gao	Self-Supervised Monocular 4D Scene Reconstruction for Egocentric Videos	null	null	null	null	cs.CV cs.RO	http://creativecommons.org/licenses/by-sa/4.0/	Egocentric videos provide valuable insights into human interactions with the physical world, which has sparked growing interest in the computer vision and robotics communities. A critical challenge in fully understanding the geometry and dynamics of egocentric videos is dense scene reconstruction. However, the lack of high-quality labeled datasets in this field has hindered the effectiveness of current supervised learning methods. In this work, we aim to address this issue by exploring an self-supervised dynamic scene reconstruction approach. We introduce EgoMono4D, a novel model that unifies the estimation of multiple variables necessary for Egocentric Monocular 4D reconstruction, including camera intrinsic, camera poses, and video depth, all within a fast feed-forward framework. Starting from pretrained single-frame depth and intrinsic estimation model, we extend it with camera poses estimation and align multi-frame results on large-scale unlabeled egocentric videos. We evaluate EgoMono4D in both in-domain and zero-shot generalization settings, achieving superior performance in dense pointclouds sequence reconstruction compared to all baselines. EgoMono4D represents the first attempt to apply self-supervised learning for pointclouds sequence reconstruction to the label-scarce egocentric field, enabling fast, dense, and generalizable reconstruction. The interactable visualization, code and trained models are released https://egomono4d.github.io/	[ { "version": "v1", "created": "Thu, 14 Nov 2024 02:57:11 GMT" }, { "version": "v2", "created": "Fri, 15 Nov 2024 12:27:39 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 15:05:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Yuan", "Chengbo", "" ], [ "Chen", "Geng", "" ], [ "Yi", "Li", "" ], [ "Gao", "Yang", "" ] ]	TITLE: Self-Supervised Monocular 4D Scene Reconstruction for Egocentric Videos ABSTRACT: Egocentric videos provide valuable insights into human interactions with the physical world, which has sparked growing interest in the computer vision and robotics communities. A critical challenge in fully understanding the geometry and dynamics of egocentric videos is dense scene reconstruction. However, the lack of high-quality labeled datasets in this field has hindered the effectiveness of current supervised learning methods. In this work, we aim to address this issue by exploring an self-supervised dynamic scene reconstruction approach. We introduce EgoMono4D, a novel model that unifies the estimation of multiple variables necessary for Egocentric Monocular 4D reconstruction, including camera intrinsic, camera poses, and video depth, all within a fast feed-forward framework. Starting from pretrained single-frame depth and intrinsic estimation model, we extend it with camera poses estimation and align multi-frame results on large-scale unlabeled egocentric videos. We evaluate EgoMono4D in both in-domain and zero-shot generalization settings, achieving superior performance in dense pointclouds sequence reconstruction compared to all baselines. EgoMono4D represents the first attempt to apply self-supervised learning for pointclouds sequence reconstruction to the label-scarce egocentric field, enabling fast, dense, and generalizable reconstruction. The interactable visualization, code and trained models are released https://egomono4d.github.io/
2411.10962	Lei Yang	Lei Yang, Xinyu Zhang, Chen Wang, Jun Li, Jiaqi Ma, Zhiying Song, Tong Zhao, Ziying Song, Li Wang, Mo Zhou, Yang Shen, Kai Wu, Chen Lv	V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception	15 pages, 9 figures	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Modern autonomous vehicle perception systems often struggle with occlusions and limited perception range. Previous studies have demonstrated the effectiveness of cooperative perception in extending the perception range and overcoming occlusions, thereby enhancing the safety of autonomous driving. In recent years, a series of cooperative perception datasets have emerged; however, these datasets primarily focus on cameras and LiDAR, neglecting 4D Radar, a sensor used in single-vehicle autonomous driving to provide robust perception in adverse weather conditions. In this paper, to bridge the gap created by the absence of 4D Radar datasets in cooperative perception, we present V2X-Radar, the first large-scale, real-world multi-modal dataset featuring 4D Radar. V2X-Radar dataset is collected using a connected vehicle platform and an intelligent roadside unit equipped with 4D Radar, LiDAR, and multi-view cameras. The collected data encompasses sunny and rainy weather conditions, spanning daytime, dusk, and nighttime, as well as various typical challenging scenarios. The dataset consists of 20K LiDAR frames, 40K camera images, and 20K 4D Radar data, including 350K annotated boxes across five categories. To support various research domains, we have established V2X-Radar-C for cooperative perception, V2X-Radar-I for roadside perception, and V2X-Radar-V for single-vehicle perception. Furthermore, we provide comprehensive benchmarks across these three sub-datasets. We will release all datasets and benchmark codebase at http://openmpd.com/column/V2X-Radar and https://github.com/yanglei18/V2X-Radar.	[ { "version": "v1", "created": "Sun, 17 Nov 2024 04:59:00 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 02:06:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Yang", "Lei", "" ], [ "Zhang", "Xinyu", "" ], [ "Wang", "Chen", "" ], [ "Li", "Jun", "" ], [ "Ma", "Jiaqi", "" ], [ "Song", "Zhiying", "" ], [ "Zhao", "Tong", "" ], [ "Song", "Ziying", "" ], [ "Wang", "Li", "" ], [ "Zhou", "Mo", "" ], [ "Shen", "Yang", "" ], [ "Wu", "Kai", "" ], [ "Lv", "Chen", "" ] ]	TITLE: V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception ABSTRACT: Modern autonomous vehicle perception systems often struggle with occlusions and limited perception range. Previous studies have demonstrated the effectiveness of cooperative perception in extending the perception range and overcoming occlusions, thereby enhancing the safety of autonomous driving. In recent years, a series of cooperative perception datasets have emerged; however, these datasets primarily focus on cameras and LiDAR, neglecting 4D Radar, a sensor used in single-vehicle autonomous driving to provide robust perception in adverse weather conditions. In this paper, to bridge the gap created by the absence of 4D Radar datasets in cooperative perception, we present V2X-Radar, the first large-scale, real-world multi-modal dataset featuring 4D Radar. V2X-Radar dataset is collected using a connected vehicle platform and an intelligent roadside unit equipped with 4D Radar, LiDAR, and multi-view cameras. The collected data encompasses sunny and rainy weather conditions, spanning daytime, dusk, and nighttime, as well as various typical challenging scenarios. The dataset consists of 20K LiDAR frames, 40K camera images, and 20K 4D Radar data, including 350K annotated boxes across five categories. To support various research domains, we have established V2X-Radar-C for cooperative perception, V2X-Radar-I for roadside perception, and V2X-Radar-V for single-vehicle perception. Furthermore, we provide comprehensive benchmarks across these three sub-datasets. We will release all datasets and benchmark codebase at http://openmpd.com/column/V2X-Radar and https://github.com/yanglei18/V2X-Radar.
2411.11886	Lu Wang-NÃ¶th	Lu Wang-N\"oth, Philipp Heiler, Hai Huang, Daniel Lichtenstern, Alexandra Reichenbach, Luis Flacke, Linus Maisch, Helmut Mayer	How Much Data is Enough? Optimization of Data Collection for Artifact Detection in EEG Recordings	Several changes of wording. Caption of figure 10 corrected	null	10.1088/1741-2552/adbebe	null	eess.SP cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Objective. Electroencephalography (EEG) is a widely used neuroimaging technique known for its cost-effectiveness and user-friendliness. However, various artifacts, particularly biological artifacts like Electromyography (EMG) signals, lead to a poor signal-to-noise ratio, limiting the precision of analyses and applications. The currently reported EEG data cleaning performance largely depends on the data used for validation, and in the case of machine learning approaches, also on the data used for training. The data are typically gathered either by recruiting subjects to perform specific artifact tasks or by integrating existing datasets. Prevailing approaches, however, tend to rely on intuitive, concept-oriented data collection with minimal justification for the selection of artifacts and their quantities. Given the substantial costs associated with biological data collection and the pressing need for effective data utilization, we propose an optimization procedure for data-oriented data collection design using deep learning-based artifact detection. Approach. We apply a binary classification between artifact epochs (time intervals containing artifacts) and non-artifact epochs (time intervals containing no artifact) using three different neural architectures. Our aim is to minimize data collection efforts while preserving the cleaning efficiency. Main results. We were able to reduce the number of artifact tasks from twelve to three and decrease repetitions of isometric contraction tasks from ten to three or sometimes even just one. Significance. Our work addresses the need for effective data utilization in biological data collection, offering a systematic and dynamic quantitative approach. By providing clear justifications for the choices of artifacts and their quantity, we aim to guide future studies toward more effective and economical data collection in EEG and EMG research.	[ { "version": "v1", "created": "Tue, 5 Nov 2024 11:47:59 GMT" }, { "version": "v2", "created": "Wed, 20 Nov 2024 10:38:55 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang-Nöth", "Lu", "" ], [ "Heiler", "Philipp", "" ], [ "Huang", "Hai", "" ], [ "Lichtenstern", "Daniel", "" ], [ "Reichenbach", "Alexandra", "" ], [ "Flacke", "Luis", "" ], [ "Maisch", "Linus", "" ], [ "Mayer", "Helmut", "" ] ]	TITLE: How Much Data is Enough? Optimization of Data Collection for Artifact Detection in EEG Recordings ABSTRACT: Objective. Electroencephalography (EEG) is a widely used neuroimaging technique known for its cost-effectiveness and user-friendliness. However, various artifacts, particularly biological artifacts like Electromyography (EMG) signals, lead to a poor signal-to-noise ratio, limiting the precision of analyses and applications. The currently reported EEG data cleaning performance largely depends on the data used for validation, and in the case of machine learning approaches, also on the data used for training. The data are typically gathered either by recruiting subjects to perform specific artifact tasks or by integrating existing datasets. Prevailing approaches, however, tend to rely on intuitive, concept-oriented data collection with minimal justification for the selection of artifacts and their quantities. Given the substantial costs associated with biological data collection and the pressing need for effective data utilization, we propose an optimization procedure for data-oriented data collection design using deep learning-based artifact detection. Approach. We apply a binary classification between artifact epochs (time intervals containing artifacts) and non-artifact epochs (time intervals containing no artifact) using three different neural architectures. Our aim is to minimize data collection efforts while preserving the cleaning efficiency. Main results. We were able to reduce the number of artifact tasks from twelve to three and decrease repetitions of isometric contraction tasks from ten to three or sometimes even just one. Significance. Our work addresses the need for effective data utilization in biological data collection, offering a systematic and dynamic quantitative approach. By providing clear justifications for the choices of artifacts and their quantity, we aim to guide future studies toward more effective and economical data collection in EEG and EMG research.
2411.13376	Ricardo Monta\~nana G\'omez	Ricardo Monta\~nana and Jos\'e A. G\'amez and Jos\'e M. Puerta	ODTE -- An ensemble of multi-class SVM-based oblique decision trees	Accepted version	Ricardo Monta\~nana, Jos\'e A. G\'amez, Jos\'e M. Puerta(2025). ODTE-An ensemble of multi-class SVM-based oblique decision trees. Expert Systems with Applications 273:126833	10.1016/j.eswa.2025.126833	null	cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We propose ODTE, a new ensemble that uses oblique decision trees as base classifiers. Additionally, we introduce STree, the base algorithm for growing oblique decision trees, which leverages support vector machines to define hyperplanes within the decision nodes. We embed a multiclass strategy -- one-vs-one or one-vs-rest -- at the decision nodes, allowing the model to directly handle non-binary classification tasks without the need to cluster instances into two groups, as is common in other approaches from the literature. In each decision node, only the best-performing model SVM -- the one that minimizes an impurity measure for the n-ary classification -- is retained, even if the learned SVM addresses a binary classification subtask. An extensive experimental study involving 49 datasets and various state-of-the-art algorithms for oblique decision tree ensembles has been conducted. Our results show that ODTE ranks consistently above its competitors, achieving significant performance gains when hyperparameters are carefully tuned. Moreover, the oblique decision trees learned through STree are more compact than those produced by other algorithms evaluated in our experiments.	[ { "version": "v1", "created": "Wed, 20 Nov 2024 14:58:32 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 11:34:32 GMT" } ]	2025-03-18T00:00:00	[ [ "Montañana", "Ricardo", "" ], [ "Gámez", "José A.", "" ], [ "Puerta", "José M.", "" ] ]	TITLE: ODTE -- An ensemble of multi-class SVM-based oblique decision trees ABSTRACT: We propose ODTE, a new ensemble that uses oblique decision trees as base classifiers. Additionally, we introduce STree, the base algorithm for growing oblique decision trees, which leverages support vector machines to define hyperplanes within the decision nodes. We embed a multiclass strategy -- one-vs-one or one-vs-rest -- at the decision nodes, allowing the model to directly handle non-binary classification tasks without the need to cluster instances into two groups, as is common in other approaches from the literature. In each decision node, only the best-performing model SVM -- the one that minimizes an impurity measure for the n-ary classification -- is retained, even if the learned SVM addresses a binary classification subtask. An extensive experimental study involving 49 datasets and various state-of-the-art algorithms for oblique decision tree ensembles has been conducted. Our results show that ODTE ranks consistently above its competitors, achieving significant performance gains when hyperparameters are carefully tuned. Moreover, the oblique decision trees learned through STree are more compact than those produced by other algorithms evaluated in our experiments.
2411.16106	Xingyu Liu	Xingyu Liu, Gu Wang, Ruida Zhang, Chenyangguang Zhang, Federico Tombari, Xiangyang Ji	UNOPose: Unseen Object Pose Estimation with an Unposed RGB-D Reference Image	Accepted by CVPR'25	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Unseen object pose estimation methods often rely on CAD models or multiple reference views, making the onboarding stage costly. To simplify reference acquisition, we aim to estimate the unseen object's pose through a single unposed RGB-D reference image. While previous works leverage reference images as pose anchors to limit the range of relative pose, our scenario presents significant challenges since the relative transformation could vary across the entire SE(3) space. Moreover, factors like occlusion, sensor noise, and extreme geometry could result in low viewpoint overlap. To address these challenges, we present a novel approach and benchmark, termed UNOPose, for unseen one-reference-based object pose estimation. Building upon a coarse-to-fine paradigm, UNOPose constructs an SE(3)-invariant reference frame to standardize object representation despite pose and size variations. To alleviate small overlap across viewpoints, we recalibrate the weight of each correspondence based on its predicted likelihood of being within the overlapping region. Evaluated on our proposed benchmark based on the BOP Challenge, UNOPose demonstrates superior performance, significantly outperforming traditional and learning-based methods in the one-reference setting and remaining competitive with CAD-model-based methods. The code and dataset are available at https://github.com/shanice-l/UNOPose.	[ { "version": "v1", "created": "Mon, 25 Nov 2024 05:36:00 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 07:54:18 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Xingyu", "" ], [ "Wang", "Gu", "" ], [ "Zhang", "Ruida", "" ], [ "Zhang", "Chenyangguang", "" ], [ "Tombari", "Federico", "" ], [ "Ji", "Xiangyang", "" ] ]	TITLE: UNOPose: Unseen Object Pose Estimation with an Unposed RGB-D Reference Image ABSTRACT: Unseen object pose estimation methods often rely on CAD models or multiple reference views, making the onboarding stage costly. To simplify reference acquisition, we aim to estimate the unseen object's pose through a single unposed RGB-D reference image. While previous works leverage reference images as pose anchors to limit the range of relative pose, our scenario presents significant challenges since the relative transformation could vary across the entire SE(3) space. Moreover, factors like occlusion, sensor noise, and extreme geometry could result in low viewpoint overlap. To address these challenges, we present a novel approach and benchmark, termed UNOPose, for unseen one-reference-based object pose estimation. Building upon a coarse-to-fine paradigm, UNOPose constructs an SE(3)-invariant reference frame to standardize object representation despite pose and size variations. To alleviate small overlap across viewpoints, we recalibrate the weight of each correspondence based on its predicted likelihood of being within the overlapping region. Evaluated on our proposed benchmark based on the BOP Challenge, UNOPose demonstrates superior performance, significantly outperforming traditional and learning-based methods in the one-reference setting and remaining competitive with CAD-model-based methods. The code and dataset are available at https://github.com/shanice-l/UNOPose.
2411.16446	Changjian Li	Jiawei Wang, Zhiming Cui, Changjian Li	VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation	null	null	null	null	cs.CV cs.GR	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper presents VQ-SGen, a novel algorithm for high-quality creative sketch generation. Recent approaches have framed the task as pixel-based generation either as a whole or part-by-part, neglecting the intrinsic and contextual relationships among individual strokes, such as the shape and spatial positioning of both proximal and distant strokes. To overcome these limitations, we propose treating each stroke within a sketch as an entity and introducing a vector-quantized (VQ) stroke representation for fine-grained sketch generation. Our method follows a two-stage framework - in stage one, we decouple each stroke's shape and location information to ensure the VQ representation prioritizes stroke shape learning. In stage two, we feed the precise and compact representation into an auto-decoding Transformer to incorporate stroke semantics, positions, and shapes into the generation process. By utilizing tokenized stroke representation, our approach generates strokes with high fidelity and facilitates novel applications, such as text or class label conditioned generation and sketch completion. Comprehensive experiments demonstrate our method surpasses existing state-of-the-art techniques on the CreativeSketch dataset, underscoring its effectiveness.	[ { "version": "v1", "created": "Mon, 25 Nov 2024 14:51:22 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 09:19:45 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Jiawei", "" ], [ "Cui", "Zhiming", "" ], [ "Li", "Changjian", "" ] ]	TITLE: VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation ABSTRACT: This paper presents VQ-SGen, a novel algorithm for high-quality creative sketch generation. Recent approaches have framed the task as pixel-based generation either as a whole or part-by-part, neglecting the intrinsic and contextual relationships among individual strokes, such as the shape and spatial positioning of both proximal and distant strokes. To overcome these limitations, we propose treating each stroke within a sketch as an entity and introducing a vector-quantized (VQ) stroke representation for fine-grained sketch generation. Our method follows a two-stage framework - in stage one, we decouple each stroke's shape and location information to ensure the VQ representation prioritizes stroke shape learning. In stage two, we feed the precise and compact representation into an auto-decoding Transformer to incorporate stroke semantics, positions, and shapes into the generation process. By utilizing tokenized stroke representation, our approach generates strokes with high fidelity and facilitates novel applications, such as text or class label conditioned generation and sketch completion. Comprehensive experiments demonstrate our method surpasses existing state-of-the-art techniques on the CreativeSketch dataset, underscoring its effectiveness.
2411.17386	Bastian Wittmann	Bastian Wittmann, Yannick Wattenberg, Tamaz Amiranashvili, Suprosanna Shit, Bjoern Menze	vesselFM: A Foundation Model for Universal 3D Blood Vessel Segmentation	null	null	null	null	eess.IV cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Segmenting 3D blood vessels is a critical yet challenging task in medical image analysis. This is due to significant imaging modality-specific variations in artifacts, vascular patterns and scales, signal-to-noise ratios, and background tissues. These variations, along with domain gaps arising from varying imaging protocols, limit the generalization of existing supervised learning-based methods, requiring tedious voxel-level annotations for each dataset separately. While foundation models promise to alleviate this limitation, they typically fail to generalize to the task of blood vessel segmentation, posing a unique, complex problem. In this work, we present vesselFM, a foundation model designed specifically for the broad task of 3D blood vessel segmentation. Unlike previous models, vesselFM can effortlessly generalize to unseen domains. To achieve zero-shot generalization, we train vesselFM on three heterogeneous data sources: a large, curated annotated dataset, data generated by a domain randomization scheme, and data sampled from a flow matching-based generative model. Extensive evaluations show that vesselFM outperforms state-of-the-art medical image segmentation foundation models across four (pre-)clinically relevant imaging modalities in zero-, one-, and few-shot scenarios, therefore providing a universal solution for 3D blood vessel segmentation.	[ { "version": "v1", "created": "Tue, 26 Nov 2024 12:44:42 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 18:56:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Wittmann", "Bastian", "" ], [ "Wattenberg", "Yannick", "" ], [ "Amiranashvili", "Tamaz", "" ], [ "Shit", "Suprosanna", "" ], [ "Menze", "Bjoern", "" ] ]	TITLE: vesselFM: A Foundation Model for Universal 3D Blood Vessel Segmentation ABSTRACT: Segmenting 3D blood vessels is a critical yet challenging task in medical image analysis. This is due to significant imaging modality-specific variations in artifacts, vascular patterns and scales, signal-to-noise ratios, and background tissues. These variations, along with domain gaps arising from varying imaging protocols, limit the generalization of existing supervised learning-based methods, requiring tedious voxel-level annotations for each dataset separately. While foundation models promise to alleviate this limitation, they typically fail to generalize to the task of blood vessel segmentation, posing a unique, complex problem. In this work, we present vesselFM, a foundation model designed specifically for the broad task of 3D blood vessel segmentation. Unlike previous models, vesselFM can effortlessly generalize to unseen domains. To achieve zero-shot generalization, we train vesselFM on three heterogeneous data sources: a large, curated annotated dataset, data generated by a domain randomization scheme, and data sampled from a flow matching-based generative model. Extensive evaluations show that vesselFM outperforms state-of-the-art medical image segmentation foundation models across four (pre-)clinically relevant imaging modalities in zero-, one-, and few-shot scenarios, therefore providing a universal solution for 3D blood vessel segmentation.
2411.17388	Haoyu Huang	Haoyu Huang, Chong Chen, Conghui He, Yang Li, Jiawei Jiang, Wentao Zhang	Can LLMs be Good Graph Judger for Knowledge Graph Construction?	null	null	null	null	cs.CL cs.AI	http://creativecommons.org/licenses/by/4.0/	In real-world scenarios, most of the data obtained from information retrieval (IR) system is unstructured. Converting natural language sentences into structured Knowledge Graphs (KGs) remains a critical challenge. The quality of constructed KGs may also impact the performance of some KG-dependent domains like GraphRAG systems and recommendation systems. Recently, Large Language Models (LLMs) have demonstrated impressive capabilities in addressing a wide range of natural language processing tasks. However, there are still challenges when utilizing LLMs to address the task of generating structured KGs. And we have identified three limitations with respect to existing KG construction methods. (1)There is a large amount of information and excessive noise in real-world documents, which could result in extracting messy information. (2)Native LLMs struggle to effectively extract accuracy knowledge from some domain-specific documents. (3)Hallucinations phenomenon cannot be overlooked when utilizing LLMs directly as an unsupervised method for constructing KGs. In this paper, we propose GraphJudger, a knowledge graph construction framework to address the aforementioned challenges. We introduce three innovative modules in our method, which are entity-centric iterative text denoising, knowledge aware instruction tuning and graph judgement, respectively. We seek to utilize the capacity of LLMs to function as a graph judger, a capability superior to their role only as a predictor for KG construction problems. Experiments conducted on two general text-graph pair datasets and one domain-specific text-graph pair dataset show superior performances compared to baseline methods. The code of our proposed method is available at https://github.com/hhy-huang/GraphJudger.	[ { "version": "v1", "created": "Tue, 26 Nov 2024 12:46:57 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 11:49:45 GMT" } ]	2025-03-18T00:00:00	[ [ "Huang", "Haoyu", "" ], [ "Chen", "Chong", "" ], [ "He", "Conghui", "" ], [ "Li", "Yang", "" ], [ "Jiang", "Jiawei", "" ], [ "Zhang", "Wentao", "" ] ]	TITLE: Can LLMs be Good Graph Judger for Knowledge Graph Construction? ABSTRACT: In real-world scenarios, most of the data obtained from information retrieval (IR) system is unstructured. Converting natural language sentences into structured Knowledge Graphs (KGs) remains a critical challenge. The quality of constructed KGs may also impact the performance of some KG-dependent domains like GraphRAG systems and recommendation systems. Recently, Large Language Models (LLMs) have demonstrated impressive capabilities in addressing a wide range of natural language processing tasks. However, there are still challenges when utilizing LLMs to address the task of generating structured KGs. And we have identified three limitations with respect to existing KG construction methods. (1)There is a large amount of information and excessive noise in real-world documents, which could result in extracting messy information. (2)Native LLMs struggle to effectively extract accuracy knowledge from some domain-specific documents. (3)Hallucinations phenomenon cannot be overlooked when utilizing LLMs directly as an unsupervised method for constructing KGs. In this paper, we propose GraphJudger, a knowledge graph construction framework to address the aforementioned challenges. We introduce three innovative modules in our method, which are entity-centric iterative text denoising, knowledge aware instruction tuning and graph judgement, respectively. We seek to utilize the capacity of LLMs to function as a graph judger, a capability superior to their role only as a predictor for KG construction problems. Experiments conducted on two general text-graph pair datasets and one domain-specific text-graph pair dataset show superior performances compared to baseline methods. The code of our proposed method is available at https://github.com/hhy-huang/GraphJudger.
2411.17698	Ziyang Chen	Ziyang Chen, Prem Seetharaman, Bryan Russell, Oriol Nieto, David Bourgin, Andrew Owens, Justin Salamon	Video-Guided Foley Sound Generation with Multimodal Controls	Accepted at CVPR 2025. Project site: https://ificl.github.io/MultiFoley/	null	null	null	cs.CV cs.MM cs.SD eess.AS	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Generating sound effects for videos often requires creating artistic sound effects that diverge significantly from real-life sources and flexible control in the sound design. To address this problem, we introduce MultiFoley, a model designed for video-guided sound generation that supports multimodal conditioning through text, audio, and video. Given a silent video and a text prompt, MultiFoley allows users to create clean sounds (e.g., skateboard wheels spinning without wind noise) or more whimsical sounds (e.g., making a lion's roar sound like a cat's meow). MultiFoley also allows users to choose reference audio from sound effects (SFX) libraries or partial videos for conditioning. A key novelty of our model lies in its joint training on both internet video datasets with low-quality audio and professional SFX recordings, enabling high-quality, full-bandwidth (48kHz) audio generation. Through automated evaluations and human studies, we demonstrate that MultiFoley successfully generates synchronized high-quality sounds across varied conditional inputs and outperforms existing methods. Please see our project page for video results: https://ificl.github.io/MultiFoley/	[ { "version": "v1", "created": "Tue, 26 Nov 2024 18:59:58 GMT" }, { "version": "v2", "created": "Thu, 28 Nov 2024 13:25:04 GMT" }, { "version": "v3", "created": "Wed, 22 Jan 2025 20:03:04 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 17:44:37 GMT" } ]	2025-03-18T00:00:00	[ [ "Chen", "Ziyang", "" ], [ "Seetharaman", "Prem", "" ], [ "Russell", "Bryan", "" ], [ "Nieto", "Oriol", "" ], [ "Bourgin", "David", "" ], [ "Owens", "Andrew", "" ], [ "Salamon", "Justin", "" ] ]	TITLE: Video-Guided Foley Sound Generation with Multimodal Controls ABSTRACT: Generating sound effects for videos often requires creating artistic sound effects that diverge significantly from real-life sources and flexible control in the sound design. To address this problem, we introduce MultiFoley, a model designed for video-guided sound generation that supports multimodal conditioning through text, audio, and video. Given a silent video and a text prompt, MultiFoley allows users to create clean sounds (e.g., skateboard wheels spinning without wind noise) or more whimsical sounds (e.g., making a lion's roar sound like a cat's meow). MultiFoley also allows users to choose reference audio from sound effects (SFX) libraries or partial videos for conditioning. A key novelty of our model lies in its joint training on both internet video datasets with low-quality audio and professional SFX recordings, enabling high-quality, full-bandwidth (48kHz) audio generation. Through automated evaluations and human studies, we demonstrate that MultiFoley successfully generates synchronized high-quality sounds across varied conditional inputs and outperforms existing methods. Please see our project page for video results: https://ificl.github.io/MultiFoley/
2411.18412	David Serrano-Lozano	David Serrano-Lozano, Luis Herranz, Shaolin Su and Javier Vazquez-Corral	Adaptive Blind All-in-One Image Restoration	17 pages	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Blind all-in-one image restoration models aim to recover a high-quality image from an input degraded with unknown distortions. However, these models require all the possible degradation types to be defined during the training stage while showing limited generalization to unseen degradations, which limits their practical application in complex cases. In this paper, we introduce ABAIR, a simple yet effective adaptive blind all-in-one restoration model that not only handles multiple degradations and generalizes well to unseen distortions but also efficiently integrates new degradations by training only a small subset of parameters. We first train our baseline model on a large dataset of natural images with multiple synthetic degradations. To enhance its ability to recognize distortions, we incorporate a segmentation head that estimates per-pixel degradation types. Second, we adapt our initial model to varying image restoration tasks using independent low-rank adapters. Third, we learn to adaptively combine adapters to versatile images via a flexible and lightweight degradation estimator. This specialize-then-merge approach is both powerful in addressing specific distortions and flexible in adapting to complex tasks. Moreover, our model not only surpasses state-of-the-art performance on five- and three-task IR setups but also demonstrates superior generalization to unseen degradations and composite distortions.	[ { "version": "v1", "created": "Wed, 27 Nov 2024 14:58:08 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 08:04:17 GMT" } ]	2025-03-18T00:00:00	[ [ "Serrano-Lozano", "David", "" ], [ "Herranz", "Luis", "" ], [ "Su", "Shaolin", "" ], [ "Vazquez-Corral", "Javier", "" ] ]	TITLE: Adaptive Blind All-in-One Image Restoration ABSTRACT: Blind all-in-one image restoration models aim to recover a high-quality image from an input degraded with unknown distortions. However, these models require all the possible degradation types to be defined during the training stage while showing limited generalization to unseen degradations, which limits their practical application in complex cases. In this paper, we introduce ABAIR, a simple yet effective adaptive blind all-in-one restoration model that not only handles multiple degradations and generalizes well to unseen distortions but also efficiently integrates new degradations by training only a small subset of parameters. We first train our baseline model on a large dataset of natural images with multiple synthetic degradations. To enhance its ability to recognize distortions, we incorporate a segmentation head that estimates per-pixel degradation types. Second, we adapt our initial model to varying image restoration tasks using independent low-rank adapters. Third, we learn to adaptively combine adapters to versatile images via a flexible and lightweight degradation estimator. This specialize-then-merge approach is both powerful in addressing specific distortions and flexible in adapting to complex tasks. Moreover, our model not only surpasses state-of-the-art performance on five- and three-task IR setups but also demonstrates superior generalization to unseen degradations and composite distortions.
2411.19921	Wenjia Wang	Wenjia Wang, Liang Pan, Zhiyang Dou, Jidong Mei, Zhouyingcheng Liao, Yuke Lou, Yifan Wu, Lei Yang, Jingbo Wang, Taku Komura	SIMS: Simulating Stylized Human-Scene Interactions with Retrieval-Augmented Script Generation	null	null	null	null	cs.CV cs.AI cs.CL cs.GR	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Simulating stylized human-scene interactions (HSI) in physical environments is a challenging yet fascinating task. Prior works emphasize long-term execution but fall short in achieving both diverse style and physical plausibility. To tackle this challenge, we introduce a novel hierarchical framework named SIMS that seamlessly bridges highlevel script-driven intent with a low-level control policy, enabling more expressive and diverse human-scene interactions. Specifically, we employ Large Language Models with Retrieval-Augmented Generation (RAG) to generate coherent and diverse long-form scripts, providing a rich foundation for motion planning. A versatile multicondition physics-based control policy is also developed, which leverages text embeddings from the generated scripts to encode stylistic cues, simultaneously perceiving environmental geometries and accomplishing task goals. By integrating the retrieval-augmented script generation with the multi-condition controller, our approach provides a unified solution for generating stylized HSI motions. We further introduce a comprehensive planning dataset produced by RAG and a stylized motion dataset featuring diverse locomotions and interactions. Extensive experiments demonstrate SIMS's effectiveness in executing various tasks and generalizing across different scenarios, significantly outperforming previous methods.	[ { "version": "v1", "created": "Fri, 29 Nov 2024 18:36:15 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 04:09:27 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Wenjia", "" ], [ "Pan", "Liang", "" ], [ "Dou", "Zhiyang", "" ], [ "Mei", "Jidong", "" ], [ "Liao", "Zhouyingcheng", "" ], [ "Lou", "Yuke", "" ], [ "Wu", "Yifan", "" ], [ "Yang", "Lei", "" ], [ "Wang", "Jingbo", "" ], [ "Komura", "Taku", "" ] ]	TITLE: SIMS: Simulating Stylized Human-Scene Interactions with Retrieval-Augmented Script Generation ABSTRACT: Simulating stylized human-scene interactions (HSI) in physical environments is a challenging yet fascinating task. Prior works emphasize long-term execution but fall short in achieving both diverse style and physical plausibility. To tackle this challenge, we introduce a novel hierarchical framework named SIMS that seamlessly bridges highlevel script-driven intent with a low-level control policy, enabling more expressive and diverse human-scene interactions. Specifically, we employ Large Language Models with Retrieval-Augmented Generation (RAG) to generate coherent and diverse long-form scripts, providing a rich foundation for motion planning. A versatile multicondition physics-based control policy is also developed, which leverages text embeddings from the generated scripts to encode stylistic cues, simultaneously perceiving environmental geometries and accomplishing task goals. By integrating the retrieval-augmented script generation with the multi-condition controller, our approach provides a unified solution for generating stylized HSI motions. We further introduce a comprehensive planning dataset produced by RAG and a stylized motion dataset featuring diverse locomotions and interactions. Extensive experiments demonstrate SIMS's effectiveness in executing various tasks and generalizing across different scenarios, significantly outperforming previous methods.
2412.00622	Heitor Medeiros Mr.	Heitor R. Medeiros, Atif Belal, Srikanth Muralidharan, Eric Granger and Marco Pedersoli	Visual Modality Prompt for Adapting Vision-Language Object Detectors	null	null	null	null	cs.CV cs.AI cs.LG	http://creativecommons.org/licenses/by/4.0/	The zero-shot performance of object detectors degrades when tested on different modalities, such as infrared and depth. While recent work has explored image translation techniques to adapt detectors to new modalities, these methods are limited to a single modality and apply only to traditional detectors. Recently, vision-language detectors, such as YOLO-World and Grounding DINO, have shown promising zero-shot capabilities, however, they have not yet been adapted for other visual modalities. Traditional fine-tuning approaches compromise the zero-shot capabilities of the detectors. The visual prompt strategies commonly used for classification with vision-language models apply the same linear prompt translation to each image, making them less effective. To address these limitations, we propose ModPrompt, a visual prompt strategy to adapt vision-language detectors to new modalities without degrading zero-shot performance. In particular, an encoder-decoder visual prompt strategy is proposed, further enhanced by the integration of inference-friendly modality prompt decoupled residual, facilitating a more robust adaptation. Empirical benchmarking results show our method for modality adaptation on two vision-language detectors, YOLO-World and Grounding DINO, and on challenging infrared (LLVIP, FLIR) and depth (NYUv2) datasets, achieving performance comparable to full fine-tuning while preserving the model's zero-shot capability. Code available at: https://github.com/heitorrapela/ModPrompt.	[ { "version": "v1", "created": "Sun, 1 Dec 2024 00:19:59 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 20:32:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Medeiros", "Heitor R.", "" ], [ "Belal", "Atif", "" ], [ "Muralidharan", "Srikanth", "" ], [ "Granger", "Eric", "" ], [ "Pedersoli", "Marco", "" ] ]	TITLE: Visual Modality Prompt for Adapting Vision-Language Object Detectors ABSTRACT: The zero-shot performance of object detectors degrades when tested on different modalities, such as infrared and depth. While recent work has explored image translation techniques to adapt detectors to new modalities, these methods are limited to a single modality and apply only to traditional detectors. Recently, vision-language detectors, such as YOLO-World and Grounding DINO, have shown promising zero-shot capabilities, however, they have not yet been adapted for other visual modalities. Traditional fine-tuning approaches compromise the zero-shot capabilities of the detectors. The visual prompt strategies commonly used for classification with vision-language models apply the same linear prompt translation to each image, making them less effective. To address these limitations, we propose ModPrompt, a visual prompt strategy to adapt vision-language detectors to new modalities without degrading zero-shot performance. In particular, an encoder-decoder visual prompt strategy is proposed, further enhanced by the integration of inference-friendly modality prompt decoupled residual, facilitating a more robust adaptation. Empirical benchmarking results show our method for modality adaptation on two vision-language detectors, YOLO-World and Grounding DINO, and on challenging infrared (LLVIP, FLIR) and depth (NYUv2) datasets, achieving performance comparable to full fine-tuning while preserving the model's zero-shot capability. Code available at: https://github.com/heitorrapela/ModPrompt.
2412.00678	Mahdi S. Hosseini Dr.	Jingwei Zhang and Anh Tien Nguyen and Xi Han and Vincent Quoc-Huy Trinh and Hong Qin and Dimitris Samaras and Mahdi S. Hosseini	2DMamba: Efficient State Space Model for Image Representation with Applications on Giga-Pixel Whole Slide Image Classification	Accepted in CVPR 2025 Main Conference	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Efficiently modeling large 2D contexts is essential for various fields including Giga-Pixel Whole Slide Imaging (WSI) and remote sensing. Transformer-based models offer high parallelism but face challenges due to their quadratic complexity for handling long sequences. Recently, Mamba introduced a selective State Space Model (SSM) with linear complexity and high parallelism, enabling effective and efficient modeling of wide context in 1D sequences. However, extending Mamba to vision tasks, which inherently involve 2D structures, results in spatial discrepancies due to the limitations of 1D sequence processing. On the other hand, current 2D SSMs inherently model 2D structures but they suffer from prohibitively slow computation due to the lack of efficient parallel algorithms. In this work, we propose 2DMamba, a novel 2D selective SSM framework that incorporates the 2D spatial structure of images into Mamba, with a highly optimized hardware-aware operator, adopting both spatial continuity and computational efficiency. We validate the versatility of our approach on both WSIs and natural images. Extensive experiments on 10 public datasets for WSI classification and survival analysis show that 2DMamba improves up to 2.48% in AUC, 3.11% in F1 score, 2.47% in accuracy and 5.52% in C-index. Additionally, integrating our method with VMamba for natural imaging yields 0.5 to 0.7 improvements in mIoU on the ADE20k semantic segmentation dataset, and 0.2% accuracy improvement on ImageNet-1K classification dataset. Our code is available at https://github.com/AtlasAnalyticsLab/2DMamba.	[ { "version": "v1", "created": "Sun, 1 Dec 2024 05:42:58 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 22:54:10 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Jingwei", "" ], [ "Nguyen", "Anh Tien", "" ], [ "Han", "Xi", "" ], [ "Trinh", "Vincent Quoc-Huy", "" ], [ "Qin", "Hong", "" ], [ "Samaras", "Dimitris", "" ], [ "Hosseini", "Mahdi S.", "" ] ]	TITLE: 2DMamba: Efficient State Space Model for Image Representation with Applications on Giga-Pixel Whole Slide Image Classification ABSTRACT: Efficiently modeling large 2D contexts is essential for various fields including Giga-Pixel Whole Slide Imaging (WSI) and remote sensing. Transformer-based models offer high parallelism but face challenges due to their quadratic complexity for handling long sequences. Recently, Mamba introduced a selective State Space Model (SSM) with linear complexity and high parallelism, enabling effective and efficient modeling of wide context in 1D sequences. However, extending Mamba to vision tasks, which inherently involve 2D structures, results in spatial discrepancies due to the limitations of 1D sequence processing. On the other hand, current 2D SSMs inherently model 2D structures but they suffer from prohibitively slow computation due to the lack of efficient parallel algorithms. In this work, we propose 2DMamba, a novel 2D selective SSM framework that incorporates the 2D spatial structure of images into Mamba, with a highly optimized hardware-aware operator, adopting both spatial continuity and computational efficiency. We validate the versatility of our approach on both WSIs and natural images. Extensive experiments on 10 public datasets for WSI classification and survival analysis show that 2DMamba improves up to 2.48% in AUC, 3.11% in F1 score, 2.47% in accuracy and 5.52% in C-index. Additionally, integrating our method with VMamba for natural imaging yields 0.5 to 0.7 improvements in mIoU on the ADE20k semantic segmentation dataset, and 0.2% accuracy improvement on ImageNet-1K classification dataset. Our code is available at https://github.com/AtlasAnalyticsLab/2DMamba.
2412.02857	Youssef Mansour	Youssef Mansour and Reinhard Heckel	Measuring Bias of Web-filtered Text Datasets and Bias Propagation Through Training	null	null	null	null	cs.LG	http://creativecommons.org/licenses/by/4.0/	We investigate biases in pretraining datasets for large language models (LLMs) through dataset classification experiments. Building on prior work demonstrating the existence of biases in popular computer vision datasets, we analyze popular open-source pretraining datasets for LLMs derived from CommonCrawl including C4, RefinedWeb, DolmaCC, RedPajama-V2, FineWeb, and DCLM-Baseline. Despite those datasets being obtained with similar curation steps, neural networks can classify surprisingly well which dataset a single text sequence belongs to, significantly better than a human can. This indicates that small differences in filtering and processing pipelines induce fingerprints evident in formatting, vocabulary, and content distributions. Those biases remain even when the text is rewritten with LLMs. Moreover, these biases propagate through training: Random sequences generated by models trained on those datasets can be classified well by a classifier trained on the original datasets. This can be leveraged to estimate the pretraining mixture proportions of the data sources.	[ { "version": "v1", "created": "Tue, 3 Dec 2024 21:43:58 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 23:07:45 GMT" } ]	2025-03-18T00:00:00	[ [ "Mansour", "Youssef", "" ], [ "Heckel", "Reinhard", "" ] ]	TITLE: Measuring Bias of Web-filtered Text Datasets and Bias Propagation Through Training ABSTRACT: We investigate biases in pretraining datasets for large language models (LLMs) through dataset classification experiments. Building on prior work demonstrating the existence of biases in popular computer vision datasets, we analyze popular open-source pretraining datasets for LLMs derived from CommonCrawl including C4, RefinedWeb, DolmaCC, RedPajama-V2, FineWeb, and DCLM-Baseline. Despite those datasets being obtained with similar curation steps, neural networks can classify surprisingly well which dataset a single text sequence belongs to, significantly better than a human can. This indicates that small differences in filtering and processing pipelines induce fingerprints evident in formatting, vocabulary, and content distributions. Those biases remain even when the text is rewritten with LLMs. Moreover, these biases propagate through training: Random sequences generated by models trained on those datasets can be classified well by a classifier trained on the original datasets. This can be leveraged to estimate the pretraining mixture proportions of the data sources.
2412.04606	Chenyu Wang	Chenyu Wang, Weichao Zhou, Shantanu Ghosh, Kayhan Batmanghelich, Wenchao Li	Semantic Consistency-Based Uncertainty Quantification for Factuality in Radiology Report Generation	null	null	null	null	cs.AI cs.CL	http://creativecommons.org/licenses/by/4.0/	Radiology report generation (RRG) has shown great potential in assisting radiologists by automating the labor-intensive task of report writing. While recent advancements have improved the quality and coherence of generated reports, ensuring their factual correctness remains a critical challenge. Although generative medical Vision Large Language Models (VLLMs) have been proposed to address this issue, these models are prone to hallucinations and can produce inaccurate diagnostic information. To address these concerns, we introduce a novel Semantic Consistency-Based Uncertainty Quantification framework that provides both report-level and sentence-level uncertainties. Unlike existing approaches, our method does not require modifications to the underlying model or access to its inner state, such as output token logits, thus serving as a plug-and-play module that can be seamlessly integrated with state-of-the-art models. Extensive experiments demonstrate the efficacy of our method in detecting hallucinations and enhancing the factual accuracy of automatically generated radiology reports. By abstaining from high-uncertainty reports, our approach improves factuality scores by $10$\%, achieved by rejecting $20$\% of reports using the \texttt{Radialog} model on the MIMIC-CXR dataset. Furthermore, sentence-level uncertainty flags the lowest-precision sentence in each report with an $82.9$\% success rate. Our implementation is open-source and available at https://github.com/BU-DEPEND-Lab/SCUQ-RRG.	[ { "version": "v1", "created": "Thu, 5 Dec 2024 20:43:39 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 19:19:05 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Chenyu", "" ], [ "Zhou", "Weichao", "" ], [ "Ghosh", "Shantanu", "" ], [ "Batmanghelich", "Kayhan", "" ], [ "Li", "Wenchao", "" ] ]	TITLE: Semantic Consistency-Based Uncertainty Quantification for Factuality in Radiology Report Generation ABSTRACT: Radiology report generation (RRG) has shown great potential in assisting radiologists by automating the labor-intensive task of report writing. While recent advancements have improved the quality and coherence of generated reports, ensuring their factual correctness remains a critical challenge. Although generative medical Vision Large Language Models (VLLMs) have been proposed to address this issue, these models are prone to hallucinations and can produce inaccurate diagnostic information. To address these concerns, we introduce a novel Semantic Consistency-Based Uncertainty Quantification framework that provides both report-level and sentence-level uncertainties. Unlike existing approaches, our method does not require modifications to the underlying model or access to its inner state, such as output token logits, thus serving as a plug-and-play module that can be seamlessly integrated with state-of-the-art models. Extensive experiments demonstrate the efficacy of our method in detecting hallucinations and enhancing the factual accuracy of automatically generated radiology reports. By abstaining from high-uncertainty reports, our approach improves factuality scores by $10$\%, achieved by rejecting $20$\% of reports using the \texttt{Radialog} model on the MIMIC-CXR dataset. Furthermore, sentence-level uncertainty flags the lowest-precision sentence in each report with an $82.9$\% success rate. Our implementation is open-source and available at https://github.com/BU-DEPEND-Lab/SCUQ-RRG.

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.