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
2503.14064	Xinhao Xiang	Xinhao Xiang, Xiao Liu, Zizhong Li, Zhuosheng Liu, Jiawei Zhang	AIGVE-Tool: AI-Generated Video Evaluation Toolkit with Multifaceted Benchmark	null	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	The rapid advancement in AI-generated video synthesis has led to a growth demand for standardized and effective evaluation metrics. Existing metrics lack a unified framework for systematically categorizing methodologies, limiting a holistic understanding of the evaluation landscape. Additionally, fragmented implementations and the absence of standardized interfaces lead to redundant processing overhead. Furthermore, many prior approaches are constrained by dataset-specific dependencies, limiting their applicability across diverse video domains. To address these challenges, we introduce AIGVE-Tool (AI-Generated Video Evaluation Toolkit), a unified framework that provides a structured and extensible evaluation pipeline for a comprehensive AI-generated video evaluation. Organized within a novel five-category taxonomy, AIGVE-Tool integrates multiple evaluation methodologies while allowing flexible customization through a modular configuration system. Additionally, we propose AIGVE-Bench, a large-scale benchmark dataset created with five SOTA video generation models based on hand-crafted instructions and prompts. This dataset systematically evaluates various video generation models across nine critical quality dimensions. Extensive experiments demonstrate the effectiveness of AIGVE-Tool in providing standardized and reliable evaluation results, highlighting specific strengths and limitations of current models and facilitating the advancements of next-generation AI-generated video techniques.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 09:36:33 GMT" } ]	2025-03-19T00:00:00	[ [ "Xiang", "Xinhao", "" ], [ "Liu", "Xiao", "" ], [ "Li", "Zizhong", "" ], [ "Liu", "Zhuosheng", "" ], [ "Zhang", "Jiawei", "" ] ]	TITLE: AIGVE-Tool: AI-Generated Video Evaluation Toolkit with Multifaceted Benchmark ABSTRACT: The rapid advancement in AI-generated video synthesis has led to a growth demand for standardized and effective evaluation metrics. Existing metrics lack a unified framework for systematically categorizing methodologies, limiting a holistic understanding of the evaluation landscape. Additionally, fragmented implementations and the absence of standardized interfaces lead to redundant processing overhead. Furthermore, many prior approaches are constrained by dataset-specific dependencies, limiting their applicability across diverse video domains. To address these challenges, we introduce AIGVE-Tool (AI-Generated Video Evaluation Toolkit), a unified framework that provides a structured and extensible evaluation pipeline for a comprehensive AI-generated video evaluation. Organized within a novel five-category taxonomy, AIGVE-Tool integrates multiple evaluation methodologies while allowing flexible customization through a modular configuration system. Additionally, we propose AIGVE-Bench, a large-scale benchmark dataset created with five SOTA video generation models based on hand-crafted instructions and prompts. This dataset systematically evaluates various video generation models across nine critical quality dimensions. Extensive experiments demonstrate the effectiveness of AIGVE-Tool in providing standardized and reliable evaluation results, highlighting specific strengths and limitations of current models and facilitating the advancements of next-generation AI-generated video techniques.
2503.14070	Junliang Guo	Yang Ye, Junliang Guo, Haoyu Wu, Tianyu He, Tim Pearce, Tabish Rashid, Katja Hofmann, Jiang Bian	Fast Autoregressive Video Generation with Diagonal Decoding	null	null	null	null	cs.CV cs.AI	http://creativecommons.org/licenses/by-nc-nd/4.0/	Autoregressive Transformer models have demonstrated impressive performance in video generation, but their sequential token-by-token decoding process poses a major bottleneck, particularly for long videos represented by tens of thousands of tokens. In this paper, we propose Diagonal Decoding (DiagD), a training-free inference acceleration algorithm for autoregressively pre-trained models that exploits spatial and temporal correlations in videos. Our method generates tokens along diagonal paths in the spatial-temporal token grid, enabling parallel decoding within each frame as well as partially overlapping across consecutive frames. The proposed algorithm is versatile and adaptive to various generative models and tasks, while providing flexible control over the trade-off between inference speed and visual quality. Furthermore, we propose a cost-effective finetuning strategy that aligns the attention patterns of the model with our decoding order, further mitigating the training-inference gap on small-scale models. Experiments on multiple autoregressive video generation models and datasets demonstrate that DiagD achieves up to $10\times$ speedup compared to naive sequential decoding, while maintaining comparable visual fidelity.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 09:42:55 GMT" } ]	2025-03-19T00:00:00	[ [ "Ye", "Yang", "" ], [ "Guo", "Junliang", "" ], [ "Wu", "Haoyu", "" ], [ "He", "Tianyu", "" ], [ "Pearce", "Tim", "" ], [ "Rashid", "Tabish", "" ], [ "Hofmann", "Katja", "" ], [ "Bian", "Jiang", "" ] ]	TITLE: Fast Autoregressive Video Generation with Diagonal Decoding ABSTRACT: Autoregressive Transformer models have demonstrated impressive performance in video generation, but their sequential token-by-token decoding process poses a major bottleneck, particularly for long videos represented by tens of thousands of tokens. In this paper, we propose Diagonal Decoding (DiagD), a training-free inference acceleration algorithm for autoregressively pre-trained models that exploits spatial and temporal correlations in videos. Our method generates tokens along diagonal paths in the spatial-temporal token grid, enabling parallel decoding within each frame as well as partially overlapping across consecutive frames. The proposed algorithm is versatile and adaptive to various generative models and tasks, while providing flexible control over the trade-off between inference speed and visual quality. Furthermore, we propose a cost-effective finetuning strategy that aligns the attention patterns of the model with our decoding order, further mitigating the training-inference gap on small-scale models. Experiments on multiple autoregressive video generation models and datasets demonstrate that DiagD achieves up to $10\times$ speedup compared to naive sequential decoding, while maintaining comparable visual fidelity.
2503.14072	Rossana Mastrandrea	Rossana Mastrandrea, Fabio Montobbio, Gabriele Pellegrino, Massimo Riccaboni, Valerio Sterzi	Leveraging Knowledge Networks: Rethinking Technological Value Distribution in mRNA Vaccine Innovations	null	null	null	null	physics.soc-ph econ.GN q-fin.EC	http://creativecommons.org/licenses/by/4.0/	This study examines the roles of public and private sector actors in the development of mRNA vaccines, a breakthrough innovation in modern medicine. Using a dataset of 151 core patent families and 2,416 antecedent (cited) patents, we analyze the structure and dynamics of the mRNA vaccine knowledge network through network theory. Our findings highlight the central role of biotechnology firms, such as Moderna and BioNTech, alongside the crucial contributions of universities and public research organizations (PROs) in providing foundational knowledge.We develop a novel credit allocation framework, showing that universities, PROs, government and research centers account for at least 27% of the external technological knowledge base behind mRNA vaccine breakthroughs - representing a minimum threshold of their overall contribution. Our study offers new insights into pharmaceutical and biotechnology innovation dynamics, emphasizing how Moderna and BioNTech's mRNA technologies have benefited from academic institutions, with notable differences in their institutional knowledge sources.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 09:45:19 GMT" } ]	2025-03-19T00:00:00	[ [ "Mastrandrea", "Rossana", "" ], [ "Montobbio", "Fabio", "" ], [ "Pellegrino", "Gabriele", "" ], [ "Riccaboni", "Massimo", "" ], [ "Sterzi", "Valerio", "" ] ]	TITLE: Leveraging Knowledge Networks: Rethinking Technological Value Distribution in mRNA Vaccine Innovations ABSTRACT: This study examines the roles of public and private sector actors in the development of mRNA vaccines, a breakthrough innovation in modern medicine. Using a dataset of 151 core patent families and 2,416 antecedent (cited) patents, we analyze the structure and dynamics of the mRNA vaccine knowledge network through network theory. Our findings highlight the central role of biotechnology firms, such as Moderna and BioNTech, alongside the crucial contributions of universities and public research organizations (PROs) in providing foundational knowledge.We develop a novel credit allocation framework, showing that universities, PROs, government and research centers account for at least 27% of the external technological knowledge base behind mRNA vaccine breakthroughs - representing a minimum threshold of their overall contribution. Our study offers new insights into pharmaceutical and biotechnology innovation dynamics, emphasizing how Moderna and BioNTech's mRNA technologies have benefited from academic institutions, with notable differences in their institutional knowledge sources.
2503.14084	Rongfei Fan	Xingrun Yan, Shiyuan Zuo, Yifeng Lyu, Rongfei Fan, Han Hu	Semantic Communication in Dynamic Channel Scenarios: Collaborative Optimization of Dual-Pipeline Joint Source-Channel Coding and Personalized Federated Learning	null	null	null	null	eess.IV cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Semantic communication is designed to tackle issues like bandwidth constraints and high latency in communication systems. However, in complex network topologies with multiple users, the enormous combinations of client data and channel state information (CSI) pose significant challenges for existing semantic communication architectures. To improve the generalization ability of semantic communication models in complex scenarios while meeting the personalized needs of each user in their local environments, we propose a novel personalized federated learning framework with dual-pipeline joint source-channel coding based on channel awareness model (PFL-DPJSCCA). Within this framework, we present a method that achieves zero optimization gap for non-convex loss functions. Experiments conducted under varying SNR distributions validate the outstanding performance of our framework across diverse datasets.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:02:22 GMT" } ]	2025-03-19T00:00:00	[ [ "Yan", "Xingrun", "" ], [ "Zuo", "Shiyuan", "" ], [ "Lyu", "Yifeng", "" ], [ "Fan", "Rongfei", "" ], [ "Hu", "Han", "" ] ]	TITLE: Semantic Communication in Dynamic Channel Scenarios: Collaborative Optimization of Dual-Pipeline Joint Source-Channel Coding and Personalized Federated Learning ABSTRACT: Semantic communication is designed to tackle issues like bandwidth constraints and high latency in communication systems. However, in complex network topologies with multiple users, the enormous combinations of client data and channel state information (CSI) pose significant challenges for existing semantic communication architectures. To improve the generalization ability of semantic communication models in complex scenarios while meeting the personalized needs of each user in their local environments, we propose a novel personalized federated learning framework with dual-pipeline joint source-channel coding based on channel awareness model (PFL-DPJSCCA). Within this framework, we present a method that achieves zero optimization gap for non-convex loss functions. Experiments conducted under varying SNR distributions validate the outstanding performance of our framework across diverse datasets.
2503.14090	Jan G\"opfert	Jan G\"opfert, Patrick Kuckertz, Jann M. Weinand, Detlef Stolten	Wiki-Quantities and Wiki-Measurements: Datasets of Quantities and their Measurement Context from Wikipedia	null	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	To cope with the large number of publications, more and more researchers are automatically extracting data of interest using natural language processing methods based on supervised learning. Much data, especially in the natural and engineering sciences, is quantitative, but there is a lack of datasets for identifying quantities and their context in text. To address this issue, we present two large datasets based on Wikipedia and Wikidata: Wiki-Quantities is a dataset consisting of over 1.2 million annotated quantities in the English-language Wikipedia. Wiki-Measurements is a dataset of 38,738 annotated quantities in the English-language Wikipedia along with their respective measured entity, property, and optional qualifiers. Manual validation of 100 samples each of Wiki-Quantities and Wiki-Measurements found 100% and 84-94% correct, respectively. The datasets can be used in pipeline approaches to measurement extraction, where quantities are first identified and then their measurement context. To allow reproduction of this work using newer or different versions of Wikipedia and Wikidata, we publish the code used to create the datasets along with the data.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:09:10 GMT" } ]	2025-03-19T00:00:00	[ [ "Göpfert", "Jan", "" ], [ "Kuckertz", "Patrick", "" ], [ "Weinand", "Jann M.", "" ], [ "Stolten", "Detlef", "" ] ]	TITLE: Wiki-Quantities and Wiki-Measurements: Datasets of Quantities and their Measurement Context from Wikipedia ABSTRACT: To cope with the large number of publications, more and more researchers are automatically extracting data of interest using natural language processing methods based on supervised learning. Much data, especially in the natural and engineering sciences, is quantitative, but there is a lack of datasets for identifying quantities and their context in text. To address this issue, we present two large datasets based on Wikipedia and Wikidata: Wiki-Quantities is a dataset consisting of over 1.2 million annotated quantities in the English-language Wikipedia. Wiki-Measurements is a dataset of 38,738 annotated quantities in the English-language Wikipedia along with their respective measured entity, property, and optional qualifiers. Manual validation of 100 samples each of Wiki-Quantities and Wiki-Measurements found 100% and 84-94% correct, respectively. The datasets can be used in pipeline approaches to measurement extraction, where quantities are first identified and then their measurement context. To allow reproduction of this work using newer or different versions of Wikipedia and Wikidata, we publish the code used to create the datasets along with the data.
2503.14095	Bipin Kumar Dr.	Bipin Kumar, Bhvisy Kumar Yadav, Soumypdeep Mukhopadhyay, Rakshit Rohan, Bhupendra Bahadur Singh, Rajib Chattopadhyay, Nagraju Chilukoti, Atul Kumar Sahai	Towards Location-Specific Precipitation Projections Using Deep Neural Networks	21 pages, 9 figures	null	null	null	physics.ao-ph cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Accurate precipitation estimates at individual locations are crucial for weather forecasting and spatial analysis. This study presents a paradigm shift by leveraging Deep Neural Networks (DNNs) to surpass traditional methods like Kriging for station-specific precipitation approximation. We propose two innovative NN architectures: one utilizing precipitation, elevation, and location, and another incorporating additional meteorological parameters like humidity, temperature, and wind speed. Trained on a vast dataset (1980-2019), these models outperform Kriging across various evaluation metrics (correlation coefficient, root mean square error, bias, and skill score) on a five-year validation set. This compelling evidence demonstrates the transformative power of deep learning for spatial prediction, offering a robust and precise alternative for station-specific precipitation estimation.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:12:17 GMT" } ]	2025-03-19T00:00:00	[ [ "Kumar", "Bipin", "" ], [ "Yadav", "Bhvisy Kumar", "" ], [ "Mukhopadhyay", "Soumypdeep", "" ], [ "Rohan", "Rakshit", "" ], [ "Singh", "Bhupendra Bahadur", "" ], [ "Chattopadhyay", "Rajib", "" ], [ "Chilukoti", "Nagraju", "" ], [ "Sahai", "Atul Kumar", "" ] ]	TITLE: Towards Location-Specific Precipitation Projections Using Deep Neural Networks ABSTRACT: Accurate precipitation estimates at individual locations are crucial for weather forecasting and spatial analysis. This study presents a paradigm shift by leveraging Deep Neural Networks (DNNs) to surpass traditional methods like Kriging for station-specific precipitation approximation. We propose two innovative NN architectures: one utilizing precipitation, elevation, and location, and another incorporating additional meteorological parameters like humidity, temperature, and wind speed. Trained on a vast dataset (1980-2019), these models outperform Kriging across various evaluation metrics (correlation coefficient, root mean square error, bias, and skill score) on a five-year validation set. This compelling evidence demonstrates the transformative power of deep learning for spatial prediction, offering a robust and precise alternative for station-specific precipitation estimation.
2503.14106	Jef Jonkers	Jef Jonkers, Frank Coopman, Luc Duchateau, Glenn Van Wallendael, Sofie Van Hoecke	Reliable uncertainty quantification for 2D/3D anatomical landmark localization using multi-output conformal prediction	33 pages, 10 figures	null	null	null	cs.CV cs.AI stat.ML	http://creativecommons.org/licenses/by/4.0/	Automatic anatomical landmark localization in medical imaging requires not just accurate predictions but reliable uncertainty quantification for effective clinical decision support. Current uncertainty quantification approaches often fall short, particularly when combined with normality assumptions, systematically underestimating total predictive uncertainty. This paper introduces conformal prediction as a framework for reliable uncertainty quantification in anatomical landmark localization, addressing a critical gap in automatic landmark localization. We present two novel approaches guaranteeing finite-sample validity for multi-output prediction: Multi-output Regression-as-Classification Conformal Prediction (M-R2CCP) and its variant Multi-output Regression to Classification Conformal Prediction set to Region (M-R2C2R). Unlike conventional methods that produce axis-aligned hyperrectangular or ellipsoidal regions, our approaches generate flexible, non-convex prediction regions that better capture the underlying uncertainty structure of landmark predictions. Through extensive empirical evaluation across multiple 2D and 3D datasets, we demonstrate that our methods consistently outperform existing multi-output conformal prediction approaches in both validity and efficiency. This work represents a significant advancement in reliable uncertainty estimation for anatomical landmark localization, providing clinicians with trustworthy confidence measures for their diagnoses. While developed for medical imaging, these methods show promise for broader applications in multi-output regression problems.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:21:32 GMT" } ]	2025-03-19T00:00:00	[ [ "Jonkers", "Jef", "" ], [ "Coopman", "Frank", "" ], [ "Duchateau", "Luc", "" ], [ "Van Wallendael", "Glenn", "" ], [ "Van Hoecke", "Sofie", "" ] ]	TITLE: Reliable uncertainty quantification for 2D/3D anatomical landmark localization using multi-output conformal prediction ABSTRACT: Automatic anatomical landmark localization in medical imaging requires not just accurate predictions but reliable uncertainty quantification for effective clinical decision support. Current uncertainty quantification approaches often fall short, particularly when combined with normality assumptions, systematically underestimating total predictive uncertainty. This paper introduces conformal prediction as a framework for reliable uncertainty quantification in anatomical landmark localization, addressing a critical gap in automatic landmark localization. We present two novel approaches guaranteeing finite-sample validity for multi-output prediction: Multi-output Regression-as-Classification Conformal Prediction (M-R2CCP) and its variant Multi-output Regression to Classification Conformal Prediction set to Region (M-R2C2R). Unlike conventional methods that produce axis-aligned hyperrectangular or ellipsoidal regions, our approaches generate flexible, non-convex prediction regions that better capture the underlying uncertainty structure of landmark predictions. Through extensive empirical evaluation across multiple 2D and 3D datasets, we demonstrate that our methods consistently outperform existing multi-output conformal prediction approaches in both validity and efficiency. This work represents a significant advancement in reliable uncertainty estimation for anatomical landmark localization, providing clinicians with trustworthy confidence measures for their diagnoses. While developed for medical imaging, these methods show promise for broader applications in multi-output regression problems.
2503.14109	Nicolas Gonthier	Nicolas Gonthier	Operational Change Detection for Geographical Information: Overview and Challenges	Preprint under review	null	null	null	cs.CV cs.AI	http://creativecommons.org/licenses/by/4.0/	Rapid evolution of territories due to climate change and human impact requires prompt and effective updates to geospatial databases maintained by the National Mapping Agency. This paper presents a comprehensive overview of change detection methods tailored for the operational updating of large-scale geographic databases. This review first outlines the fundamental definition of change, emphasizing its multifaceted nature, from temporal to semantic characterization. It categorizes automatic change detection methods into four main families: rule-based, statistical, machine learning, and simulation methods. The strengths, limitations, and applicability of every family are discussed in the context of various input data. Then, key applications for National Mapping Agencies are identified, particularly the optimization of geospatial database updating, change-based phenomena, and dynamics monitoring. Finally, the paper highlights the current challenges for leveraging change detection such as the variability of change definition, the missing of relevant large-scale datasets, the diversity of input data, the unstudied no-change detection, the human in the loop integration and the operational constraints. The discussion underscores the necessity for ongoing innovation in change detection techniques to address the future needs of geographic information systems for national mapping agencies.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:25:28 GMT" } ]	2025-03-19T00:00:00	[ [ "Gonthier", "Nicolas", "" ] ]	TITLE: Operational Change Detection for Geographical Information: Overview and Challenges ABSTRACT: Rapid evolution of territories due to climate change and human impact requires prompt and effective updates to geospatial databases maintained by the National Mapping Agency. This paper presents a comprehensive overview of change detection methods tailored for the operational updating of large-scale geographic databases. This review first outlines the fundamental definition of change, emphasizing its multifaceted nature, from temporal to semantic characterization. It categorizes automatic change detection methods into four main families: rule-based, statistical, machine learning, and simulation methods. The strengths, limitations, and applicability of every family are discussed in the context of various input data. Then, key applications for National Mapping Agencies are identified, particularly the optimization of geospatial database updating, change-based phenomena, and dynamics monitoring. Finally, the paper highlights the current challenges for leveraging change detection such as the variability of change definition, the missing of relevant large-scale datasets, the diversity of input data, the unstudied no-change detection, the human in the loop integration and the operational constraints. The discussion underscores the necessity for ongoing innovation in change detection techniques to address the future needs of geographic information systems for national mapping agencies.
2503.14112	Guodong Ding Dr.	Guodong Ding, Rongyu Chen and Angela Yao	Condensing Action Segmentation Datasets via Generative Network Inversion	10 pages, 3 figures, 5 tables, Accepted to CVPR2025	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	This work presents the first condensation approach for procedural video datasets used in temporal action segmentation. We propose a condensation framework that leverages generative prior learned from the dataset and network inversion to condense data into compact latent codes with significant storage reduced across temporal and channel aspects. Orthogonally, we propose sampling diverse and representative action sequences to minimize video-wise redundancy. Our evaluation on standard benchmarks demonstrates consistent effectiveness in condensing TAS datasets and achieving competitive performances. Specifically, on the Breakfast dataset, our approach reduces storage by over 500$\times$ while retaining 83% of the performance compared to training with the full dataset. Furthermore, when applied to a downstream incremental learning task, it yields superior performance compared to the state-of-the-art.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:29:47 GMT" } ]	2025-03-19T00:00:00	[ [ "Ding", "Guodong", "" ], [ "Chen", "Rongyu", "" ], [ "Yao", "Angela", "" ] ]	TITLE: Condensing Action Segmentation Datasets via Generative Network Inversion ABSTRACT: This work presents the first condensation approach for procedural video datasets used in temporal action segmentation. We propose a condensation framework that leverages generative prior learned from the dataset and network inversion to condense data into compact latent codes with significant storage reduced across temporal and channel aspects. Orthogonally, we propose sampling diverse and representative action sequences to minimize video-wise redundancy. Our evaluation on standard benchmarks demonstrates consistent effectiveness in condensing TAS datasets and achieving competitive performances. Specifically, on the Breakfast dataset, our approach reduces storage by over 500$\times$ while retaining 83% of the performance compared to training with the full dataset. Furthermore, when applied to a downstream incremental learning task, it yields superior performance compared to the state-of-the-art.
2503.14118	Michele Ceriotti	Arslan Mazitov, Filippo Bigi, Matthias Kellner, Paolo Pegolo, Davide Tisi, Guillaume Fraux, Sergey Pozdnyakov, Philip Loche, and Michele Ceriotti	PET-MAD, a universal interatomic potential for advanced materials modeling	null	null	null	null	cond-mat.mtrl-sci cs.LG physics.chem-ph	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Machine-learning interatomic potentials (MLIPs) have greatly extended the reach of atomic-scale simulations, offering the accuracy of first-principles calculations at a fraction of the effort. Leveraging large quantum mechanical databases and expressive architectures, recent "universal" models deliver qualitative accuracy across the periodic table but are often biased toward low-energy configurations. We introduce PET-MAD, a generally applicable MLIP trained on a dataset combining stable inorganic and organic solids, systematically modified to enhance atomic diversity. Using a moderate but highly-consistent level of electronic-structure theory, we assess PET-MAD's accuracy on established benchmarks and advanced simulations of six materials. PET-MAD rivals state-of-the-art MLIPs for inorganic solids, while also being reliable for molecules, organic materials, and surfaces. It is stable and fast, enabling, out-of-the-box, the near-quantitative study of thermal and quantum mechanical fluctuations, functional properties, and phase transitions. It can be efficiently fine-tuned to deliver full quantum mechanical accuracy with a minimal number of targeted calculations.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:35:30 GMT" } ]	2025-03-19T00:00:00	[ [ "Mazitov", "Arslan", "" ], [ "Bigi", "Filippo", "" ], [ "Kellner", "Matthias", "" ], [ "Pegolo", "Paolo", "" ], [ "Tisi", "Davide", "" ], [ "Fraux", "Guillaume", "" ], [ "Pozdnyakov", "Sergey", "" ], [ "Loche", "Philip", "" ], [ "Ceriotti", "Michele", "" ] ]	TITLE: PET-MAD, a universal interatomic potential for advanced materials modeling ABSTRACT: Machine-learning interatomic potentials (MLIPs) have greatly extended the reach of atomic-scale simulations, offering the accuracy of first-principles calculations at a fraction of the effort. Leveraging large quantum mechanical databases and expressive architectures, recent "universal" models deliver qualitative accuracy across the periodic table but are often biased toward low-energy configurations. We introduce PET-MAD, a generally applicable MLIP trained on a dataset combining stable inorganic and organic solids, systematically modified to enhance atomic diversity. Using a moderate but highly-consistent level of electronic-structure theory, we assess PET-MAD's accuracy on established benchmarks and advanced simulations of six materials. PET-MAD rivals state-of-the-art MLIPs for inorganic solids, while also being reliable for molecules, organic materials, and surfaces. It is stable and fast, enabling, out-of-the-box, the near-quantitative study of thermal and quantum mechanical fluctuations, functional properties, and phase transitions. It can be efficiently fine-tuned to deliver full quantum mechanical accuracy with a minimal number of targeted calculations.
2503.14136	Ankit Dutta	Ankit Dutta, Nabarup Ghosh, Ankush Chatterjee	CARE: A QLoRA-Fine Tuned Multi-Domain Chatbot With Fast Learning On Minimal Hardware	null	null	null	null	cs.CL cs.AI	http://creativecommons.org/licenses/by-nc-nd/4.0/	Large Language models have demonstrated excellent domain-specific question-answering capabilities when finetuned with a particular dataset of that specific domain. However, fine-tuning the models requires a significant amount of training time and a considerable amount of hardware. In this work, we propose CARE (Customer Assistance and Response Engine), a lightweight model made by fine-tuning Phi3.5-mini on very minimal hardware and data, designed to handle queries primarily across three domains: telecommunications support, medical support, and banking support. For telecommunications and banking, the chatbot addresses issues and problems faced by customers regularly in the above-mentioned domains. In the medical domain, CARE provides preliminary support by offering basic diagnoses and medical suggestions that a user might take before consulting a healthcare professional. Since CARE is built on Phi3.5-mini, it can be used even on mobile devices, increasing its usability. Our research also shows that CARE performs relatively well on various medical benchmarks, indicating that it can be used to make basic medical suggestions.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 10:58:10 GMT" } ]	2025-03-19T00:00:00	[ [ "Dutta", "Ankit", "" ], [ "Ghosh", "Nabarup", "" ], [ "Chatterjee", "Ankush", "" ] ]	TITLE: CARE: A QLoRA-Fine Tuned Multi-Domain Chatbot With Fast Learning On Minimal Hardware ABSTRACT: Large Language models have demonstrated excellent domain-specific question-answering capabilities when finetuned with a particular dataset of that specific domain. However, fine-tuning the models requires a significant amount of training time and a considerable amount of hardware. In this work, we propose CARE (Customer Assistance and Response Engine), a lightweight model made by fine-tuning Phi3.5-mini on very minimal hardware and data, designed to handle queries primarily across three domains: telecommunications support, medical support, and banking support. For telecommunications and banking, the chatbot addresses issues and problems faced by customers regularly in the above-mentioned domains. In the medical domain, CARE provides preliminary support by offering basic diagnoses and medical suggestions that a user might take before consulting a healthcare professional. Since CARE is built on Phi3.5-mini, it can be used even on mobile devices, increasing its usability. Our research also shows that CARE performs relatively well on various medical benchmarks, indicating that it can be used to make basic medical suggestions.
2503.14138	Siddharth Jaiswal	Siddharth D Jaiswal, Sagnik Basu, Sandipan Sikdar, Animesh Mukherjee	Exploring Disparity-Accuracy Trade-offs in Face Recognition Systems: The Role of Datasets, Architectures, and Loss Functions	This work has been accepted for publication at AAAI ICWSM 2025	null	null	null	cs.CV cs.AI cs.CY	http://creativecommons.org/licenses/by/4.0/	Automated Face Recognition Systems (FRSs), developed using deep learning models, are deployed worldwide for identity verification and facial attribute analysis. The performance of these models is determined by a complex interdependence among the model architecture, optimization/loss function and datasets. Although FRSs have surpassed human-level accuracy, they continue to be disparate against certain demographics. Due to the ubiquity of applications, it is extremely important to understand the impact of the three components -- model architecture, loss function and face image dataset on the accuracy-disparity trade-off to design better, unbiased platforms. In this work, we perform an in-depth analysis of three FRSs for the task of gender prediction, with various architectural modifications resulting in ten deep-learning models coupled with four loss functions and benchmark them on seven face datasets across 266 evaluation configurations. Our results show that all three components have an individual as well as a combined impact on both accuracy and disparity. We identify that datasets have an inherent property that causes them to perform similarly across models, independent of the choice of loss functions. Moreover, the choice of dataset determines the model's perceived bias -- the same model reports bias in opposite directions for three gender-balanced datasets of ``in-the-wild'' face images of popular individuals. Studying the facial embeddings shows that the models are unable to generalize a uniform definition of what constitutes a ``female face'' as opposed to a ``male face'', due to dataset diversity. We provide recommendations to model developers on using our study as a blueprint for model development and subsequent deployment.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:04:57 GMT" } ]	2025-03-19T00:00:00	[ [ "Jaiswal", "Siddharth D", "" ], [ "Basu", "Sagnik", "" ], [ "Sikdar", "Sandipan", "" ], [ "Mukherjee", "Animesh", "" ] ]	TITLE: Exploring Disparity-Accuracy Trade-offs in Face Recognition Systems: The Role of Datasets, Architectures, and Loss Functions ABSTRACT: Automated Face Recognition Systems (FRSs), developed using deep learning models, are deployed worldwide for identity verification and facial attribute analysis. The performance of these models is determined by a complex interdependence among the model architecture, optimization/loss function and datasets. Although FRSs have surpassed human-level accuracy, they continue to be disparate against certain demographics. Due to the ubiquity of applications, it is extremely important to understand the impact of the three components -- model architecture, loss function and face image dataset on the accuracy-disparity trade-off to design better, unbiased platforms. In this work, we perform an in-depth analysis of three FRSs for the task of gender prediction, with various architectural modifications resulting in ten deep-learning models coupled with four loss functions and benchmark them on seven face datasets across 266 evaluation configurations. Our results show that all three components have an individual as well as a combined impact on both accuracy and disparity. We identify that datasets have an inherent property that causes them to perform similarly across models, independent of the choice of loss functions. Moreover, the choice of dataset determines the model's perceived bias -- the same model reports bias in opposite directions for three gender-balanced datasets of ``in-the-wild'' face images of popular individuals. Studying the facial embeddings shows that the models are unable to generalize a uniform definition of what constitutes a ``female face'' as opposed to a ``male face'', due to dataset diversity. We provide recommendations to model developers on using our study as a blueprint for model development and subsequent deployment.
2503.14140	Zining Wang	Zining Wang, Tongkun Guan, Pei Fu, Chen Duan, Qianyi Jiang, Zhentao Guo, Shan Guo, Junfeng Luo, Wei Shen, Xiaokang Yang	Marten: Visual Question Answering with Mask Generation for Multi-modal Document Understanding	Accepted by CVPR2025	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Multi-modal Large Language Models (MLLMs) have introduced a novel dimension to document understanding, i.e., they endow large language models with visual comprehension capabilities; however, how to design a suitable image-text pre-training task for bridging the visual and language modality in document-level MLLMs remains underexplored. In this study, we introduce a novel visual-language alignment method that casts the key issue as a Visual Question Answering with Mask generation (VQAMask) task, optimizing two tasks simultaneously: VQA-based text parsing and mask generation. The former allows the model to implicitly align images and text at the semantic level. The latter introduces an additional mask generator (discarded during inference) to explicitly ensure alignment between visual texts within images and their corresponding image regions at a spatially-aware level. Together, they can prevent model hallucinations when parsing visual text and effectively promote spatially-aware feature representation learning. To support the proposed VQAMask task, we construct a comprehensive image-mask generation pipeline and provide a large-scale dataset with 6M data (MTMask6M). Subsequently, we demonstrate that introducing the proposed mask generation task yields competitive document-level understanding performance. Leveraging the proposed VQAMask, we introduce Marten, a training-efficient MLLM tailored for document-level understanding. Extensive experiments show that our Marten consistently achieves significant improvements among 8B-MLLMs in document-centric tasks. Code and datasets are available at https://github.com/PriNing/Marten.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:07:14 GMT" } ]	2025-03-19T00:00:00	[ [ "Wang", "Zining", "" ], [ "Guan", "Tongkun", "" ], [ "Fu", "Pei", "" ], [ "Duan", "Chen", "" ], [ "Jiang", "Qianyi", "" ], [ "Guo", "Zhentao", "" ], [ "Guo", "Shan", "" ], [ "Luo", "Junfeng", "" ], [ "Shen", "Wei", "" ], [ "Yang", "Xiaokang", "" ] ]	TITLE: Marten: Visual Question Answering with Mask Generation for Multi-modal Document Understanding ABSTRACT: Multi-modal Large Language Models (MLLMs) have introduced a novel dimension to document understanding, i.e., they endow large language models with visual comprehension capabilities; however, how to design a suitable image-text pre-training task for bridging the visual and language modality in document-level MLLMs remains underexplored. In this study, we introduce a novel visual-language alignment method that casts the key issue as a Visual Question Answering with Mask generation (VQAMask) task, optimizing two tasks simultaneously: VQA-based text parsing and mask generation. The former allows the model to implicitly align images and text at the semantic level. The latter introduces an additional mask generator (discarded during inference) to explicitly ensure alignment between visual texts within images and their corresponding image regions at a spatially-aware level. Together, they can prevent model hallucinations when parsing visual text and effectively promote spatially-aware feature representation learning. To support the proposed VQAMask task, we construct a comprehensive image-mask generation pipeline and provide a large-scale dataset with 6M data (MTMask6M). Subsequently, we demonstrate that introducing the proposed mask generation task yields competitive document-level understanding performance. Leveraging the proposed VQAMask, we introduce Marten, a training-efficient MLLM tailored for document-level understanding. Extensive experiments show that our Marten consistently achieves significant improvements among 8B-MLLMs in document-centric tasks. Code and datasets are available at https://github.com/PriNing/Marten.
2503.14150	Yihang Zhou	Yihang Zhou, Ruige Kong, Zhengsen Xu, Linlin Xu, Sibo Cheng	Comparative and Interpretative Analysis of CNN and Transformer Models in Predicting Wildfire Spread Using Remote Sensing Data	null	null	10.1029/2024JH000409	null	cs.CV eess.IV	http://creativecommons.org/licenses/by-nc-sa/4.0/	Facing the escalating threat of global wildfires, numerous computer vision techniques using remote sensing data have been applied in this area. However, the selection of deep learning methods for wildfire prediction remains uncertain due to the lack of comparative analysis in a quantitative and explainable manner, crucial for improving prevention measures and refining models. This study aims to thoroughly compare the performance, efficiency, and explainability of four prevalent deep learning architectures: Autoencoder, ResNet, UNet, and Transformer-based Swin-UNet. Employing a real-world dataset that includes nearly a decade of remote sensing data from California, U.S., these models predict the spread of wildfires for the following day. Through detailed quantitative comparison analysis, we discovered that Transformer-based Swin-UNet and UNet generally outperform Autoencoder and ResNet, particularly due to the advanced attention mechanisms in Transformer-based Swin-UNet and the efficient use of skip connections in both UNet and Transformer-based Swin-UNet, which contribute to superior predictive accuracy and model interpretability. Then we applied XAI techniques on all four models, this not only enhances the clarity and trustworthiness of models but also promotes focused improvements in wildfire prediction capabilities. The XAI analysis reveals that UNet and Transformer-based Swin-UNet are able to focus on critical features such as 'Previous Fire Mask', 'Drought', and 'Vegetation' more effectively than the other two models, while also maintaining balanced attention to the remaining features, leading to their superior performance. The insights from our thorough comparative analysis offer substantial implications for future model design and also provide guidance for model selection in different scenarios.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:16:48 GMT" } ]	2025-03-19T00:00:00	[ [ "Zhou", "Yihang", "" ], [ "Kong", "Ruige", "" ], [ "Xu", "Zhengsen", "" ], [ "Xu", "Linlin", "" ], [ "Cheng", "Sibo", "" ] ]	TITLE: Comparative and Interpretative Analysis of CNN and Transformer Models in Predicting Wildfire Spread Using Remote Sensing Data ABSTRACT: Facing the escalating threat of global wildfires, numerous computer vision techniques using remote sensing data have been applied in this area. However, the selection of deep learning methods for wildfire prediction remains uncertain due to the lack of comparative analysis in a quantitative and explainable manner, crucial for improving prevention measures and refining models. This study aims to thoroughly compare the performance, efficiency, and explainability of four prevalent deep learning architectures: Autoencoder, ResNet, UNet, and Transformer-based Swin-UNet. Employing a real-world dataset that includes nearly a decade of remote sensing data from California, U.S., these models predict the spread of wildfires for the following day. Through detailed quantitative comparison analysis, we discovered that Transformer-based Swin-UNet and UNet generally outperform Autoencoder and ResNet, particularly due to the advanced attention mechanisms in Transformer-based Swin-UNet and the efficient use of skip connections in both UNet and Transformer-based Swin-UNet, which contribute to superior predictive accuracy and model interpretability. Then we applied XAI techniques on all four models, this not only enhances the clarity and trustworthiness of models but also promotes focused improvements in wildfire prediction capabilities. The XAI analysis reveals that UNet and Transformer-based Swin-UNet are able to focus on critical features such as 'Previous Fire Mask', 'Drought', and 'Vegetation' more effectively than the other two models, while also maintaining balanced attention to the remaining features, leading to their superior performance. The insights from our thorough comparative analysis offer substantial implications for future model design and also provide guidance for model selection in different scenarios.
2503.14153	Changran Xu	Changran Xu, Yi Liu, Yunhao Zhou, Shan Huang, Ningyi Xu, Qiang Xu	Speculative Decoding for Verilog: Speed and Quality, All in One	Accepted by the 62nd Design Automation Conference (DAC 2025)	null	null	null	cs.LG cs.AR cs.CL	http://creativecommons.org/licenses/by/4.0/	The rapid advancement of large language models (LLMs) has revolutionized code generation tasks across various programming languages. However, the unique characteristics of programming languages, particularly those like Verilog with specific syntax and lower representation in training datasets, pose significant challenges for conventional tokenization and decoding approaches. In this paper, we introduce a novel application of speculative decoding for Verilog code generation, showing that it can improve both inference speed and output quality, effectively achieving speed and quality all in one. Unlike standard LLM tokenization schemes, which often fragment meaningful code structures, our approach aligns decoding stops with syntactically significant tokens, making it easier for models to learn the token distribution. This refinement addresses inherent tokenization issues and enhances the model's ability to capture Verilog's logical constructs more effectively. Our experimental results show that our method achieves up to a 5.05x speedup in Verilog code generation and increases pass@10 functional accuracy on RTLLM by up to 17.19% compared to conventional training strategies. These findings highlight speculative decoding as a promising approach to bridge the quality gap in code generation for specialized programming languages.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:21:53 GMT" } ]	2025-03-19T00:00:00	[ [ "Xu", "Changran", "" ], [ "Liu", "Yi", "" ], [ "Zhou", "Yunhao", "" ], [ "Huang", "Shan", "" ], [ "Xu", "Ningyi", "" ], [ "Xu", "Qiang", "" ] ]	TITLE: Speculative Decoding for Verilog: Speed and Quality, All in One ABSTRACT: The rapid advancement of large language models (LLMs) has revolutionized code generation tasks across various programming languages. However, the unique characteristics of programming languages, particularly those like Verilog with specific syntax and lower representation in training datasets, pose significant challenges for conventional tokenization and decoding approaches. In this paper, we introduce a novel application of speculative decoding for Verilog code generation, showing that it can improve both inference speed and output quality, effectively achieving speed and quality all in one. Unlike standard LLM tokenization schemes, which often fragment meaningful code structures, our approach aligns decoding stops with syntactically significant tokens, making it easier for models to learn the token distribution. This refinement addresses inherent tokenization issues and enhances the model's ability to capture Verilog's logical constructs more effectively. Our experimental results show that our method achieves up to a 5.05x speedup in Verilog code generation and increases pass@10 functional accuracy on RTLLM by up to 17.19% compared to conventional training strategies. These findings highlight speculative decoding as a promising approach to bridge the quality gap in code generation for specialized programming languages.
2503.14154	Zhang Chen	Zhang Chen, Shuai Wan, Siyu Ren, Fuzheng Yang, Mengting Yu, and Junhui Hou	RBFIM: Perceptual Quality Assessment for Compressed Point Clouds Using Radial Basis Function Interpolation	null	null	null	null	cs.CV cs.MM eess.IV	http://creativecommons.org/licenses/by-nc-nd/4.0/	One of the main challenges in point cloud compression (PCC) is how to evaluate the perceived distortion so that the codec can be optimized for perceptual quality. Current standard practices in PCC highlight a primary issue: while single-feature metrics are widely used to assess compression distortion, the classic method of searching point-to-point nearest neighbors frequently fails to adequately build precise correspondences between point clouds, resulting in an ineffective capture of human perceptual features. To overcome the related limitations, we propose a novel assessment method called RBFIM, utilizing radial basis function (RBF) interpolation to convert discrete point features into a continuous feature function for the distorted point cloud. By substituting the geometry coordinates of the original point cloud into the feature function, we obtain the bijective sets of point features. This enables an establishment of precise corresponding features between distorted and original point clouds and significantly improves the accuracy of quality assessments. Moreover, this method avoids the complexity caused by bidirectional searches. Extensive experiments on multiple subjective quality datasets of compressed point clouds demonstrate that our RBFIM excels in addressing human perception tasks, thereby providing robust support for PCC optimization efforts.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:25:55 GMT" } ]	2025-03-19T00:00:00	[ [ "Chen", "Zhang", "" ], [ "Wan", "Shuai", "" ], [ "Ren", "Siyu", "" ], [ "Yang", "Fuzheng", "" ], [ "Yu", "Mengting", "" ], [ "Hou", "Junhui", "" ] ]	TITLE: RBFIM: Perceptual Quality Assessment for Compressed Point Clouds Using Radial Basis Function Interpolation ABSTRACT: One of the main challenges in point cloud compression (PCC) is how to evaluate the perceived distortion so that the codec can be optimized for perceptual quality. Current standard practices in PCC highlight a primary issue: while single-feature metrics are widely used to assess compression distortion, the classic method of searching point-to-point nearest neighbors frequently fails to adequately build precise correspondences between point clouds, resulting in an ineffective capture of human perceptual features. To overcome the related limitations, we propose a novel assessment method called RBFIM, utilizing radial basis function (RBF) interpolation to convert discrete point features into a continuous feature function for the distorted point cloud. By substituting the geometry coordinates of the original point cloud into the feature function, we obtain the bijective sets of point features. This enables an establishment of precise corresponding features between distorted and original point clouds and significantly improves the accuracy of quality assessments. Moreover, this method avoids the complexity caused by bidirectional searches. Extensive experiments on multiple subjective quality datasets of compressed point clouds demonstrate that our RBFIM excels in addressing human perception tasks, thereby providing robust support for PCC optimization efforts.
2503.14162	Zongyun Zhang	Zongyun Zhang, Jiacheng Ruan, Xian Gao, Ting Liu, Yuzhuo Fu	EIAD: Explainable Industrial Anomaly Detection Via Multi-Modal Large Language Models	null	null	null	null	cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Industrial Anomaly Detection (IAD) is critical to ensure product quality during manufacturing. Although existing zero-shot defect segmentation and detection methods have shown effectiveness, they cannot provide detailed descriptions of the defects. Furthermore, the application of large multi-modal models in IAD remains in its infancy, facing challenges in balancing question-answering (QA) performance and mask-based grounding capabilities, often owing to overfitting during the fine-tuning process. To address these challenges, we propose a novel approach that introduces a dedicated multi-modal defect localization module to decouple the dialog functionality from the core feature extraction. This decoupling is achieved through independent optimization objectives and tailored learning strategies. Additionally, we contribute to the first multi-modal industrial anomaly detection training dataset, named Defect Detection Question Answering (DDQA), encompassing a wide range of defect types and industrial scenarios. Unlike conventional datasets that rely on GPT-generated data, DDQA ensures authenticity and reliability and offers a robust foundation for model training. Experimental results demonstrate that our proposed method, Explainable Industrial Anomaly Detection Assistant (EIAD), achieves outstanding performance in defect detection and localization tasks. It not only significantly enhances accuracy but also improves interpretability. These advancements highlight the potential of EIAD for practical applications in industrial settings.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:33:29 GMT" } ]	2025-03-19T00:00:00	[ [ "Zhang", "Zongyun", "" ], [ "Ruan", "Jiacheng", "" ], [ "Gao", "Xian", "" ], [ "Liu", "Ting", "" ], [ "Fu", "Yuzhuo", "" ] ]	TITLE: EIAD: Explainable Industrial Anomaly Detection Via Multi-Modal Large Language Models ABSTRACT: Industrial Anomaly Detection (IAD) is critical to ensure product quality during manufacturing. Although existing zero-shot defect segmentation and detection methods have shown effectiveness, they cannot provide detailed descriptions of the defects. Furthermore, the application of large multi-modal models in IAD remains in its infancy, facing challenges in balancing question-answering (QA) performance and mask-based grounding capabilities, often owing to overfitting during the fine-tuning process. To address these challenges, we propose a novel approach that introduces a dedicated multi-modal defect localization module to decouple the dialog functionality from the core feature extraction. This decoupling is achieved through independent optimization objectives and tailored learning strategies. Additionally, we contribute to the first multi-modal industrial anomaly detection training dataset, named Defect Detection Question Answering (DDQA), encompassing a wide range of defect types and industrial scenarios. Unlike conventional datasets that rely on GPT-generated data, DDQA ensures authenticity and reliability and offers a robust foundation for model training. Experimental results demonstrate that our proposed method, Explainable Industrial Anomaly Detection Assistant (EIAD), achieves outstanding performance in defect detection and localization tasks. It not only significantly enhances accuracy but also improves interpretability. These advancements highlight the potential of EIAD for practical applications in industrial settings.
2503.14167	Christian Poelitz	Christian Poelitz, Nick McKenna	Synthetic Clarification and Correction Dialogues about Data-Centric Tasks -- A Teacher-Student Approach	null	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Real dialogues with AI assistants for solving data-centric tasks often follow dynamic, unpredictable paths due to imperfect information provided by the user or in the data, which must be caught and handled. Developing datasets which capture such user-AI interactions is difficult and time-consuming. In this work, we develop a novel framework for synthetically generating controlled, multi-turn conversations between a user and AI assistant for the task of table-based question answering, which can be generated from an existing dataset with fully specified table QA examples for any target domain. Each conversation aims to solve a table-based reasoning question through collaborative effort, modeling one of two real-world scenarios: (1) an AI-initiated clarification, or (2) a user-initiated correction. Critically, we employ a strong teacher LLM to verify the correctness of our synthetic conversations, ensuring high quality. We demonstrate synthetic datasets generated from TAT-QA and WikiTableQuestions as benchmarks of frontier LLMs. We find that even larger models struggle to effectively issuing clarification questions and accurately integrate user feedback for corrections.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:37:25 GMT" } ]	2025-03-19T00:00:00	[ [ "Poelitz", "Christian", "" ], [ "McKenna", "Nick", "" ] ]	TITLE: Synthetic Clarification and Correction Dialogues about Data-Centric Tasks -- A Teacher-Student Approach ABSTRACT: Real dialogues with AI assistants for solving data-centric tasks often follow dynamic, unpredictable paths due to imperfect information provided by the user or in the data, which must be caught and handled. Developing datasets which capture such user-AI interactions is difficult and time-consuming. In this work, we develop a novel framework for synthetically generating controlled, multi-turn conversations between a user and AI assistant for the task of table-based question answering, which can be generated from an existing dataset with fully specified table QA examples for any target domain. Each conversation aims to solve a table-based reasoning question through collaborative effort, modeling one of two real-world scenarios: (1) an AI-initiated clarification, or (2) a user-initiated correction. Critically, we employ a strong teacher LLM to verify the correctness of our synthetic conversations, ensuring high quality. We demonstrate synthetic datasets generated from TAT-QA and WikiTableQuestions as benchmarks of frontier LLMs. We find that even larger models struggle to effectively issuing clarification questions and accurately integrate user feedback for corrections.
2503.14171	Simon Niedermayr	Simon Niedermayr, Christoph Neuhauser R\"udiger Westermann	Lightweight Gradient-Aware Upscaling of 3D Gaussian Splatting Images	null	null	null	null	cs.CV eess.IV	http://creativecommons.org/licenses/by-sa/4.0/	We introduce an image upscaling technique tailored for 3D Gaussian Splatting (3DGS) on lightweight GPUs. Compared to 3DGS, it achieves significantly higher rendering speeds and reduces artifacts commonly observed in 3DGS reconstructions. Our technique upscales low-resolution 3DGS renderings with a marginal increase in cost by directly leveraging the analytical image gradients of Gaussians for gradient-based bicubic spline interpolation. The technique is agnostic to the specific 3DGS implementation, achieving novel view synthesis at rates 3x-4x higher than the baseline implementation. Through extensive experiments on multiple datasets, we showcase the performance improvements and high reconstruction fidelity attainable with gradient-aware upscaling of 3DGS images. We further demonstrate the integration of gradient-aware upscaling into the gradient-based optimization of a 3DGS model and analyze its effects on reconstruction quality and performance.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:42:52 GMT" } ]	2025-03-19T00:00:00	[ [ "Niedermayr", "Simon", "" ], [ "Westermann", "Christoph Neuhauser Rüdiger", "" ] ]	TITLE: Lightweight Gradient-Aware Upscaling of 3D Gaussian Splatting Images ABSTRACT: We introduce an image upscaling technique tailored for 3D Gaussian Splatting (3DGS) on lightweight GPUs. Compared to 3DGS, it achieves significantly higher rendering speeds and reduces artifacts commonly observed in 3DGS reconstructions. Our technique upscales low-resolution 3DGS renderings with a marginal increase in cost by directly leveraging the analytical image gradients of Gaussians for gradient-based bicubic spline interpolation. The technique is agnostic to the specific 3DGS implementation, achieving novel view synthesis at rates 3x-4x higher than the baseline implementation. Through extensive experiments on multiple datasets, we showcase the performance improvements and high reconstruction fidelity attainable with gradient-aware upscaling of 3DGS images. We further demonstrate the integration of gradient-aware upscaling into the gradient-based optimization of a 3DGS model and analyze its effects on reconstruction quality and performance.
2503.14173	Raul Quijada	Guillem Cadevall Ferreres, Marc Serrano Sanz, Marc Bardeli G\'amez, Pol Gerdt Basullas, Francesc Tarres Ruiz, Raul Quijada Ferrero	NERCat: Fine-Tuning for Enhanced Named Entity Recognition in Catalan	7 pages, 1 table	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Named Entity Recognition (NER) is a critical component of Natural Language Processing (NLP) for extracting structured information from unstructured text. However, for low-resource languages like Catalan, the performance of NER systems often suffers due to the lack of high-quality annotated datasets. This paper introduces NERCat, a fine-tuned version of the GLiNER[1] model, designed to improve NER performance specifically for Catalan text. We used a dataset of manually annotated Catalan television transcriptions to train and fine-tune the model, focusing on domains such as politics, sports, and culture. The evaluation results show significant improvements in precision, recall, and F1-score, particularly for underrepresented named entity categories such as Law, Product, and Facility. This study demonstrates the effectiveness of domain-specific fine-tuning in low-resource languages and highlights the potential for enhancing Catalan NLP applications through manual annotation and high-quality datasets.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:44:19 GMT" } ]	2025-03-19T00:00:00	[ [ "Ferreres", "Guillem Cadevall", "" ], [ "Sanz", "Marc Serrano", "" ], [ "Gámez", "Marc Bardeli", "" ], [ "Basullas", "Pol Gerdt", "" ], [ "Ruiz", "Francesc Tarres", "" ], [ "Ferrero", "Raul Quijada", "" ] ]	TITLE: NERCat: Fine-Tuning for Enhanced Named Entity Recognition in Catalan ABSTRACT: Named Entity Recognition (NER) is a critical component of Natural Language Processing (NLP) for extracting structured information from unstructured text. However, for low-resource languages like Catalan, the performance of NER systems often suffers due to the lack of high-quality annotated datasets. This paper introduces NERCat, a fine-tuned version of the GLiNER[1] model, designed to improve NER performance specifically for Catalan text. We used a dataset of manually annotated Catalan television transcriptions to train and fine-tune the model, focusing on domains such as politics, sports, and culture. The evaluation results show significant improvements in precision, recall, and F1-score, particularly for underrepresented named entity categories such as Law, Product, and Facility. This study demonstrates the effectiveness of domain-specific fine-tuning in low-resource languages and highlights the potential for enhancing Catalan NLP applications through manual annotation and high-quality datasets.
2503.14182	Bozhou Zhang	Bozhou Zhang, Nan Song, Xin Jin, Li Zhang	Bridging Past and Future: End-to-End Autonomous Driving with Historical Prediction and Planning	CVPR 2025	null	null	null	cs.RO cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	End-to-end autonomous driving unifies tasks in a differentiable framework, enabling planning-oriented optimization and attracting growing attention. Current methods aggregate historical information either through dense historical bird's-eye-view (BEV) features or by querying a sparse memory bank, following paradigms inherited from detection. However, we argue that these paradigms either omit historical information in motion planning or fail to align with its multi-step nature, which requires predicting or planning multiple future time steps. In line with the philosophy of future is a continuation of past, we propose BridgeAD, which reformulates motion and planning queries as multi-step queries to differentiate the queries for each future time step. This design enables the effective use of historical prediction and planning by applying them to the appropriate parts of the end-to-end system based on the time steps, which improves both perception and motion planning. Specifically, historical queries for the current frame are combined with perception, while queries for future frames are integrated with motion planning. In this way, we bridge the gap between past and future by aggregating historical insights at every time step, enhancing the overall coherence and accuracy of the end-to-end autonomous driving pipeline. Extensive experiments on the nuScenes dataset in both open-loop and closed-loop settings demonstrate that BridgeAD achieves state-of-the-art performance.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:57:31 GMT" } ]	2025-03-19T00:00:00	[ [ "Zhang", "Bozhou", "" ], [ "Song", "Nan", "" ], [ "Jin", "Xin", "" ], [ "Zhang", "Li", "" ] ]	TITLE: Bridging Past and Future: End-to-End Autonomous Driving with Historical Prediction and Planning ABSTRACT: End-to-end autonomous driving unifies tasks in a differentiable framework, enabling planning-oriented optimization and attracting growing attention. Current methods aggregate historical information either through dense historical bird's-eye-view (BEV) features or by querying a sparse memory bank, following paradigms inherited from detection. However, we argue that these paradigms either omit historical information in motion planning or fail to align with its multi-step nature, which requires predicting or planning multiple future time steps. In line with the philosophy of future is a continuation of past, we propose BridgeAD, which reformulates motion and planning queries as multi-step queries to differentiate the queries for each future time step. This design enables the effective use of historical prediction and planning by applying them to the appropriate parts of the end-to-end system based on the time steps, which improves both perception and motion planning. Specifically, historical queries for the current frame are combined with perception, while queries for future frames are integrated with motion planning. In this way, we bridge the gap between past and future by aggregating historical insights at every time step, enhancing the overall coherence and accuracy of the end-to-end autonomous driving pipeline. Extensive experiments on the nuScenes dataset in both open-loop and closed-loop settings demonstrate that BridgeAD achieves state-of-the-art performance.
2503.14183	Ekaterina Verbitskaia	Aleksandr Shefer, Igor Engel, Stanislav Alekseev, Daniil Berezun, Ekaterina Verbitskaia, Anton Podkopaev	Can LLMs Enable Verification in Mainstream Programming?	null	null	null	null	cs.SE cs.AI cs.PL	http://creativecommons.org/licenses/by/4.0/	Although formal methods are capable of producing reliable software, they have seen minimal adoption in everyday programming. Automatic code generation using large language models is becoming increasingly widespread, but it rarely considers producing strong correctness guarantees. In this study, we explore the ability of LLMs to produce verified code in three verification languages (Dafny, Nagini, and Verus). To do so, we use manually curated datasets derived from the state-ofthe-art Python benchmark, HumanEval. We also assess what types of information are sufficient to achieve good-quality results.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:58:00 GMT" } ]	2025-03-19T00:00:00	[ [ "Shefer", "Aleksandr", "" ], [ "Engel", "Igor", "" ], [ "Alekseev", "Stanislav", "" ], [ "Berezun", "Daniil", "" ], [ "Verbitskaia", "Ekaterina", "" ], [ "Podkopaev", "Anton", "" ] ]	TITLE: Can LLMs Enable Verification in Mainstream Programming? ABSTRACT: Although formal methods are capable of producing reliable software, they have seen minimal adoption in everyday programming. Automatic code generation using large language models is becoming increasingly widespread, but it rarely considers producing strong correctness guarantees. In this study, we explore the ability of LLMs to produce verified code in three verification languages (Dafny, Nagini, and Verus). To do so, we use manually curated datasets derived from the state-ofthe-art Python benchmark, HumanEval. We also assess what types of information are sufficient to achieve good-quality results.
2503.14185	Wuwei Huang	Wuwei Huang, Dexin Wang, Deyi Xiong	AdaST: Dynamically Adapting Encoder States in the Decoder for End-to-End Speech-to-Text Translation	ACL 2021 Findings	null	null	null	cs.CL cs.SD eess.AS	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In end-to-end speech translation, acoustic representations learned by the encoder are usually fixed and static, from the perspective of the decoder, which is not desirable for dealing with the cross-modal and cross-lingual challenge in speech translation. In this paper, we show the benefits of varying acoustic states according to decoder hidden states and propose an adaptive speech-to-text translation model that is able to dynamically adapt acoustic states in the decoder. We concatenate the acoustic state and target word embedding sequence and feed the concatenated sequence into subsequent blocks in the decoder. In order to model the deep interaction between acoustic states and target hidden states, a speech-text mixed attention sublayer is introduced to replace the conventional cross-attention network. Experiment results on two widely-used datasets show that the proposed method significantly outperforms state-of-the-art neural speech translation models.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 11:59:27 GMT" } ]	2025-03-19T00:00:00	[ [ "Huang", "Wuwei", "" ], [ "Wang", "Dexin", "" ], [ "Xiong", "Deyi", "" ] ]	TITLE: AdaST: Dynamically Adapting Encoder States in the Decoder for End-to-End Speech-to-Text Translation ABSTRACT: In end-to-end speech translation, acoustic representations learned by the encoder are usually fixed and static, from the perspective of the decoder, which is not desirable for dealing with the cross-modal and cross-lingual challenge in speech translation. In this paper, we show the benefits of varying acoustic states according to decoder hidden states and propose an adaptive speech-to-text translation model that is able to dynamically adapt acoustic states in the decoder. We concatenate the acoustic state and target word embedding sequence and feed the concatenated sequence into subsequent blocks in the decoder. In order to model the deep interaction between acoustic states and target hidden states, a speech-text mixed attention sublayer is introduced to replace the conventional cross-attention network. Experiment results on two widely-used datasets show that the proposed method significantly outperforms state-of-the-art neural speech translation models.
2503.14189	Yongqi Li	Yongqi Li, Lu Yang, Jian Wang, Runyang You, Wenjie Li, Liqiang Nie	Towards Harmless Multimodal Assistants with Blind Preference Optimization	null	null	null	null	cs.CL cs.CV	http://creativecommons.org/licenses/by/4.0/	Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in multimodal understanding, reasoning, and interaction. Given the extensive applications of MLLMs, the associated safety issues have become increasingly critical. Due to the effectiveness of preference optimization in aligning MLLMs with human preferences, there is an urgent need for safety-related preference data for MLLMs. To address this, we construct the MMSafe-PO preference dataset towards harmless multimodal assistants, featuring multimodal instructions, the conversational format, and ranked paired responses from human feedback. We also identify two insightful observations: modality co-defense and modality cheating, which illustrate that MLLMs possess a certain level of inherent defense while still presenting unique safety challenges. Based on these observations, we propose the Blind Preference Optimization (BPO) approach. Comprehensive experiments on three benchmarks show that BPO effectively enhances the safety capabilities of MLLMs. Notably, BPO significantly improves the safety rate of the base MLLM by 45.0%, outperforming the DPO approach. Additionally, applying BPO to the MMSafe-PO dataset greatly reduces the base MLLM's unsafe rate on other safety benchmarks (14.5% on MM-SafetyBench and 82.9% on HarmEval, demonstrating the effectiveness and robustness of both the dataset and the approach. We release code and data at https://lu-yang666.github.io/MMsafe-PO-Web/.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:02:38 GMT" } ]	2025-03-19T00:00:00	[ [ "Li", "Yongqi", "" ], [ "Yang", "Lu", "" ], [ "Wang", "Jian", "" ], [ "You", "Runyang", "" ], [ "Li", "Wenjie", "" ], [ "Nie", "Liqiang", "" ] ]	TITLE: Towards Harmless Multimodal Assistants with Blind Preference Optimization ABSTRACT: Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in multimodal understanding, reasoning, and interaction. Given the extensive applications of MLLMs, the associated safety issues have become increasingly critical. Due to the effectiveness of preference optimization in aligning MLLMs with human preferences, there is an urgent need for safety-related preference data for MLLMs. To address this, we construct the MMSafe-PO preference dataset towards harmless multimodal assistants, featuring multimodal instructions, the conversational format, and ranked paired responses from human feedback. We also identify two insightful observations: modality co-defense and modality cheating, which illustrate that MLLMs possess a certain level of inherent defense while still presenting unique safety challenges. Based on these observations, we propose the Blind Preference Optimization (BPO) approach. Comprehensive experiments on three benchmarks show that BPO effectively enhances the safety capabilities of MLLMs. Notably, BPO significantly improves the safety rate of the base MLLM by 45.0%, outperforming the DPO approach. Additionally, applying BPO to the MMSafe-PO dataset greatly reduces the base MLLM's unsafe rate on other safety benchmarks (14.5% on MM-SafetyBench and 82.9% on HarmEval, demonstrating the effectiveness and robustness of both the dataset and the approach. We release code and data at https://lu-yang666.github.io/MMsafe-PO-Web/.
2503.14194	Yilin Wang	Yilin Wang	Driving behavior recognition via self-discovery learning	9 pages	null	null	null	cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Autonomous driving systems require a deep understanding of human driving behaviors to achieve higher intelligence and safety.Despite advancements in deep learning, challenges such as long-tail distribution due to scarce samples and confusion from similar behaviors hinder effective driving behavior detection.Existing methods often fail to address sample confusion adequately, as datasets frequently contain ambiguous samples that obscure unique semantic information.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:13:08 GMT" } ]	2025-03-19T00:00:00	[ [ "Wang", "Yilin", "" ] ]	TITLE: Driving behavior recognition via self-discovery learning ABSTRACT: Autonomous driving systems require a deep understanding of human driving behaviors to achieve higher intelligence and safety.Despite advancements in deep learning, challenges such as long-tail distribution due to scarce samples and confusion from similar behaviors hinder effective driving behavior detection.Existing methods often fail to address sample confusion adequately, as datasets frequently contain ambiguous samples that obscure unique semantic information.
2503.14198	Junjin Xiao	Junjin Xiao, Qing Zhang, Yonewei Nie, Lei Zhu, Wei-Shi Zheng	RoGSplat: Learning Robust Generalizable Human Gaussian Splatting from Sparse Multi-View Images	Accepted to CVPR2025	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper presents RoGSplat, a novel approach for synthesizing high-fidelity novel views of unseen human from sparse multi-view images, while requiring no cumbersome per-subject optimization. Unlike previous methods that typically struggle with sparse views with few overlappings and are less effective in reconstructing complex human geometry, the proposed method enables robust reconstruction in such challenging conditions. Our key idea is to lift SMPL vertices to dense and reliable 3D prior points representing accurate human body geometry, and then regress human Gaussian parameters based on the points. To account for possible misalignment between SMPL model and images, we propose to predict image-aligned 3D prior points by leveraging both pixel-level features and voxel-level features, from which we regress the coarse Gaussians. To enhance the ability to capture high-frequency details, we further render depth maps from the coarse 3D Gaussians to help regress fine-grained pixel-wise Gaussians. Experiments on several benchmark datasets demonstrate that our method outperforms state-of-the-art methods in novel view synthesis and cross-dataset generalization. Our code is available at https://github.com/iSEE-Laboratory/RoGSplat.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:18:34 GMT" } ]	2025-03-19T00:00:00	[ [ "Xiao", "Junjin", "" ], [ "Zhang", "Qing", "" ], [ "Nie", "Yonewei", "" ], [ "Zhu", "Lei", "" ], [ "Zheng", "Wei-Shi", "" ] ]	TITLE: RoGSplat: Learning Robust Generalizable Human Gaussian Splatting from Sparse Multi-View Images ABSTRACT: This paper presents RoGSplat, a novel approach for synthesizing high-fidelity novel views of unseen human from sparse multi-view images, while requiring no cumbersome per-subject optimization. Unlike previous methods that typically struggle with sparse views with few overlappings and are less effective in reconstructing complex human geometry, the proposed method enables robust reconstruction in such challenging conditions. Our key idea is to lift SMPL vertices to dense and reliable 3D prior points representing accurate human body geometry, and then regress human Gaussian parameters based on the points. To account for possible misalignment between SMPL model and images, we propose to predict image-aligned 3D prior points by leveraging both pixel-level features and voxel-level features, from which we regress the coarse Gaussians. To enhance the ability to capture high-frequency details, we further render depth maps from the coarse 3D Gaussians to help regress fine-grained pixel-wise Gaussians. Experiments on several benchmark datasets demonstrate that our method outperforms state-of-the-art methods in novel view synthesis and cross-dataset generalization. Our code is available at https://github.com/iSEE-Laboratory/RoGSplat.
2503.14201	Alberto Martin-Lopez	Alessandro Giagnorio, Alberto Martin-Lopez, Gabriele Bavota	Why Personalizing Deep Learning-Based Code Completion Tools Matters	Accepted for publication at ACM TOSEM	null	null	null	cs.SE	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Deep learning (DL)-based code completion tools have transformed software development by enabling advanced code generation. These tools leverage models trained on vast amounts of code from numerous repositories, capturing general coding patterns. However, the impact of fine-tuning these models for specific organizations or developers to boost their performance on such subjects remains unexplored. In this work, we fill this gap by presenting solid empirical evidence answering this question. More specifically, we consider 136 developers from two organizations (Apache and Spring), two model architectures (T5 and Code Llama), and three model sizes (60M, 750M, and 7B trainable parameters). T5 models (60M, 750M) were pre-trained and fine-tuned on over 2,000 open-source projects, excluding the subject organizations' data, and compared against versions fine-tuned on organization- and developer-specific datasets. For the Code Llama model (7B), we compared the performance of the already pre-trained model publicly available online with the same model fine-tuned via parameter-efficient fine-tuning on organization- and developer-specific datasets. Our results show that there is a boost in prediction capabilities provided by both an organization-specific and a developer-specific additional fine-tuning, with the former being particularly performant. Such a finding generalizes across (i) the two subject organizations (i.e., Apache and Spring) and (ii) models of completely different magnitude (from 60M to 7B trainable parameters). Finally, we show that DL models fine-tuned on an organization-specific dataset achieve the same completion performance of pre-trained code models used out of the box and being $\sim$10$\times$ larger, with consequent savings in terms of deployment and inference cost (e.g., smaller GPUs needed).	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:26:06 GMT" } ]	2025-03-19T00:00:00	[ [ "Giagnorio", "Alessandro", "" ], [ "Martin-Lopez", "Alberto", "" ], [ "Bavota", "Gabriele", "" ] ]	TITLE: Why Personalizing Deep Learning-Based Code Completion Tools Matters ABSTRACT: Deep learning (DL)-based code completion tools have transformed software development by enabling advanced code generation. These tools leverage models trained on vast amounts of code from numerous repositories, capturing general coding patterns. However, the impact of fine-tuning these models for specific organizations or developers to boost their performance on such subjects remains unexplored. In this work, we fill this gap by presenting solid empirical evidence answering this question. More specifically, we consider 136 developers from two organizations (Apache and Spring), two model architectures (T5 and Code Llama), and three model sizes (60M, 750M, and 7B trainable parameters). T5 models (60M, 750M) were pre-trained and fine-tuned on over 2,000 open-source projects, excluding the subject organizations' data, and compared against versions fine-tuned on organization- and developer-specific datasets. For the Code Llama model (7B), we compared the performance of the already pre-trained model publicly available online with the same model fine-tuned via parameter-efficient fine-tuning on organization- and developer-specific datasets. Our results show that there is a boost in prediction capabilities provided by both an organization-specific and a developer-specific additional fine-tuning, with the former being particularly performant. Such a finding generalizes across (i) the two subject organizations (i.e., Apache and Spring) and (ii) models of completely different magnitude (from 60M to 7B trainable parameters). Finally, we show that DL models fine-tuned on an organization-specific dataset achieve the same completion performance of pre-trained code models used out of the box and being $\sim$10$\times$ larger, with consequent savings in terms of deployment and inference cost (e.g., smaller GPUs needed).
2503.14209	Saif Ur Rehman Khan Mr	Saif Ur Rehman Khan, Muhammad Nabeel Asim, Sebastian Vollmer, Andreas Dengel	AI-Driven Diabetic Retinopathy Diagnosis Enhancement through Image Processing and Salp Swarm Algorithm-Optimized Ensemble Network	null	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Diabetic retinopathy is a leading cause of blindness in diabetic patients and early detection plays a crucial role in preventing vision loss. Traditional diagnostic methods are often time-consuming and prone to errors. The emergence of deep learning techniques has provided innovative solutions to improve diagnostic efficiency. However, single deep learning models frequently face issues related to extracting key features from complex retinal images. To handle this problem, we present an effective ensemble method for DR diagnosis comprising four main phases: image pre-processing, selection of backbone pre-trained models, feature enhancement, and optimization. Our methodology initiates with the pre-processing phase, where we apply CLAHE to enhance image contrast and Gamma correction is then used to adjust the brightness for better feature recognition. We then apply Discrete Wavelet Transform (DWT) for image fusion by combining multi-resolution details to create a richer dataset. Then, we selected three pre-trained models with the best performance named DenseNet169, MobileNetV1, and Xception for diverse feature extraction. To further improve feature extraction, an improved residual block is integrated into each model. Finally, the predictions from these base models are then aggregated using weighted ensemble approach, with the weights optimized by using Salp Swarm Algorithm (SSA).SSA intelligently explores the weight space and finds the optimal configuration of base architectures to maximize the performance of the ensemble model. The proposed model is evaluated on the multiclass Kaggle APTOS 2019 dataset and obtained 88.52% accuracy.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:35:56 GMT" } ]	2025-03-19T00:00:00	[ [ "Khan", "Saif Ur Rehman", "" ], [ "Asim", "Muhammad Nabeel", "" ], [ "Vollmer", "Sebastian", "" ], [ "Dengel", "Andreas", "" ] ]	TITLE: AI-Driven Diabetic Retinopathy Diagnosis Enhancement through Image Processing and Salp Swarm Algorithm-Optimized Ensemble Network ABSTRACT: Diabetic retinopathy is a leading cause of blindness in diabetic patients and early detection plays a crucial role in preventing vision loss. Traditional diagnostic methods are often time-consuming and prone to errors. The emergence of deep learning techniques has provided innovative solutions to improve diagnostic efficiency. However, single deep learning models frequently face issues related to extracting key features from complex retinal images. To handle this problem, we present an effective ensemble method for DR diagnosis comprising four main phases: image pre-processing, selection of backbone pre-trained models, feature enhancement, and optimization. Our methodology initiates with the pre-processing phase, where we apply CLAHE to enhance image contrast and Gamma correction is then used to adjust the brightness for better feature recognition. We then apply Discrete Wavelet Transform (DWT) for image fusion by combining multi-resolution details to create a richer dataset. Then, we selected three pre-trained models with the best performance named DenseNet169, MobileNetV1, and Xception for diverse feature extraction. To further improve feature extraction, an improved residual block is integrated into each model. Finally, the predictions from these base models are then aggregated using weighted ensemble approach, with the weights optimized by using Salp Swarm Algorithm (SSA).SSA intelligently explores the weight space and finds the optimal configuration of base architectures to maximize the performance of the ensemble model. The proposed model is evaluated on the multiclass Kaggle APTOS 2019 dataset and obtained 88.52% accuracy.
2503.14213	Ashraf Ghiye	Ashraf Ghiye, Baptiste Barreau, Laurent Carlier, Michalis Vazirgiannis	Rolling Forward: Enhancing LightGCN with Causal Graph Convolution for Credit Bond Recommendation	8 pages, published in the international conference for AI in Finance (ACM ICAIF'24)	null	10.1145/3677052.3698683	null	cs.IR cs.LG q-fin.CP	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Graph Neural Networks have significantly advanced research in recommender systems over the past few years. These methods typically capture global interests using aggregated past interactions and rely on static embeddings of users and items over extended periods of time. While effective in some domains, these methods fall short in many real-world scenarios, especially in finance, where user interests and item popularity evolve rapidly over time. To address these challenges, we introduce a novel extension to Light Graph Convolutional Network (LightGCN) designed to learn temporal node embeddings that capture dynamic interests. Our approach employs causal convolution to maintain a forward-looking model architecture. By preserving the chronological order of user-item interactions and introducing a dynamic update mechanism for embeddings through a sliding window, the proposed model generates well-timed and contextually relevant recommendations. Extensive experiments on a real-world dataset from BNP Paribas demonstrate that our approach significantly enhances the performance of LightGCN while maintaining the simplicity and efficiency of its architecture. Our findings provide new insights into designing graph-based recommender systems in time-sensitive applications, particularly for financial product recommendations.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 12:47:01 GMT" } ]	2025-03-19T00:00:00	[ [ "Ghiye", "Ashraf", "" ], [ "Barreau", "Baptiste", "" ], [ "Carlier", "Laurent", "" ], [ "Vazirgiannis", "Michalis", "" ] ]	TITLE: Rolling Forward: Enhancing LightGCN with Causal Graph Convolution for Credit Bond Recommendation ABSTRACT: Graph Neural Networks have significantly advanced research in recommender systems over the past few years. These methods typically capture global interests using aggregated past interactions and rely on static embeddings of users and items over extended periods of time. While effective in some domains, these methods fall short in many real-world scenarios, especially in finance, where user interests and item popularity evolve rapidly over time. To address these challenges, we introduce a novel extension to Light Graph Convolutional Network (LightGCN) designed to learn temporal node embeddings that capture dynamic interests. Our approach employs causal convolution to maintain a forward-looking model architecture. By preserving the chronological order of user-item interactions and introducing a dynamic update mechanism for embeddings through a sliding window, the proposed model generates well-timed and contextually relevant recommendations. Extensive experiments on a real-world dataset from BNP Paribas demonstrate that our approach significantly enhances the performance of LightGCN while maintaining the simplicity and efficiency of its architecture. Our findings provide new insights into designing graph-based recommender systems in time-sensitive applications, particularly for financial product recommendations.
2503.14228	Yasunori Ishii Mr	Nobuhiko Wakai, Satoshi Sato, Yasunori Ishii, Takayoshi Yamashita	Panoramic Distortion-Aware Tokenization for Person Detection and Localization Using Transformers in Overhead Fisheye Images	null	null	null	null	cs.CV cs.AI	http://creativecommons.org/licenses/by/4.0/	Person detection methods are used widely in applications including visual surveillance, pedestrian detection, and robotics. However, accurate detection of persons from overhead fisheye images remains an open challenge because of factors including person rotation and small-sized persons. To address the person rotation problem, we convert the fisheye images into panoramic images. For smaller people, we focused on the geometry of the panoramas. Conventional detection methods tend to focus on larger people because these larger people yield large significant areas for feature maps. In equirectangular panoramic images, we find that a person's height decreases linearly near the top of the images. Using this finding, we leverage the significance values and aggregate tokens that are sorted based on these values to balance the significant areas. In this leveraging process, we introduce panoramic distortion-aware tokenization. This tokenization procedure divides a panoramic image using self-similarity figures that enable determination of optimal divisions without gaps, and we leverage the maximum significant values in each tile of token groups to preserve the significant areas of smaller people. To achieve higher detection accuracy, we propose a person detection and localization method that combines panoramic-image remapping and the tokenization procedure. Extensive experiments demonstrated that our method outperforms conventional methods when applied to large-scale datasets.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 13:05:41 GMT" } ]	2025-03-19T00:00:00	[ [ "Wakai", "Nobuhiko", "" ], [ "Sato", "Satoshi", "" ], [ "Ishii", "Yasunori", "" ], [ "Yamashita", "Takayoshi", "" ] ]	TITLE: Panoramic Distortion-Aware Tokenization for Person Detection and Localization Using Transformers in Overhead Fisheye Images ABSTRACT: Person detection methods are used widely in applications including visual surveillance, pedestrian detection, and robotics. However, accurate detection of persons from overhead fisheye images remains an open challenge because of factors including person rotation and small-sized persons. To address the person rotation problem, we convert the fisheye images into panoramic images. For smaller people, we focused on the geometry of the panoramas. Conventional detection methods tend to focus on larger people because these larger people yield large significant areas for feature maps. In equirectangular panoramic images, we find that a person's height decreases linearly near the top of the images. Using this finding, we leverage the significance values and aggregate tokens that are sorted based on these values to balance the significant areas. In this leveraging process, we introduce panoramic distortion-aware tokenization. This tokenization procedure divides a panoramic image using self-similarity figures that enable determination of optimal divisions without gaps, and we leverage the maximum significant values in each tile of token groups to preserve the significant areas of smaller people. To achieve higher detection accuracy, we propose a person detection and localization method that combines panoramic-image remapping and the tokenization procedure. Extensive experiments demonstrated that our method outperforms conventional methods when applied to large-scale datasets.
2503.14229	Yifei Dong	Yifei Dong, Fengyi Wu, Qi He, Heng Li, Minghan Li, Zebang Cheng, Yuxuan Zhou, Jingdong Sun, Qi Dai, Zhi-Qi Cheng, Alexander G Hauptmann	HA-VLN: A Benchmark for Human-Aware Navigation in Discrete-Continuous Environments with Dynamic Multi-Human Interactions, Real-World Validation, and an Open Leaderboard	27 pages, website: https://ha-vln-project.vercel.app/	null	null	null	cs.AI cs.CV cs.RO	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Vision-and-Language Navigation (VLN) systems often focus on either discrete (panoramic) or continuous (free-motion) paradigms alone, overlooking the complexities of human-populated, dynamic environments. We introduce a unified Human-Aware VLN (HA-VLN) benchmark that merges these paradigms under explicit social-awareness constraints. Our contributions include: 1. A standardized task definition that balances discrete-continuous navigation with personal-space requirements; 2. An enhanced human motion dataset (HAPS 2.0) and upgraded simulators capturing realistic multi-human interactions, outdoor contexts, and refined motion-language alignment; 3. Extensive benchmarking on 16,844 human-centric instructions, revealing how multi-human dynamics and partial observability pose substantial challenges for leading VLN agents; 4. Real-world robot tests validating sim-to-real transfer in crowded indoor spaces; and 5. A public leaderboard supporting transparent comparisons across discrete and continuous tasks. Empirical results show improved navigation success and fewer collisions when social context is integrated, underscoring the need for human-centric design. By releasing all datasets, simulators, agent code, and evaluation tools, we aim to advance safer, more capable, and socially responsible VLN research.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 13:05:55 GMT" } ]	2025-03-19T00:00:00	[ [ "Dong", "Yifei", "" ], [ "Wu", "Fengyi", "" ], [ "He", "Qi", "" ], [ "Li", "Heng", "" ], [ "Li", "Minghan", "" ], [ "Cheng", "Zebang", "" ], [ "Zhou", "Yuxuan", "" ], [ "Sun", "Jingdong", "" ], [ "Dai", "Qi", "" ], [ "Cheng", "Zhi-Qi", "" ], [ "Hauptmann", "Alexander G", "" ] ]	TITLE: HA-VLN: A Benchmark for Human-Aware Navigation in Discrete-Continuous Environments with Dynamic Multi-Human Interactions, Real-World Validation, and an Open Leaderboard ABSTRACT: Vision-and-Language Navigation (VLN) systems often focus on either discrete (panoramic) or continuous (free-motion) paradigms alone, overlooking the complexities of human-populated, dynamic environments. We introduce a unified Human-Aware VLN (HA-VLN) benchmark that merges these paradigms under explicit social-awareness constraints. Our contributions include: 1. A standardized task definition that balances discrete-continuous navigation with personal-space requirements; 2. An enhanced human motion dataset (HAPS 2.0) and upgraded simulators capturing realistic multi-human interactions, outdoor contexts, and refined motion-language alignment; 3. Extensive benchmarking on 16,844 human-centric instructions, revealing how multi-human dynamics and partial observability pose substantial challenges for leading VLN agents; 4. Real-world robot tests validating sim-to-real transfer in crowded indoor spaces; and 5. A public leaderboard supporting transparent comparisons across discrete and continuous tasks. Empirical results show improved navigation success and fewer collisions when social context is integrated, underscoring the need for human-centric design. By releasing all datasets, simulators, agent code, and evaluation tools, we aim to advance safer, more capable, and socially responsible VLN research.
2503.14231	Giovanni Delnevo	Ziyao Ling, Giovanni Delnevo, Paola Salomoni, Silvia Mirri	Multi-task Learning for Identification of Porcelain in Song and Yuan Dynasties	null	null	null	null	cs.CV cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Chinese porcelain holds immense historical and cultural value, making its accurate classification essential for archaeological research and cultural heritage preservation. Traditional classification methods rely heavily on expert analysis, which is time-consuming, subjective, and difficult to scale. This paper explores the application of DL and transfer learning techniques to automate the classification of porcelain artifacts across four key attributes: dynasty, glaze, ware, and type. We evaluate four Convolutional Neural Networks (CNNs) - ResNet50, MobileNetV2, VGG16, and InceptionV3 - comparing their performance with and without pre-trained weights. Our results demonstrate that transfer learning significantly enhances classification accuracy, particularly for complex tasks like type classification, where models trained from scratch exhibit lower performance. MobileNetV2 and ResNet50 consistently achieve high accuracy and robustness across all tasks, while VGG16 struggles with more diverse classifications. We further discuss the impact of dataset limitations and propose future directions, including domain-specific pre-training, integration of attention mechanisms, explainable AI methods, and generalization to other cultural artifacts.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 13:09:00 GMT" } ]	2025-03-19T00:00:00	[ [ "Ling", "Ziyao", "" ], [ "Delnevo", "Giovanni", "" ], [ "Salomoni", "Paola", "" ], [ "Mirri", "Silvia", "" ] ]	TITLE: Multi-task Learning for Identification of Porcelain in Song and Yuan Dynasties ABSTRACT: Chinese porcelain holds immense historical and cultural value, making its accurate classification essential for archaeological research and cultural heritage preservation. Traditional classification methods rely heavily on expert analysis, which is time-consuming, subjective, and difficult to scale. This paper explores the application of DL and transfer learning techniques to automate the classification of porcelain artifacts across four key attributes: dynasty, glaze, ware, and type. We evaluate four Convolutional Neural Networks (CNNs) - ResNet50, MobileNetV2, VGG16, and InceptionV3 - comparing their performance with and without pre-trained weights. Our results demonstrate that transfer learning significantly enhances classification accuracy, particularly for complex tasks like type classification, where models trained from scratch exhibit lower performance. MobileNetV2 and ResNet50 consistently achieve high accuracy and robustness across all tasks, while VGG16 struggles with more diverse classifications. We further discuss the impact of dataset limitations and propose future directions, including domain-specific pre-training, integration of attention mechanisms, explainable AI methods, and generalization to other cultural artifacts.
2503.14247	Tingyang Xiao	Tingyang Xiao, Xiaolin Zhou, Liu Liu, Wei Sui, Wei Feng, Jiaxiong Qiu, Xinjie Wang, and Zhizhong Su	GeoFlow-SLAM: A Robust Tightly-Coupled RGBD-Inertial Fusion SLAM for Dynamic Legged Robotics	8 pages	null	null	null	cs.RO cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper presents GeoFlow-SLAM, a robust and effective Tightly-Coupled RGBD-inertial SLAM for legged robots operating in highly dynamic environments.By integrating geometric consistency, legged odometry constraints, and dual-stream optical flow (GeoFlow), our method addresses three critical challenges:feature matching and pose initialization failures during fast locomotion and visual feature scarcity in texture-less scenes.Specifically, in rapid motion scenarios, feature matching is notably enhanced by leveraging dual-stream optical flow, which combines prior map points and poses. Additionally, we propose a robust pose initialization method for fast locomotion and IMU error in legged robots, integrating IMU/Legged odometry, inter-frame Perspective-n-Point (PnP), and Generalized Iterative Closest Point (GICP). Furthermore, a novel optimization framework that tightly couples depth-to-map and GICP geometric constraints is first introduced to improve the robustness and accuracy in long-duration, visually texture-less environments. The proposed algorithms achieve state-of-the-art (SOTA) on collected legged robots and open-source datasets. To further promote research and development, the open-source datasets and code will be made publicly available at https://github.com/NSN-Hello/GeoFlow-SLAM	[ { "version": "v1", "created": "Tue, 18 Mar 2025 13:35:49 GMT" } ]	2025-03-19T00:00:00	[ [ "Xiao", "Tingyang", "" ], [ "Zhou", "Xiaolin", "" ], [ "Liu", "Liu", "" ], [ "Sui", "Wei", "" ], [ "Feng", "Wei", "" ], [ "Qiu", "Jiaxiong", "" ], [ "Wang", "Xinjie", "" ], [ "Su", "Zhizhong", "" ] ]	TITLE: GeoFlow-SLAM: A Robust Tightly-Coupled RGBD-Inertial Fusion SLAM for Dynamic Legged Robotics ABSTRACT: This paper presents GeoFlow-SLAM, a robust and effective Tightly-Coupled RGBD-inertial SLAM for legged robots operating in highly dynamic environments.By integrating geometric consistency, legged odometry constraints, and dual-stream optical flow (GeoFlow), our method addresses three critical challenges:feature matching and pose initialization failures during fast locomotion and visual feature scarcity in texture-less scenes.Specifically, in rapid motion scenarios, feature matching is notably enhanced by leveraging dual-stream optical flow, which combines prior map points and poses. Additionally, we propose a robust pose initialization method for fast locomotion and IMU error in legged robots, integrating IMU/Legged odometry, inter-frame Perspective-n-Point (PnP), and Generalized Iterative Closest Point (GICP). Furthermore, a novel optimization framework that tightly couples depth-to-map and GICP geometric constraints is first introduced to improve the robustness and accuracy in long-duration, visually texture-less environments. The proposed algorithms achieve state-of-the-art (SOTA) on collected legged robots and open-source datasets. To further promote research and development, the open-source datasets and code will be made publicly available at https://github.com/NSN-Hello/GeoFlow-SLAM
2503.14284	Jiacen Xu	Jiacen Xu, Chenang Li, Yu Zheng, Zhou Li	Entente: Cross-silo Intrusion Detection on Network Log Graphs with Federated Learning	null	null	null	null	cs.CR	http://creativecommons.org/licenses/by-nc-nd/4.0/	Graph-based Network Intrusion Detection System (GNIDS) has gained significant momentum in detecting sophisticated cyber-attacks, like Advanced Persistent Threat (APT), in an organization or across organizations. Though achieving satisfying detection accuracy and adapting to ever-changing attacks and normal patterns, all prior GNIDSs assume the centralized data settings directly, but non-trivial data collection is not always practical under privacy regulations nowadays. We argue that training a GNIDS model has to consider privacy regulations, and propose to leverage federated learning (FL) to address this prominent challenge. Yet, directly applying FL to GNIDS is unlikely to succeed, due to issues like non-IID (independent and identically distributed) graph data over clients and the diverse design choices taken by different GNIDS. We address these issues with a set of novel techniques tailored to the graph datasets, including reference graph synthesis, graph sketching and adaptive contribution scaling, and develop a new system Entente. We evaluate Entente on the large-scale LANL, OpTC and Pivoting datasets. The result shows Entente outperforms the other baseline FL algorithms and sometimes even the non-FL GNIDS. We also evaluate Entente under FL poisoning attacks tailored to the GNIDS setting, and show Entente is able to bound the attack success rate to low values. Overall, our result suggests building cross-silo GNIDS is feasible and we hope to encourage more efforts in this direction.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 14:21:24 GMT" } ]	2025-03-19T00:00:00	[ [ "Xu", "Jiacen", "" ], [ "Li", "Chenang", "" ], [ "Zheng", "Yu", "" ], [ "Li", "Zhou", "" ] ]	TITLE: Entente: Cross-silo Intrusion Detection on Network Log Graphs with Federated Learning ABSTRACT: Graph-based Network Intrusion Detection System (GNIDS) has gained significant momentum in detecting sophisticated cyber-attacks, like Advanced Persistent Threat (APT), in an organization or across organizations. Though achieving satisfying detection accuracy and adapting to ever-changing attacks and normal patterns, all prior GNIDSs assume the centralized data settings directly, but non-trivial data collection is not always practical under privacy regulations nowadays. We argue that training a GNIDS model has to consider privacy regulations, and propose to leverage federated learning (FL) to address this prominent challenge. Yet, directly applying FL to GNIDS is unlikely to succeed, due to issues like non-IID (independent and identically distributed) graph data over clients and the diverse design choices taken by different GNIDS. We address these issues with a set of novel techniques tailored to the graph datasets, including reference graph synthesis, graph sketching and adaptive contribution scaling, and develop a new system Entente. We evaluate Entente on the large-scale LANL, OpTC and Pivoting datasets. The result shows Entente outperforms the other baseline FL algorithms and sometimes even the non-FL GNIDS. We also evaluate Entente under FL poisoning attacks tailored to the GNIDS setting, and show Entente is able to bound the attack success rate to low values. Overall, our result suggests building cross-silo GNIDS is feasible and we hope to encourage more efforts in this direction.
2503.14287	Enrico Tosi	Enrico Tosi, Panwei Hu, Aleksandar Ichkov, Marina Petrova, Ljiljana Simi\'c	Cross-Environment Transfer Learning for Location-Aided Beam Prediction in 5G and Beyond Millimeter-Wave Networks	null	null	null	null	eess.SP cs.SY eess.SY	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Millimeter-wave (mm-wave) communications requirebeamforming and consequent precise beam alignmentbetween the gNodeB (gNB) and the user equipment (UE) toovercome high propagation losses. This beam alignment needs tobe constantly updated for different UE locations based on beamsweepingradio frequency measurements, leading to significantbeam management overhead. One potential solution involvesusing machine learning (ML) beam prediction algorithms thatleverage UE position information to select the serving beamwithout the overhead of beam sweeping. However, the highlysite-specific nature of mm-wave propagation means that MLmodels require training from scratch for each scenario, whichis inefficient in practice. In this paper, we propose a robustcross-environment transfer learning solution for location-aidedbeam prediction, whereby the ML model trained on a referencegNB is transferred to a target gNB by fine-tuning with a limiteddataset. Extensive simulation results based on ray-tracing in twourban environments show the effectiveness of our solution forboth inter- and intra-city model transfer. Our results show thatby training the model on a reference gNB and transferring themodel by fine-tuning with only 5% of the target gNB dataset,we can achieve 80% accuracy in predicting the best beamfor the target gNB. Importantly, our approach improves thepoor generalization accuracy of transferring the model to newenvironments without fine-tuning by around 75 percentage points.This demonstrates that transfer learning enables high predictionaccuracy while reducing the computational and training datasetcollection burden of ML-based beam prediction, making itpractical for 5G-and-beyond deployments.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 14:24:50 GMT" } ]	2025-03-19T00:00:00	[ [ "Tosi", "Enrico", "" ], [ "Hu", "Panwei", "" ], [ "Ichkov", "Aleksandar", "" ], [ "Petrova", "Marina", "" ], [ "Simić", "Ljiljana", "" ] ]	TITLE: Cross-Environment Transfer Learning for Location-Aided Beam Prediction in 5G and Beyond Millimeter-Wave Networks ABSTRACT: Millimeter-wave (mm-wave) communications requirebeamforming and consequent precise beam alignmentbetween the gNodeB (gNB) and the user equipment (UE) toovercome high propagation losses. This beam alignment needs tobe constantly updated for different UE locations based on beamsweepingradio frequency measurements, leading to significantbeam management overhead. One potential solution involvesusing machine learning (ML) beam prediction algorithms thatleverage UE position information to select the serving beamwithout the overhead of beam sweeping. However, the highlysite-specific nature of mm-wave propagation means that MLmodels require training from scratch for each scenario, whichis inefficient in practice. In this paper, we propose a robustcross-environment transfer learning solution for location-aidedbeam prediction, whereby the ML model trained on a referencegNB is transferred to a target gNB by fine-tuning with a limiteddataset. Extensive simulation results based on ray-tracing in twourban environments show the effectiveness of our solution forboth inter- and intra-city model transfer. Our results show thatby training the model on a reference gNB and transferring themodel by fine-tuning with only 5% of the target gNB dataset,we can achieve 80% accuracy in predicting the best beamfor the target gNB. Importantly, our approach improves thepoor generalization accuracy of transferring the model to newenvironments without fine-tuning by around 75 percentage points.This demonstrates that transfer learning enables high predictionaccuracy while reducing the computational and training datasetcollection burden of ML-based beam prediction, making itpractical for 5G-and-beyond deployments.
2503.14304	Yuheng Li	Yuheng Li, Mingzhe Hu, Richard L.J. Qiu, Maria Thor, Andre Williams, Deborah Marshall and Xiaofeng Yang	RoMedFormer: A Rotary-Embedding Transformer Foundation Model for 3D Genito-Pelvic Structure Segmentation in MRI and CT	null	null	null	null	eess.IV cs.CV	http://creativecommons.org/licenses/by/4.0/	Deep learning-based segmentation of genito-pelvic structures in MRI and CT is crucial for applications such as radiation therapy, surgical planning, and disease diagnosis. However, existing segmentation models often struggle with generalizability across imaging modalities, and anatomical variations. In this work, we propose RoMedFormer, a rotary-embedding transformer-based foundation model designed for 3D female genito-pelvic structure segmentation in both MRI and CT. RoMedFormer leverages self-supervised learning and rotary positional embeddings to enhance spatial feature representation and capture long-range dependencies in 3D medical data. We pre-train our model using a diverse dataset of 3D MRI and CT scans and fine-tune it for downstream segmentation tasks. Experimental results demonstrate that RoMedFormer achieves superior performance segmenting genito-pelvic organs. Our findings highlight the potential of transformer-based architectures in medical image segmentation and pave the way for more transferable segmentation frameworks.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 14:45:05 GMT" } ]	2025-03-19T00:00:00	[ [ "Li", "Yuheng", "" ], [ "Hu", "Mingzhe", "" ], [ "Qiu", "Richard L. J.", "" ], [ "Thor", "Maria", "" ], [ "Williams", "Andre", "" ], [ "Marshall", "Deborah", "" ], [ "Yang", "Xiaofeng", "" ] ]	TITLE: RoMedFormer: A Rotary-Embedding Transformer Foundation Model for 3D Genito-Pelvic Structure Segmentation in MRI and CT ABSTRACT: Deep learning-based segmentation of genito-pelvic structures in MRI and CT is crucial for applications such as radiation therapy, surgical planning, and disease diagnosis. However, existing segmentation models often struggle with generalizability across imaging modalities, and anatomical variations. In this work, we propose RoMedFormer, a rotary-embedding transformer-based foundation model designed for 3D female genito-pelvic structure segmentation in both MRI and CT. RoMedFormer leverages self-supervised learning and rotary positional embeddings to enhance spatial feature representation and capture long-range dependencies in 3D medical data. We pre-train our model using a diverse dataset of 3D MRI and CT scans and fine-tune it for downstream segmentation tasks. Experimental results demonstrate that RoMedFormer achieves superior performance segmenting genito-pelvic organs. Our findings highlight the potential of transformer-based architectures in medical image segmentation and pave the way for more transferable segmentation frameworks.
2503.14322	Florian Heinrichs	Tiago Vasconcelos Afonso, Florian Heinrichs	Consumer-grade EEG-based Eye Tracking	Data descriptor, 13 pages, 8 figures, 5 tables	null	null	null	eess.SP cs.HC cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Electroencephalography-based eye tracking (EEG-ET) leverages eye movement artifacts in EEG signals as an alternative to camera-based tracking. While EEG-ET offers advantages such as robustness in low-light conditions and better integration with brain-computer interfaces, its development lags behind traditional methods, particularly in consumer-grade settings. To support research in this area, we present a dataset comprising simultaneous EEG and eye-tracking recordings from 113 participants across 116 sessions, amounting to 11 hours and 45 minutes of recordings. Data was collected using a consumer-grade EEG headset and webcam-based eye tracking, capturing eye movements under four experimental paradigms with varying complexity. The dataset enables the evaluation of EEG-ET methods across different gaze conditions and serves as a benchmark for assessing feasibility with affordable hardware. Data preprocessing includes handling of missing values and filtering to enhance usability. In addition to the dataset, code for data preprocessing and analysis is available to support reproducibility and further research.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 14:53:20 GMT" } ]	2025-03-19T00:00:00	[ [ "Afonso", "Tiago Vasconcelos", "" ], [ "Heinrichs", "Florian", "" ] ]	TITLE: Consumer-grade EEG-based Eye Tracking ABSTRACT: Electroencephalography-based eye tracking (EEG-ET) leverages eye movement artifacts in EEG signals as an alternative to camera-based tracking. While EEG-ET offers advantages such as robustness in low-light conditions and better integration with brain-computer interfaces, its development lags behind traditional methods, particularly in consumer-grade settings. To support research in this area, we present a dataset comprising simultaneous EEG and eye-tracking recordings from 113 participants across 116 sessions, amounting to 11 hours and 45 minutes of recordings. Data was collected using a consumer-grade EEG headset and webcam-based eye tracking, capturing eye movements under four experimental paradigms with varying complexity. The dataset enables the evaluation of EEG-ET methods across different gaze conditions and serves as a benchmark for assessing feasibility with affordable hardware. Data preprocessing includes handling of missing values and filtering to enhance usability. In addition to the dataset, code for data preprocessing and analysis is available to support reproducibility and further research.
2503.14346	Xavier Anad\'on	X. Anad\'on, Javier Rodr\'iguez-Puigvert, J.M.M. Montiel	3D Densification for Multi-Map Monocular VSLAM in Endoscopy	null	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	Multi-map Sparse Monocular visual Simultaneous Localization and Mapping applied to monocular endoscopic sequences has proven efficient to robustly recover tracking after the frequent losses in endoscopy due to motion blur, temporal occlusion, tools interaction or water jets. The sparse multi-maps are adequate for robust camera localization, however they are very poor for environment representation, they are noisy, with a high percentage of inaccurately reconstructed 3D points, including significant outliers, and more importantly with an unacceptable low density for clinical applications. We propose a method to remove outliers and densify the maps of the state of the art for sparse endoscopy multi-map CudaSIFT-SLAM. The NN LightDepth for up-to-scale depth dense predictions are aligned with the sparse CudaSIFT submaps by means of the robust to spurious LMedS. Our system mitigates the inherent scale ambiguity in monocular depth estimation while filtering outliers, leading to reliable densified 3D maps. We provide experimental evidence of accurate densified maps 4.15 mm RMS accuracy at affordable computing time in the C3VD phantom colon dataset. We report qualitative results on the real colonoscopy from the Endomapper dataset.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:25:38 GMT" } ]	2025-03-19T00:00:00	[ [ "Anadón", "X.", "" ], [ "Rodríguez-Puigvert", "Javier", "" ], [ "Montiel", "J. M. M.", "" ] ]	TITLE: 3D Densification for Multi-Map Monocular VSLAM in Endoscopy ABSTRACT: Multi-map Sparse Monocular visual Simultaneous Localization and Mapping applied to monocular endoscopic sequences has proven efficient to robustly recover tracking after the frequent losses in endoscopy due to motion blur, temporal occlusion, tools interaction or water jets. The sparse multi-maps are adequate for robust camera localization, however they are very poor for environment representation, they are noisy, with a high percentage of inaccurately reconstructed 3D points, including significant outliers, and more importantly with an unacceptable low density for clinical applications. We propose a method to remove outliers and densify the maps of the state of the art for sparse endoscopy multi-map CudaSIFT-SLAM. The NN LightDepth for up-to-scale depth dense predictions are aligned with the sparse CudaSIFT submaps by means of the robust to spurious LMedS. Our system mitigates the inherent scale ambiguity in monocular depth estimation while filtering outliers, leading to reliable densified 3D maps. We provide experimental evidence of accurate densified maps 4.15 mm RMS accuracy at affordable computing time in the C3VD phantom colon dataset. We report qualitative results on the real colonoscopy from the Endomapper dataset.
2503.14355	Runqi Meng	Runqi Meng, Sifan Song, Pengfei Jin, Yujin Oh, Lin Teng, Yulin Wang, Yiqun Sun, Ling Chen, Xiang Li, Quanzheng Li, Ning Guo, Dinggang Shen	MAST-Pro: Dynamic Mixture-of-Experts for Adaptive Segmentation of Pan-Tumors with Knowledge-Driven Prompts	10 pages, 2 figures	null	null	null	cs.CV	http://creativecommons.org/licenses/by-nc-nd/4.0/	Accurate tumor segmentation is crucial for cancer diagnosis and treatment. While foundation models have advanced general-purpose segmentation, existing methods still struggle with: (1) limited incorporation of medical priors, (2) imbalance between generic and tumor-specific features, and (3) high computational costs for clinical adaptation. To address these challenges, we propose MAST-Pro (Mixture-of-experts for Adaptive Segmentation of pan-Tumors with knowledge-driven Prompts), a novel framework that integrates dynamic Mixture-of-Experts (D-MoE) and knowledge-driven prompts for pan-tumor segmentation. Specifically, text and anatomical prompts provide domain-specific priors, guiding tumor representation learning, while D-MoE dynamically selects experts to balance generic and tumor-specific feature learning, improving segmentation accuracy across diverse tumor types. To enhance efficiency, we employ Parameter-Efficient Fine-Tuning (PEFT), optimizing MAST-Pro with significantly reduced computational overhead. Experiments on multi-anatomical tumor datasets demonstrate that MAST-Pro outperforms state-of-the-art approaches, achieving up to a 5.20% improvement in average DSC while reducing trainable parameters by 91.04%, without compromising accuracy.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:39:44 GMT" } ]	2025-03-19T00:00:00	[ [ "Meng", "Runqi", "" ], [ "Song", "Sifan", "" ], [ "Jin", "Pengfei", "" ], [ "Oh", "Yujin", "" ], [ "Teng", "Lin", "" ], [ "Wang", "Yulin", "" ], [ "Sun", "Yiqun", "" ], [ "Chen", "Ling", "" ], [ "Li", "Xiang", "" ], [ "Li", "Quanzheng", "" ], [ "Guo", "Ning", "" ], [ "Shen", "Dinggang", "" ] ]	TITLE: MAST-Pro: Dynamic Mixture-of-Experts for Adaptive Segmentation of Pan-Tumors with Knowledge-Driven Prompts ABSTRACT: Accurate tumor segmentation is crucial for cancer diagnosis and treatment. While foundation models have advanced general-purpose segmentation, existing methods still struggle with: (1) limited incorporation of medical priors, (2) imbalance between generic and tumor-specific features, and (3) high computational costs for clinical adaptation. To address these challenges, we propose MAST-Pro (Mixture-of-experts for Adaptive Segmentation of pan-Tumors with knowledge-driven Prompts), a novel framework that integrates dynamic Mixture-of-Experts (D-MoE) and knowledge-driven prompts for pan-tumor segmentation. Specifically, text and anatomical prompts provide domain-specific priors, guiding tumor representation learning, while D-MoE dynamically selects experts to balance generic and tumor-specific feature learning, improving segmentation accuracy across diverse tumor types. To enhance efficiency, we employ Parameter-Efficient Fine-Tuning (PEFT), optimizing MAST-Pro with significantly reduced computational overhead. Experiments on multi-anatomical tumor datasets demonstrate that MAST-Pro outperforms state-of-the-art approaches, achieving up to a 5.20% improvement in average DSC while reducing trainable parameters by 91.04%, without compromising accuracy.
2503.14356	Alexander Partin	Alexander Partin (1), Priyanka Vasanthakumari (1), Oleksandr Narykov (1), Andreas Wilke (1), Natasha Koussa (2), Sara E. Jones (2), Yitan Zhu (1), Jamie C. Overbeek (1), Rajeev Jain (1), Gayara Demini Fernando (3), Cesar Sanchez-Villalobos (4), Cristina Garcia-Cardona (5), Jamaludin Mohd-Yusof (5), Nicholas Chia (1), Justin M. Wozniak (1), Souparno Ghosh (3), Ranadip Pal (4), Thomas S. Brettin (1), M. Ryan Weil (2), Rick L. Stevens (1 and 6) ((1) Division of Data Science and Learning, Argonne National Laboratory, Lemont, IL, USA, (2) Frederick National Laboratory for Cancer Research, Cancer Data Science Initiatives, Cancer Research Technology Program, Frederick, MD, USA, (3) Department of Statistics, University of Nebraska-Lincoln, Lincoln, NE, USA, (4) Department of Electrical and Computer Engineering, Texas Tech University, Lubbock, TX, USA, (5) Division of Computer, Computational and Statistical Sciences, Los Alamos National Laboratory, Los Alamos, NM, USA, (6) Department of Computer Science, The University of Chicago, Chicago, IL, USA)	Benchmarking community drug response prediction models: datasets, models, tools, and metrics for cross-dataset generalization analysis	18 pages, 9 figures	null	null	null	cs.LG q-bio.QM	http://creativecommons.org/licenses/by/4.0/	Deep learning (DL) and machine learning (ML) models have shown promise in drug response prediction (DRP), yet their ability to generalize across datasets remains an open question, raising concerns about their real-world applicability. Due to the lack of standardized benchmarking approaches, model evaluations and comparisons often rely on inconsistent datasets and evaluation criteria, making it difficult to assess true predictive capabilities. In this work, we introduce a benchmarking framework for evaluating cross-dataset prediction generalization in DRP models. Our framework incorporates five publicly available drug screening datasets, six standardized DRP models, and a scalable workflow for systematic evaluation. To assess model generalization, we introduce a set of evaluation metrics that quantify both absolute performance (e.g., predictive accuracy across datasets) and relative performance (e.g., performance drop compared to within-dataset results), enabling a more comprehensive assessment of model transferability. Our results reveal substantial performance drops when models are tested on unseen datasets, underscoring the importance of rigorous generalization assessments. While several models demonstrate relatively strong cross-dataset generalization, no single model consistently outperforms across all datasets. Furthermore, we identify CTRPv2 as the most effective source dataset for training, yielding higher generalization scores across target datasets. By sharing this standardized evaluation framework with the community, our study aims to establish a rigorous foundation for model comparison, and accelerate the development of robust DRP models for real-world applications.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:40:18 GMT" } ]	2025-03-19T00:00:00	[ [ "Partin", "Alexander", "", "1 and 6" ], [ "Vasanthakumari", "Priyanka", "", "1 and 6" ], [ "Narykov", "Oleksandr", "", "1 and 6" ], [ "Wilke", "Andreas", "", "1 and 6" ], [ "Koussa", "Natasha", "", "1 and 6" ], [ "Jones", "Sara E.", "", "1 and 6" ], [ "Zhu", "Yitan", "", "1 and 6" ], [ "Overbeek", "Jamie C.", "", "1 and 6" ], [ "Jain", "Rajeev", "", "1 and 6" ], [ "Fernando", "Gayara Demini", "", "1 and 6" ], [ "Sanchez-Villalobos", "Cesar", "", "1 and 6" ], [ "Garcia-Cardona", "Cristina", "", "1 and 6" ], [ "Mohd-Yusof", "Jamaludin", "", "1 and 6" ], [ "Chia", "Nicholas", "", "1 and 6" ], [ "Wozniak", "Justin M.", "", "1 and 6" ], [ "Ghosh", "Souparno", "", "1 and 6" ], [ "Pal", "Ranadip", "", "1 and 6" ], [ "Brettin", "Thomas S.", "", "1 and 6" ], [ "Weil", "M. Ryan", "", "1 and 6" ], [ "Stevens", "Rick L.", "", "1 and 6" ] ]	TITLE: Benchmarking community drug response prediction models: datasets, models, tools, and metrics for cross-dataset generalization analysis ABSTRACT: Deep learning (DL) and machine learning (ML) models have shown promise in drug response prediction (DRP), yet their ability to generalize across datasets remains an open question, raising concerns about their real-world applicability. Due to the lack of standardized benchmarking approaches, model evaluations and comparisons often rely on inconsistent datasets and evaluation criteria, making it difficult to assess true predictive capabilities. In this work, we introduce a benchmarking framework for evaluating cross-dataset prediction generalization in DRP models. Our framework incorporates five publicly available drug screening datasets, six standardized DRP models, and a scalable workflow for systematic evaluation. To assess model generalization, we introduce a set of evaluation metrics that quantify both absolute performance (e.g., predictive accuracy across datasets) and relative performance (e.g., performance drop compared to within-dataset results), enabling a more comprehensive assessment of model transferability. Our results reveal substantial performance drops when models are tested on unseen datasets, underscoring the importance of rigorous generalization assessments. While several models demonstrate relatively strong cross-dataset generalization, no single model consistently outperforms across all datasets. Furthermore, we identify CTRPv2 as the most effective source dataset for training, yielding higher generalization scores across target datasets. By sharing this standardized evaluation framework with the community, our study aims to establish a rigorous foundation for model comparison, and accelerate the development of robust DRP models for real-world applications.
2503.14357	Giovanni Sansavini	Alfredo Oneto, Blazhe Gjorgiev, Giovanni Sansavini	Wasserstein-based Kernels for Clustering: Application to Power Distribution Graphs	null	null	null	null	cs.LG stat.AP	http://creativecommons.org/licenses/by-nc-nd/4.0/	Many data clustering applications must handle objects that cannot be represented as vector data. In this context, the bag-of-vectors representation can be leveraged to describe complex objects through discrete distributions, and the Wasserstein distance can effectively measure the dissimilarity between them. Additionally, kernel methods can be used to embed data into feature spaces that are easier to analyze. Despite significant progress in data clustering, a method that simultaneously accounts for distributional and vectorial dissimilarity measures is still lacking. To tackle this gap, this work explores kernel methods and Wasserstein distance metrics to develop a computationally tractable clustering framework. The compositional properties of kernels allow the simultaneous handling of different metrics, enabling the integration of both vectors and discrete distributions for object representation. This approach is flexible enough to be applied in various domains, such as graph analysis and image processing. The framework consists of three main components. First, we efficiently approximate pairwise Wasserstein distances using multiple reference distributions. Second, we employ kernel functions based on Wasserstein distances and present ways of composing kernels to express different types of information. Finally, we use the kernels to cluster data and evaluate the quality of the results using scalable and distance-agnostic validity indices. A case study involving two datasets of 879 and 34,920 power distribution graphs demonstrates the framework's effectiveness and efficiency.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:40:55 GMT" } ]	2025-03-19T00:00:00	[ [ "Oneto", "Alfredo", "" ], [ "Gjorgiev", "Blazhe", "" ], [ "Sansavini", "Giovanni", "" ] ]	TITLE: Wasserstein-based Kernels for Clustering: Application to Power Distribution Graphs ABSTRACT: Many data clustering applications must handle objects that cannot be represented as vector data. In this context, the bag-of-vectors representation can be leveraged to describe complex objects through discrete distributions, and the Wasserstein distance can effectively measure the dissimilarity between them. Additionally, kernel methods can be used to embed data into feature spaces that are easier to analyze. Despite significant progress in data clustering, a method that simultaneously accounts for distributional and vectorial dissimilarity measures is still lacking. To tackle this gap, this work explores kernel methods and Wasserstein distance metrics to develop a computationally tractable clustering framework. The compositional properties of kernels allow the simultaneous handling of different metrics, enabling the integration of both vectors and discrete distributions for object representation. This approach is flexible enough to be applied in various domains, such as graph analysis and image processing. The framework consists of three main components. First, we efficiently approximate pairwise Wasserstein distances using multiple reference distributions. Second, we employ kernel functions based on Wasserstein distances and present ways of composing kernels to express different types of information. Finally, we use the kernels to cluster data and evaluate the quality of the results using scalable and distance-agnostic validity indices. A case study involving two datasets of 879 and 34,920 power distribution graphs demonstrates the framework's effectiveness and efficiency.
2503.14358	Chao Wang	Chao Wang, Giulio Franzese, Alessandro Finamore, Pietro Michiardi	RFMI: Estimating Mutual Information on Rectified Flow for Text-to-Image Alignment	to appear at ICLR 2025 Workshop on Deep Generative Model in Machine Learning: Theory, Principle and Efficacy	null	null	null	cs.CV cs.LG	http://creativecommons.org/licenses/by/4.0/	Rectified Flow (RF) models trained with a Flow matching framework have achieved state-of-the-art performance on Text-to-Image (T2I) conditional generation. Yet, multiple benchmarks show that synthetic images can still suffer from poor alignment with the prompt, i.e., images show wrong attribute binding, subject positioning, numeracy, etc. While the literature offers many methods to improve T2I alignment, they all consider only Diffusion Models, and require auxiliary datasets, scoring models, and linguistic analysis of the prompt. In this paper we aim to address these gaps. First, we introduce RFMI, a novel Mutual Information (MI) estimator for RF models that uses the pre-trained model itself for the MI estimation. Then, we investigate a self-supervised fine-tuning approach for T2I alignment based on RFMI that does not require auxiliary information other than the pre-trained model itself. Specifically, a fine-tuning set is constructed by selecting synthetic images generated from the pre-trained RF model and having high point-wise MI between images and prompts. Our experiments on MI estimation benchmarks demonstrate the validity of RFMI, and empirical fine-tuning on SD3.5-Medium confirms the effectiveness of RFMI for improving T2I alignment while maintaining image quality.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:41:45 GMT" } ]	2025-03-19T00:00:00	[ [ "Wang", "Chao", "" ], [ "Franzese", "Giulio", "" ], [ "Finamore", "Alessandro", "" ], [ "Michiardi", "Pietro", "" ] ]	TITLE: RFMI: Estimating Mutual Information on Rectified Flow for Text-to-Image Alignment ABSTRACT: Rectified Flow (RF) models trained with a Flow matching framework have achieved state-of-the-art performance on Text-to-Image (T2I) conditional generation. Yet, multiple benchmarks show that synthetic images can still suffer from poor alignment with the prompt, i.e., images show wrong attribute binding, subject positioning, numeracy, etc. While the literature offers many methods to improve T2I alignment, they all consider only Diffusion Models, and require auxiliary datasets, scoring models, and linguistic analysis of the prompt. In this paper we aim to address these gaps. First, we introduce RFMI, a novel Mutual Information (MI) estimator for RF models that uses the pre-trained model itself for the MI estimation. Then, we investigate a self-supervised fine-tuning approach for T2I alignment based on RFMI that does not require auxiliary information other than the pre-trained model itself. Specifically, a fine-tuning set is constructed by selecting synthetic images generated from the pre-trained RF model and having high point-wise MI between images and prompts. Our experiments on MI estimation benchmarks demonstrate the validity of RFMI, and empirical fine-tuning on SD3.5-Medium confirms the effectiveness of RFMI for improving T2I alignment while maintaining image quality.
2503.14359	Zhengxian Yang	Zhengxian Yang, Shi Pan, Shengqi Wang, Haoxiang Wang, Li Lin, Guanjun Li, Zhengqi Wen, Borong Lin, Jianhua Tao, Tao Yu	ImViD: Immersive Volumetric Videos for Enhanced VR Engagement	Accepted by CVPR 2025	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	User engagement is greatly enhanced by fully immersive multi-modal experiences that combine visual and auditory stimuli. Consequently, the next frontier in VR/AR technologies lies in immersive volumetric videos with complete scene capture, large 6-DoF interaction space, multi-modal feedback, and high resolution & frame-rate contents. To stimulate the reconstruction of immersive volumetric videos, we introduce ImViD, a multi-view, multi-modal dataset featuring complete space-oriented data capture and various indoor/outdoor scenarios. Our capture rig supports multi-view video-audio capture while on the move, a capability absent in existing datasets, significantly enhancing the completeness, flexibility, and efficiency of data capture. The captured multi-view videos (with synchronized audios) are in 5K resolution at 60FPS, lasting from 1-5 minutes, and include rich foreground-background elements, and complex dynamics. We benchmark existing methods using our dataset and establish a base pipeline for constructing immersive volumetric videos from multi-view audiovisual inputs for 6-DoF multi-modal immersive VR experiences. The benchmark and the reconstruction and interaction results demonstrate the effectiveness of our dataset and baseline method, which we believe will stimulate future research on immersive volumetric video production.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:42:22 GMT" } ]	2025-03-19T00:00:00	[ [ "Yang", "Zhengxian", "" ], [ "Pan", "Shi", "" ], [ "Wang", "Shengqi", "" ], [ "Wang", "Haoxiang", "" ], [ "Lin", "Li", "" ], [ "Li", "Guanjun", "" ], [ "Wen", "Zhengqi", "" ], [ "Lin", "Borong", "" ], [ "Tao", "Jianhua", "" ], [ "Yu", "Tao", "" ] ]	TITLE: ImViD: Immersive Volumetric Videos for Enhanced VR Engagement ABSTRACT: User engagement is greatly enhanced by fully immersive multi-modal experiences that combine visual and auditory stimuli. Consequently, the next frontier in VR/AR technologies lies in immersive volumetric videos with complete scene capture, large 6-DoF interaction space, multi-modal feedback, and high resolution & frame-rate contents. To stimulate the reconstruction of immersive volumetric videos, we introduce ImViD, a multi-view, multi-modal dataset featuring complete space-oriented data capture and various indoor/outdoor scenarios. Our capture rig supports multi-view video-audio capture while on the move, a capability absent in existing datasets, significantly enhancing the completeness, flexibility, and efficiency of data capture. The captured multi-view videos (with synchronized audios) are in 5K resolution at 60FPS, lasting from 1-5 minutes, and include rich foreground-background elements, and complex dynamics. We benchmark existing methods using our dataset and establish a base pipeline for constructing immersive volumetric videos from multi-view audiovisual inputs for 6-DoF multi-modal immersive VR experiences. The benchmark and the reconstruction and interaction results demonstrate the effectiveness of our dataset and baseline method, which we believe will stimulate future research on immersive volumetric video production.
2503.14362	Nicolas Menand	Nicolas Menand and Erik Waingarten	Streaming and Massively Parallel Algorithms for Euclidean Max-Cut	null	null	null	null	cs.DS	http://creativecommons.org/licenses/by/4.0/	Given a set of vectors $X = \{ x_1,\dots, x_n \} \subset \mathbb{R}^d$, the Euclidean max-cut problem asks to partition the vectors into two parts so as to maximize the sum of Euclidean distances which cross the partition. We design new algorithms for Euclidean max-cut in models for massive datasets: $\bullet$ We give a fully-scalable constant-round MPC algorithm using $O(nd) + n \cdot \text{poly}( \log(n) / \epsilon)$ total space which gives a $(1+\epsilon)$-approximate Euclidean max-cut. $\bullet$ We give a dynamic streaming algorithm using $\text{poly}(d \log \Delta / \epsilon)$ space when $X \subseteq [\Delta]^d$, which provides oracle access to a $(1+\epsilon)$-approximate Euclidean max-cut. Recently, Chen, Jiang, and Krauthgamer $[\text{STOC}~'23]$ gave a dynamic streaming algorithm with space $\text{poly}(d\log\Delta/\epsilon)$ to approximate the value of the Euclidean max-cut, but could not provide oracle access to an approximately optimal cut. This was left open in that work, and we resolve it here. Both algorithms follow from the same framework, which analyzes a ``parallel'' and ``subsampled'' (Euclidean) version of a greedy algorithm of Mathieu and Schudy $[\text{SODA}~'08]$ for dense max-cut.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:45:00 GMT" } ]	2025-03-19T00:00:00	[ [ "Menand", "Nicolas", "" ], [ "Waingarten", "Erik", "" ] ]	TITLE: Streaming and Massively Parallel Algorithms for Euclidean Max-Cut ABSTRACT: Given a set of vectors $X = \{ x_1,\dots, x_n \} \subset \mathbb{R}^d$, the Euclidean max-cut problem asks to partition the vectors into two parts so as to maximize the sum of Euclidean distances which cross the partition. We design new algorithms for Euclidean max-cut in models for massive datasets: $\bullet$ We give a fully-scalable constant-round MPC algorithm using $O(nd) + n \cdot \text{poly}( \log(n) / \epsilon)$ total space which gives a $(1+\epsilon)$-approximate Euclidean max-cut. $\bullet$ We give a dynamic streaming algorithm using $\text{poly}(d \log \Delta / \epsilon)$ space when $X \subseteq [\Delta]^d$, which provides oracle access to a $(1+\epsilon)$-approximate Euclidean max-cut. Recently, Chen, Jiang, and Krauthgamer $[\text{STOC}~'23]$ gave a dynamic streaming algorithm with space $\text{poly}(d\log\Delta/\epsilon)$ to approximate the value of the Euclidean max-cut, but could not provide oracle access to an approximately optimal cut. This was left open in that work, and we resolve it here. Both algorithms follow from the same framework, which analyzes a ``parallel'' and ``subsampled'' (Euclidean) version of a greedy algorithm of Mathieu and Schudy $[\text{SODA}~'08]$ for dense max-cut.
2503.14369	Giuseppe Bruni	Giuseppe Bruni, Sepehr Maleki, Senthil K Krishnababu	C(NN)FD -- Deep Learning Modelling of Multi-Stage Axial Compressors Aerodynamics	null	null	null	null	physics.flu-dyn cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The field of scientific machine learning and its applications to numerical analyses such as CFD has recently experienced a surge in interest. While its viability has been demonstrated in different domains, it has not yet reached a level of robustness and scalability to make it practical for industrial applications in the turbomachinery field. The highly complex, turbulent, and three-dimensional flows of multi-stage axial compressors for gas turbine applications represent a remarkably challenging case. This is due to the high-dimensionality of the regression of the flow-field from geometrical and operational variables, and the high computational cost associated with the large scale of the CFD domains. This paper demonstrates the development and application of a generalized deep learning framework for predictions of the flow field and aerodynamic performance of multi-stage axial compressors, also potentially applicable to any type of turbomachinery. A physics-based dimensionality reduction unlocks the potential for flow-field predictions for large-scale domains, re-formulating the regression problem from an unstructured to a structured one. The relevant physical equations are used to define a multi-dimensional physical loss function. Compared to "black-box" approaches, the proposed framework has the advantage of physically explainable predictions of overall performance, as the corresponding aerodynamic drivers can be identified on a 0D/1D/2D/3D level. An iterative architecture is employed, improving the accuracy of the predictions, as well as estimating the associated uncertainty. The model is trained on a series of dataset including manufacturing and build variations, different geometries, compressor designs and operating conditions. This demonstrates the capability to predict the flow-field and the overall performance in a generalizable manner, with accuracy comparable to the benchmark.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 15:58:58 GMT" } ]	2025-03-19T00:00:00	[ [ "Bruni", "Giuseppe", "" ], [ "Maleki", "Sepehr", "" ], [ "Krishnababu", "Senthil K", "" ] ]	TITLE: C(NN)FD -- Deep Learning Modelling of Multi-Stage Axial Compressors Aerodynamics ABSTRACT: The field of scientific machine learning and its applications to numerical analyses such as CFD has recently experienced a surge in interest. While its viability has been demonstrated in different domains, it has not yet reached a level of robustness and scalability to make it practical for industrial applications in the turbomachinery field. The highly complex, turbulent, and three-dimensional flows of multi-stage axial compressors for gas turbine applications represent a remarkably challenging case. This is due to the high-dimensionality of the regression of the flow-field from geometrical and operational variables, and the high computational cost associated with the large scale of the CFD domains. This paper demonstrates the development and application of a generalized deep learning framework for predictions of the flow field and aerodynamic performance of multi-stage axial compressors, also potentially applicable to any type of turbomachinery. A physics-based dimensionality reduction unlocks the potential for flow-field predictions for large-scale domains, re-formulating the regression problem from an unstructured to a structured one. The relevant physical equations are used to define a multi-dimensional physical loss function. Compared to "black-box" approaches, the proposed framework has the advantage of physically explainable predictions of overall performance, as the corresponding aerodynamic drivers can be identified on a 0D/1D/2D/3D level. An iterative architecture is employed, improving the accuracy of the predictions, as well as estimating the associated uncertainty. The model is trained on a series of dataset including manufacturing and build variations, different geometries, compressor designs and operating conditions. This demonstrates the capability to predict the flow-field and the overall performance in a generalizable manner, with accuracy comparable to the benchmark.
2503.14375	Zachary Kingston	Sai Coumar, Zachary Kingston	Evaluating Machine Learning Approaches for ASCII Art Generation	9 pages, 7 figures, 3 tables. Code available at https://github.com/saiccoumar/deep_ascii_converter	null	null	null	cs.GR cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Generating structured ASCII art using computational techniques demands a careful interplay between aesthetic representation and computational precision, requiring models that can effectively translate visual information into symbolic text characters. Although Convolutional Neural Networks (CNNs) have shown promise in this domain, the comparative performance of deep learning architectures and classical machine learning methods remains unexplored. This paper explores the application of contemporary ML and DL methods to generate structured ASCII art, focusing on three key criteria: fidelity, character classification accuracy, and output quality. We investigate deep learning architectures, including Multilayer Perceptrons (MLPs), ResNet, and MobileNetV2, alongside classical approaches such as Random Forests, Support Vector Machines (SVMs) and k-Nearest Neighbors (k-NN), trained on an augmented synthetic dataset of ASCII characters. Our results show that complex neural network architectures often fall short in producing high-quality ASCII art, whereas classical machine learning classifiers, despite their simplicity, achieve performance similar to CNNs. Our findings highlight the strength of classical methods in bridging model simplicity with output quality, offering new insights into ASCII art synthesis and machine learning on image data with low dimensionality.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:07:29 GMT" } ]	2025-03-19T00:00:00	[ [ "Coumar", "Sai", "" ], [ "Kingston", "Zachary", "" ] ]	TITLE: Evaluating Machine Learning Approaches for ASCII Art Generation ABSTRACT: Generating structured ASCII art using computational techniques demands a careful interplay between aesthetic representation and computational precision, requiring models that can effectively translate visual information into symbolic text characters. Although Convolutional Neural Networks (CNNs) have shown promise in this domain, the comparative performance of deep learning architectures and classical machine learning methods remains unexplored. This paper explores the application of contemporary ML and DL methods to generate structured ASCII art, focusing on three key criteria: fidelity, character classification accuracy, and output quality. We investigate deep learning architectures, including Multilayer Perceptrons (MLPs), ResNet, and MobileNetV2, alongside classical approaches such as Random Forests, Support Vector Machines (SVMs) and k-Nearest Neighbors (k-NN), trained on an augmented synthetic dataset of ASCII characters. Our results show that complex neural network architectures often fall short in producing high-quality ASCII art, whereas classical machine learning classifiers, despite their simplicity, achieve performance similar to CNNs. Our findings highlight the strength of classical methods in bridging model simplicity with output quality, offering new insights into ASCII art synthesis and machine learning on image data with low dimensionality.
2503.14377	Arash Afkanpour	Negin Baghbanzadeh, Adibvafa Fallahpour, Yasaman Parhizkar, Franklin Ogidi, Shuvendu Roy, Sajad Ashkezari, Vahid Reza Khazaie, Michael Colacci, Ali Etemad, Arash Afkanpour, Elham Dolatabadi	Advancing Medical Representation Learning Through High-Quality Data	null	null	null	null	eess.IV cs.CV cs.LG	http://creativecommons.org/licenses/by/4.0/	Despite the growing scale of medical Vision-Language datasets, the impact of dataset quality on model performance remains under-explored. We introduce Open-PMC, a high-quality medical dataset from PubMed Central, containing 2.2 million image-text pairs, enriched with image modality annotations, subfigures, and summarized in-text references. Notably, the in-text references provide richer medical context, extending beyond the abstract information typically found in captions. Through extensive experiments, we benchmark Open-PMC against larger datasets across retrieval and zero-shot classification tasks. Our results show that dataset quality-not just size-drives significant performance gains. We complement our benchmark with an in-depth analysis of feature representation. Our findings highlight the crucial role of data curation quality in advancing multimodal medical AI. We release Open-PMC, along with the trained models and our codebase.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:10:11 GMT" } ]	2025-03-19T00:00:00	[ [ "Baghbanzadeh", "Negin", "" ], [ "Fallahpour", "Adibvafa", "" ], [ "Parhizkar", "Yasaman", "" ], [ "Ogidi", "Franklin", "" ], [ "Roy", "Shuvendu", "" ], [ "Ashkezari", "Sajad", "" ], [ "Khazaie", "Vahid Reza", "" ], [ "Colacci", "Michael", "" ], [ "Etemad", "Ali", "" ], [ "Afkanpour", "Arash", "" ], [ "Dolatabadi", "Elham", "" ] ]	TITLE: Advancing Medical Representation Learning Through High-Quality Data ABSTRACT: Despite the growing scale of medical Vision-Language datasets, the impact of dataset quality on model performance remains under-explored. We introduce Open-PMC, a high-quality medical dataset from PubMed Central, containing 2.2 million image-text pairs, enriched with image modality annotations, subfigures, and summarized in-text references. Notably, the in-text references provide richer medical context, extending beyond the abstract information typically found in captions. Through extensive experiments, we benchmark Open-PMC against larger datasets across retrieval and zero-shot classification tasks. Our results show that dataset quality-not just size-drives significant performance gains. We complement our benchmark with an in-depth analysis of feature representation. Our findings highlight the crucial role of data curation quality in advancing multimodal medical AI. We release Open-PMC, along with the trained models and our codebase.
2503.14378	Zechen Bai	Zechen Bai, Hai Ci, Mike Zheng Shou	Impossible Videos	26 pages	null	null	null	cs.CV	http://creativecommons.org/licenses/by-nc-sa/4.0/	Synthetic videos nowadays is widely used to complement data scarcity and diversity of real-world videos. Current synthetic datasets primarily replicate real-world scenarios, leaving impossible, counterfactual and anti-reality video concepts underexplored. This work aims to answer two questions: 1) Can today's video generation models effectively follow prompts to create impossible video content? 2) Are today's video understanding models good enough for understanding impossible videos? To this end, we introduce IPV-Bench, a novel benchmark designed to evaluate and foster progress in video understanding and generation. IPV-Bench is underpinned by a comprehensive taxonomy, encompassing 4 domains, 14 categories. It features diverse scenes that defy physical, biological, geographical, or social laws. Based on the taxonomy, a prompt suite is constructed to evaluate video generation models, challenging their prompt following and creativity capabilities. In addition, a video benchmark is curated to assess Video-LLMs on their ability of understanding impossible videos, which particularly requires reasoning on temporal dynamics and world knowledge. Comprehensive evaluations reveal limitations and insights for future directions of video models, paving the way for next-generation video models.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:10:24 GMT" } ]	2025-03-19T00:00:00	[ [ "Bai", "Zechen", "" ], [ "Ci", "Hai", "" ], [ "Shou", "Mike Zheng", "" ] ]	TITLE: Impossible Videos ABSTRACT: Synthetic videos nowadays is widely used to complement data scarcity and diversity of real-world videos. Current synthetic datasets primarily replicate real-world scenarios, leaving impossible, counterfactual and anti-reality video concepts underexplored. This work aims to answer two questions: 1) Can today's video generation models effectively follow prompts to create impossible video content? 2) Are today's video understanding models good enough for understanding impossible videos? To this end, we introduce IPV-Bench, a novel benchmark designed to evaluate and foster progress in video understanding and generation. IPV-Bench is underpinned by a comprehensive taxonomy, encompassing 4 domains, 14 categories. It features diverse scenes that defy physical, biological, geographical, or social laws. Based on the taxonomy, a prompt suite is constructed to evaluate video generation models, challenging their prompt following and creativity capabilities. In addition, a video benchmark is curated to assess Video-LLMs on their ability of understanding impossible videos, which particularly requires reasoning on temporal dynamics and world knowledge. Comprehensive evaluations reveal limitations and insights for future directions of video models, paving the way for next-generation video models.
2503.14388	Yekatierina Churakova	Yekatierina Churakova Mathias Ekstedt	Vexed by VEX tools: Consistency evaluation of container vulnerability scanners	22 pages, 1 listing, 18 tables	null	null	null	cs.CR	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper presents a study that analyzed state-of-the-art vulnerability scanning tools applied to containers. We have focused the work on tools following the Vulnerability Exploitability eXchange (VEX) format, which has been introduced to complement Software Bills of Material (SBOM) with security advisories of known vulnerabilities. Being able to get an accurate understanding of vulnerabilities found in the dependencies of third-party software is critical for secure software development and risk analysis. Accepting the overwhelming challenge of estimating the precise accuracy and precision of a vulnerability scanner, we have in this study instead set out to explore how consistently different tools perform. By doing this, we aim to assess the maturity of the VEX tool field as a whole (rather than any particular tool). We have used the Jaccard and Tversky indices to produce similarity scores of tool performance for several different datasets created from container images. Overall, our results show a low level of consistency among the tools, thus indicating a low level of maturity in VEX tool space. We have performed a number of experiments to find and explanation to our results, but largely they are inconclusive and further research is needed to understand the underlying causalities of our findings.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:22:43 GMT" } ]	2025-03-19T00:00:00	[ [ "Ekstedt", "Yekatierina Churakova Mathias", "" ] ]	TITLE: Vexed by VEX tools: Consistency evaluation of container vulnerability scanners ABSTRACT: This paper presents a study that analyzed state-of-the-art vulnerability scanning tools applied to containers. We have focused the work on tools following the Vulnerability Exploitability eXchange (VEX) format, which has been introduced to complement Software Bills of Material (SBOM) with security advisories of known vulnerabilities. Being able to get an accurate understanding of vulnerabilities found in the dependencies of third-party software is critical for secure software development and risk analysis. Accepting the overwhelming challenge of estimating the precise accuracy and precision of a vulnerability scanner, we have in this study instead set out to explore how consistently different tools perform. By doing this, we aim to assess the maturity of the VEX tool field as a whole (rather than any particular tool). We have used the Jaccard and Tversky indices to produce similarity scores of tool performance for several different datasets created from container images. Overall, our results show a low level of consistency among the tools, thus indicating a low level of maturity in VEX tool space. We have performed a number of experiments to find and explanation to our results, but largely they are inconclusive and further research is needed to understand the underlying causalities of our findings.
2503.14395	Ruirui Liu	Jing Wang, Ruirui Liu, Yu Lei, Michael J. Baine, Tian Liu, Yang Lei	Weakly Supervised Spatial Implicit Neural Representation Learning for 3D MRI-Ultrasound Deformable Image Registration in HDR Prostate Brachytherapy	null	null	null	null	physics.med-ph cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Purpose: Accurate 3D MRI-ultrasound (US) deformable registration is critical for real-time guidance in high-dose-rate (HDR) prostate brachytherapy. We present a weakly supervised spatial implicit neural representation (SINR) method to address modality differences and pelvic anatomy challenges. Methods: The framework uses sparse surface supervision from MRI/US segmentations instead of dense intensity matching. SINR models deformations as continuous spatial functions, with patient-specific surface priors guiding a stationary velocity field for biologically plausible deformations. Validation included 20 public Prostate-MRI-US-Biopsy cases and 10 institutional HDR cases, evaluated via Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff distance (HD95). Results: The proposed method achieved robust registration. For the public dataset, prostate DSC was $0.93 \pm 0.05$, MSD $0.87 \pm 0.10$ mm, and HD95 $1.58 \pm 0.37$ mm. For the institutional dataset, prostate CTV achieved DSC $0.88 \pm 0.09$, MSD $1.21 \pm 0.38$ mm, and HD95 $2.09 \pm 1.48$ mm. Bladder and rectum performance was lower due to ultrasound's limited field of view. Visual assessments confirmed accurate alignment with minimal discrepancies. Conclusion: This study introduces a novel weakly supervised SINR-based approach for 3D MRI-US deformable registration. By leveraging sparse surface supervision and spatial priors, it achieves accurate, robust, and computationally efficient registration, enhancing real-time image guidance in HDR prostate brachytherapy and improving treatment precision.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:30:08 GMT" } ]	2025-03-19T00:00:00	[ [ "Wang", "Jing", "" ], [ "Liu", "Ruirui", "" ], [ "Lei", "Yu", "" ], [ "Baine", "Michael J.", "" ], [ "Liu", "Tian", "" ], [ "Lei", "Yang", "" ] ]	TITLE: Weakly Supervised Spatial Implicit Neural Representation Learning for 3D MRI-Ultrasound Deformable Image Registration in HDR Prostate Brachytherapy ABSTRACT: Purpose: Accurate 3D MRI-ultrasound (US) deformable registration is critical for real-time guidance in high-dose-rate (HDR) prostate brachytherapy. We present a weakly supervised spatial implicit neural representation (SINR) method to address modality differences and pelvic anatomy challenges. Methods: The framework uses sparse surface supervision from MRI/US segmentations instead of dense intensity matching. SINR models deformations as continuous spatial functions, with patient-specific surface priors guiding a stationary velocity field for biologically plausible deformations. Validation included 20 public Prostate-MRI-US-Biopsy cases and 10 institutional HDR cases, evaluated via Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff distance (HD95). Results: The proposed method achieved robust registration. For the public dataset, prostate DSC was $0.93 \pm 0.05$, MSD $0.87 \pm 0.10$ mm, and HD95 $1.58 \pm 0.37$ mm. For the institutional dataset, prostate CTV achieved DSC $0.88 \pm 0.09$, MSD $1.21 \pm 0.38$ mm, and HD95 $2.09 \pm 1.48$ mm. Bladder and rectum performance was lower due to ultrasound's limited field of view. Visual assessments confirmed accurate alignment with minimal discrepancies. Conclusion: This study introduces a novel weakly supervised SINR-based approach for 3D MRI-US deformable registration. By leveraging sparse surface supervision and spatial priors, it achieves accurate, robust, and computationally efficient registration, enhancing real-time image guidance in HDR prostate brachytherapy and improving treatment precision.
2503.14409	Merijn Floren	Merijn Floren, Jean-Philippe No\"el, Jan Swevers	Inference and Learning of Nonlinear LFR State-space Models	null	null	null	null	eess.SY cs.SY	http://creativecommons.org/licenses/by/4.0/	Estimating the parameters of nonlinear block-oriented state-space models from input-output data typically involves solving a highly non-convex optimization problem, making it susceptible to poor local minima and slow convergence. This paper presents a computationally efficient initialization method for fully parametrizing nonlinear linear fractional representation (NL-LFR) models using periodic data. The approach first infers the latent variables and then estimates the model parameters, yielding initial estimates that serve as a starting point for further nonlinear optimization. The proposed method shows robustness against poor local minima, and achieves a twofold error reduction compared to the state-of-the-art on a challenging benchmark dataset.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:49:56 GMT" } ]	2025-03-19T00:00:00	[ [ "Floren", "Merijn", "" ], [ "Noël", "Jean-Philippe", "" ], [ "Swevers", "Jan", "" ] ]	TITLE: Inference and Learning of Nonlinear LFR State-space Models ABSTRACT: Estimating the parameters of nonlinear block-oriented state-space models from input-output data typically involves solving a highly non-convex optimization problem, making it susceptible to poor local minima and slow convergence. This paper presents a computationally efficient initialization method for fully parametrizing nonlinear linear fractional representation (NL-LFR) models using periodic data. The approach first infers the latent variables and then estimates the model parameters, yielding initial estimates that serve as a starting point for further nonlinear optimization. The proposed method shows robustness against poor local minima, and achieves a twofold error reduction compared to the state-of-the-art on a challenging benchmark dataset.
2503.14411	Siwei Zhang	Siwei Zhang, Yun Xiong, Yateng Tang, Xi Chen, Zian Jia, Zehao Gu, Jiarong Xu, Jiawei Zhang	Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models	Submit to ICML2025	null	null	null	cs.CL cs.AI	http://creativecommons.org/licenses/by/4.0/	Temporal graph neural networks (TGNNs) have shown remarkable performance in temporal graph modeling. However, real-world temporal graphs often possess rich textual information, giving rise to temporal text-attributed graphs (TTAGs). Such combination of dynamic text semantics and evolving graph structures introduces heightened complexity. Existing TGNNs embed texts statically and rely heavily on encoding mechanisms that biasedly prioritize structural information, overlooking the temporal evolution of text semantics and the essential interplay between semantics and structures for synergistic reinforcement. To tackle these issues, we present \textbf{{Cross}}, a novel framework that seamlessly extends existing TGNNs for TTAG modeling. The key idea is to employ the advanced large language models (LLMs) to extract the dynamic semantics in text space and then generate expressive representations unifying both semantics and structures. Specifically, we propose a Temporal Semantics Extractor in the {Cross} framework, which empowers the LLM to offer the temporal semantic understanding of node's evolving contexts of textual neighborhoods, facilitating semantic dynamics. Subsequently, we introduce the Semantic-structural Co-encoder, which collaborates with the above Extractor for synthesizing illuminating representations by jointly considering both semantic and structural information while encouraging their mutual reinforcement. Extensive experimental results on four public datasets and one practical industrial dataset demonstrate {Cross}'s significant effectiveness and robustness.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:50:10 GMT" } ]	2025-03-19T00:00:00	[ [ "Zhang", "Siwei", "" ], [ "Xiong", "Yun", "" ], [ "Tang", "Yateng", "" ], [ "Chen", "Xi", "" ], [ "Jia", "Zian", "" ], [ "Gu", "Zehao", "" ], [ "Xu", "Jiarong", "" ], [ "Zhang", "Jiawei", "" ] ]	TITLE: Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models ABSTRACT: Temporal graph neural networks (TGNNs) have shown remarkable performance in temporal graph modeling. However, real-world temporal graphs often possess rich textual information, giving rise to temporal text-attributed graphs (TTAGs). Such combination of dynamic text semantics and evolving graph structures introduces heightened complexity. Existing TGNNs embed texts statically and rely heavily on encoding mechanisms that biasedly prioritize structural information, overlooking the temporal evolution of text semantics and the essential interplay between semantics and structures for synergistic reinforcement. To tackle these issues, we present \textbf{{Cross}}, a novel framework that seamlessly extends existing TGNNs for TTAG modeling. The key idea is to employ the advanced large language models (LLMs) to extract the dynamic semantics in text space and then generate expressive representations unifying both semantics and structures. Specifically, we propose a Temporal Semantics Extractor in the {Cross} framework, which empowers the LLM to offer the temporal semantic understanding of node's evolving contexts of textual neighborhoods, facilitating semantic dynamics. Subsequently, we introduce the Semantic-structural Co-encoder, which collaborates with the above Extractor for synthesizing illuminating representations by jointly considering both semantic and structural information while encouraging their mutual reinforcement. Extensive experimental results on four public datasets and one practical industrial dataset demonstrate {Cross}'s significant effectiveness and robustness.
2503.14421	Radu Tudor Ionescu	Vlad Hondru, Eduard Hogea, Darian Onchis, Radu Tudor Ionescu	ExDDV: A New Dataset for Explainable Deepfake Detection in Video	null	null	null	null	cs.CV cs.AI cs.CL cs.LG cs.MM	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The ever growing realism and quality of generated videos makes it increasingly harder for humans to spot deepfake content, who need to rely more and more on automatic deepfake detectors. However, deepfake detectors are also prone to errors, and their decisions are not explainable, leaving humans vulnerable to deepfake-based fraud and misinformation. To this end, we introduce ExDDV, the first dataset and benchmark for Explainable Deepfake Detection in Video. ExDDV comprises around 5.4K real and deepfake videos that are manually annotated with text descriptions (to explain the artifacts) and clicks (to point out the artifacts). We evaluate a number of vision-language models on ExDDV, performing experiments with various fine-tuning and in-context learning strategies. Our results show that text and click supervision are both required to develop robust explainable models for deepfake videos, which are able to localize and describe the observed artifacts. Our novel dataset and code to reproduce the results are available at https://github.com/vladhondru25/ExDDV.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 16:55:07 GMT" } ]	2025-03-19T00:00:00	[ [ "Hondru", "Vlad", "" ], [ "Hogea", "Eduard", "" ], [ "Onchis", "Darian", "" ], [ "Ionescu", "Radu Tudor", "" ] ]	TITLE: ExDDV: A New Dataset for Explainable Deepfake Detection in Video ABSTRACT: The ever growing realism and quality of generated videos makes it increasingly harder for humans to spot deepfake content, who need to rely more and more on automatic deepfake detectors. However, deepfake detectors are also prone to errors, and their decisions are not explainable, leaving humans vulnerable to deepfake-based fraud and misinformation. To this end, we introduce ExDDV, the first dataset and benchmark for Explainable Deepfake Detection in Video. ExDDV comprises around 5.4K real and deepfake videos that are manually annotated with text descriptions (to explain the artifacts) and clicks (to point out the artifacts). We evaluate a number of vision-language models on ExDDV, performing experiments with various fine-tuning and in-context learning strategies. Our results show that text and click supervision are both required to develop robust explainable models for deepfake videos, which are able to localize and describe the observed artifacts. Our novel dataset and code to reproduce the results are available at https://github.com/vladhondru25/ExDDV.
2503.14443	Yaroslav Zharov	Aleksandra Eliseeva, Alexander Kovrigin, Ilia Kholkin, Egor Bogomolov, Yaroslav Zharov	EnvBench: A Benchmark for Automated Environment Setup	Accepted at the DL4Code workshop at ICLR'25	null	null	null	cs.LG cs.SE	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Recent advances in Large Language Models (LLMs) have enabled researchers to focus on practical repository-level tasks in software engineering domain. In this work, we consider a cornerstone task for automating work with software repositories-environment setup, i.e., a task of configuring a repository-specific development environment on a system. Existing studies on environment setup introduce innovative agentic strategies, but their evaluation is often based on small datasets that may not capture the full range of configuration challenges encountered in practice. To address this gap, we introduce a comprehensive environment setup benchmark EnvBench. It encompasses 329 Python and 665 JVM-based (Java, Kotlin) repositories, with a focus on repositories that present genuine configuration challenges, excluding projects that can be fully configured by simple deterministic scripts. To enable further benchmark extension and usage for model tuning, we implement two automatic metrics: a static analysis check for missing imports in Python and a compilation check for JVM languages. We demonstrate the applicability of our benchmark by evaluating three environment setup approaches, including a simple zero-shot baseline and two agentic workflows, that we test with two powerful LLM backbones, GPT-4o and GPT-4o-mini. The best approach manages to successfully configure 6.69% repositories for Python and 29.47% repositories for JVM, suggesting that EnvBench remains challenging for current approaches. Our benchmark suite is publicly available at https://github.com/JetBrains-Research/EnvBench. The dataset and experiment trajectories are available at https://jb.gg/envbench.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:19:12 GMT" } ]	2025-03-19T00:00:00	[ [ "Eliseeva", "Aleksandra", "" ], [ "Kovrigin", "Alexander", "" ], [ "Kholkin", "Ilia", "" ], [ "Bogomolov", "Egor", "" ], [ "Zharov", "Yaroslav", "" ] ]	TITLE: EnvBench: A Benchmark for Automated Environment Setup ABSTRACT: Recent advances in Large Language Models (LLMs) have enabled researchers to focus on practical repository-level tasks in software engineering domain. In this work, we consider a cornerstone task for automating work with software repositories-environment setup, i.e., a task of configuring a repository-specific development environment on a system. Existing studies on environment setup introduce innovative agentic strategies, but their evaluation is often based on small datasets that may not capture the full range of configuration challenges encountered in practice. To address this gap, we introduce a comprehensive environment setup benchmark EnvBench. It encompasses 329 Python and 665 JVM-based (Java, Kotlin) repositories, with a focus on repositories that present genuine configuration challenges, excluding projects that can be fully configured by simple deterministic scripts. To enable further benchmark extension and usage for model tuning, we implement two automatic metrics: a static analysis check for missing imports in Python and a compilation check for JVM languages. We demonstrate the applicability of our benchmark by evaluating three environment setup approaches, including a simple zero-shot baseline and two agentic workflows, that we test with two powerful LLM backbones, GPT-4o and GPT-4o-mini. The best approach manages to successfully configure 6.69% repositories for Python and 29.47% repositories for JVM, suggesting that EnvBench remains challenging for current approaches. Our benchmark suite is publicly available at https://github.com/JetBrains-Research/EnvBench. The dataset and experiment trajectories are available at https://jb.gg/envbench.
2503.14445	Stanislaw Szymanowicz	Stanislaw Szymanowicz and Jason Y. Zhang and Pratul Srinivasan and Ruiqi Gao and Arthur Brussee and Aleksander Holynski and Ricardo Martin-Brualla and Jonathan T. Barron and Philipp Henzler	Bolt3D: Generating 3D Scenes in Seconds	Project page: https://szymanowiczs.github.io/bolt3d	null	null	null	cs.CV	http://creativecommons.org/licenses/by-sa/4.0/	We present a latent diffusion model for fast feed-forward 3D scene generation. Given one or more images, our model Bolt3D directly samples a 3D scene representation in less than seven seconds on a single GPU. We achieve this by leveraging powerful and scalable existing 2D diffusion network architectures to produce consistent high-fidelity 3D scene representations. To train this model, we create a large-scale multiview-consistent dataset of 3D geometry and appearance by applying state-of-the-art dense 3D reconstruction techniques to existing multiview image datasets. Compared to prior multiview generative models that require per-scene optimization for 3D reconstruction, Bolt3D reduces the inference cost by a factor of up to 300 times.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:24:19 GMT" } ]	2025-03-19T00:00:00	[ [ "Szymanowicz", "Stanislaw", "" ], [ "Zhang", "Jason Y.", "" ], [ "Srinivasan", "Pratul", "" ], [ "Gao", "Ruiqi", "" ], [ "Brussee", "Arthur", "" ], [ "Holynski", "Aleksander", "" ], [ "Martin-Brualla", "Ricardo", "" ], [ "Barron", "Jonathan T.", "" ], [ "Henzler", "Philipp", "" ] ]	TITLE: Bolt3D: Generating 3D Scenes in Seconds ABSTRACT: We present a latent diffusion model for fast feed-forward 3D scene generation. Given one or more images, our model Bolt3D directly samples a 3D scene representation in less than seven seconds on a single GPU. We achieve this by leveraging powerful and scalable existing 2D diffusion network architectures to produce consistent high-fidelity 3D scene representations. To train this model, we create a large-scale multiview-consistent dataset of 3D geometry and appearance by applying state-of-the-art dense 3D reconstruction techniques to existing multiview image datasets. Compared to prior multiview generative models that require per-scene optimization for 3D reconstruction, Bolt3D reduces the inference cost by a factor of up to 300 times.
2503.14459	Piersilvio De Bartolomeis	Piersilvio De Bartolomeis, Julia Kostin, Javier Abad, Yixin Wang, Fanny Yang	Doubly robust identification of treatment effects from multiple environments	Accepted for presentation at the International Conference on Learning Representations (ICLR) 2025	null	null	null	stat.ML cs.LG stat.ME	http://creativecommons.org/licenses/by/4.0/	Practical and ethical constraints often require the use of observational data for causal inference, particularly in medicine and social sciences. Yet, observational datasets are prone to confounding, potentially compromising the validity of causal conclusions. While it is possible to correct for biases if the underlying causal graph is known, this is rarely a feasible ask in practical scenarios. A common strategy is to adjust for all available covariates, yet this approach can yield biased treatment effect estimates, especially when post-treatment or unobserved variables are present. We propose RAMEN, an algorithm that produces unbiased treatment effect estimates by leveraging the heterogeneity of multiple data sources without the need to know or learn the underlying causal graph. Notably, RAMEN achieves doubly robust identification: it can identify the treatment effect whenever the causal parents of the treatment or those of the outcome are observed, and the node whose parents are observed satisfies an invariance assumption. Empirical evaluations on synthetic and real-world datasets show that our approach outperforms existing methods.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:33:10 GMT" } ]	2025-03-19T00:00:00	[ [ "De Bartolomeis", "Piersilvio", "" ], [ "Kostin", "Julia", "" ], [ "Abad", "Javier", "" ], [ "Wang", "Yixin", "" ], [ "Yang", "Fanny", "" ] ]	TITLE: Doubly robust identification of treatment effects from multiple environments ABSTRACT: Practical and ethical constraints often require the use of observational data for causal inference, particularly in medicine and social sciences. Yet, observational datasets are prone to confounding, potentially compromising the validity of causal conclusions. While it is possible to correct for biases if the underlying causal graph is known, this is rarely a feasible ask in practical scenarios. A common strategy is to adjust for all available covariates, yet this approach can yield biased treatment effect estimates, especially when post-treatment or unobserved variables are present. We propose RAMEN, an algorithm that produces unbiased treatment effect estimates by leveraging the heterogeneity of multiple data sources without the need to know or learn the underlying causal graph. Notably, RAMEN achieves doubly robust identification: it can identify the treatment effect whenever the causal parents of the treatment or those of the outcome are observed, and the node whose parents are observed satisfies an invariance assumption. Empirical evaluations on synthetic and real-world datasets show that our approach outperforms existing methods.
2503.14473	Jason Han	Jason Han, Nicholas S. DiBrita, Younghyun Cho, Hengrui Luo, Tirthak Patel	EnQode: Fast Amplitude Embedding for Quantum Machine Learning Using Classical Data	EnQode will appear in the Proceedings of the Design Automation Conference (DAC), 2025	null	null	null	quant-ph cs.ET cs.LG	http://creativecommons.org/licenses/by/4.0/	Amplitude embedding (AE) is essential in quantum machine learning (QML) for encoding classical data onto quantum circuits. However, conventional AE methods suffer from deep, variable-length circuits that introduce high output error due to extensive gate usage and variable error rates across samples, resulting in noise-driven inconsistencies that degrade model accuracy. We introduce EnQode, a fast AE technique based on symbolic representation that addresses these limitations by clustering dataset samples and solving for cluster mean states through a low-depth, machine-specific ansatz. Optimized to reduce physical gates and SWAP operations, EnQode ensures all samples face consistent, low noise levels by standardizing circuit depth and composition. With over 90% fidelity in data mapping, EnQode enables robust, high-performance QML on noisy intermediate-scale quantum (NISQ) devices. Our open-source solution provides a scalable and efficient alternative for integrating classical data with quantum models.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:48:03 GMT" } ]	2025-03-19T00:00:00	[ [ "Han", "Jason", "" ], [ "DiBrita", "Nicholas S.", "" ], [ "Cho", "Younghyun", "" ], [ "Luo", "Hengrui", "" ], [ "Patel", "Tirthak", "" ] ]	TITLE: EnQode: Fast Amplitude Embedding for Quantum Machine Learning Using Classical Data ABSTRACT: Amplitude embedding (AE) is essential in quantum machine learning (QML) for encoding classical data onto quantum circuits. However, conventional AE methods suffer from deep, variable-length circuits that introduce high output error due to extensive gate usage and variable error rates across samples, resulting in noise-driven inconsistencies that degrade model accuracy. We introduce EnQode, a fast AE technique based on symbolic representation that addresses these limitations by clustering dataset samples and solving for cluster mean states through a low-depth, machine-specific ansatz. Optimized to reduce physical gates and SWAP operations, EnQode ensures all samples face consistent, low noise levels by standardizing circuit depth and composition. With over 90% fidelity in data mapping, EnQode enables robust, high-performance QML on noisy intermediate-scale quantum (NISQ) devices. Our open-source solution provides a scalable and efficient alternative for integrating classical data with quantum models.
2503.14476	Qiying Yu	Qiying Yu, Zheng Zhang, Ruofei Zhu, Yufeng Yuan, Xiaochen Zuo, Yu Yue, Tiantian Fan, Gaohong Liu, Lingjun Liu, Xin Liu, Haibin Lin, Zhiqi Lin, Bole Ma, Guangming Sheng, Yuxuan Tong, Chi Zhang, Mofan Zhang, Wang Zhang, Hang Zhu, Jinhua Zhu, Jiaze Chen, Jiangjie Chen, Chengyi Wang, Hongli Yu, Weinan Dai, Yuxuan Song, Xiangpeng Wei, Hao Zhou, Jingjing Liu, Wei-Ying Ma, Ya-Qin Zhang, Lin Yan, Mu Qiao, Yonghui Wu, Mingxuan Wang	DAPO: An Open-Source LLM Reinforcement Learning System at Scale	Project Page: https://dapo-sia.github.io/	null	null	null	cs.LG cs.CL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Inference scaling empowers LLMs with unprecedented reasoning ability, with reinforcement learning as the core technique to elicit complex reasoning. However, key technical details of state-of-the-art reasoning LLMs are concealed (such as in OpenAI o1 blog and DeepSeek R1 technical report), thus the community still struggles to reproduce their RL training results. We propose the $\textbf{D}$ecoupled Clip and $\textbf{D}$ynamic s$\textbf{A}$mpling $\textbf{P}$olicy $\textbf{O}$ptimization ($\textbf{DAPO}$) algorithm, and fully open-source a state-of-the-art large-scale RL system that achieves 50 points on AIME 2024 using Qwen2.5-32B base model. Unlike previous works that withhold training details, we introduce four key techniques of our algorithm that make large-scale LLM RL a success. In addition, we open-source our training code, which is built on the verl framework, along with a carefully curated and processed dataset. These components of our open-source system enhance reproducibility and support future research in large-scale LLM RL.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:49:06 GMT" } ]	2025-03-19T00:00:00	[ [ "Yu", "Qiying", "" ], [ "Zhang", "Zheng", "" ], [ "Zhu", "Ruofei", "" ], [ "Yuan", "Yufeng", "" ], [ "Zuo", "Xiaochen", "" ], [ "Yue", "Yu", "" ], [ "Fan", "Tiantian", "" ], [ "Liu", "Gaohong", "" ], [ "Liu", "Lingjun", "" ], [ "Liu", "Xin", "" ], [ "Lin", "Haibin", "" ], [ "Lin", "Zhiqi", "" ], [ "Ma", "Bole", "" ], [ "Sheng", "Guangming", "" ], [ "Tong", "Yuxuan", "" ], [ "Zhang", "Chi", "" ], [ "Zhang", "Mofan", "" ], [ "Zhang", "Wang", "" ], [ "Zhu", "Hang", "" ], [ "Zhu", "Jinhua", "" ], [ "Chen", "Jiaze", "" ], [ "Chen", "Jiangjie", "" ], [ "Wang", "Chengyi", "" ], [ "Yu", "Hongli", "" ], [ "Dai", "Weinan", "" ], [ "Song", "Yuxuan", "" ], [ "Wei", "Xiangpeng", "" ], [ "Zhou", "Hao", "" ], [ "Liu", "Jingjing", "" ], [ "Ma", "Wei-Ying", "" ], [ "Zhang", "Ya-Qin", "" ], [ "Yan", "Lin", "" ], [ "Qiao", "Mu", "" ], [ "Wu", "Yonghui", "" ], [ "Wang", "Mingxuan", "" ] ]	TITLE: DAPO: An Open-Source LLM Reinforcement Learning System at Scale ABSTRACT: Inference scaling empowers LLMs with unprecedented reasoning ability, with reinforcement learning as the core technique to elicit complex reasoning. However, key technical details of state-of-the-art reasoning LLMs are concealed (such as in OpenAI o1 blog and DeepSeek R1 technical report), thus the community still struggles to reproduce their RL training results. We propose the $\textbf{D}$ecoupled Clip and $\textbf{D}$ynamic s$\textbf{A}$mpling $\textbf{P}$olicy $\textbf{O}$ptimization ($\textbf{DAPO}$) algorithm, and fully open-source a state-of-the-art large-scale RL system that achieves 50 points on AIME 2024 using Qwen2.5-32B base model. Unlike previous works that withhold training details, we introduce four key techniques of our algorithm that make large-scale LLM RL a success. In addition, we open-source our training code, which is built on the verl framework, along with a carefully curated and processed dataset. These components of our open-source system enhance reproducibility and support future research in large-scale LLM RL.
2503.14482	Yulin Pan	Yulin Pan, Xiangteng He, Chaojie Mao, Zhen Han, Zeyinzi Jiang, Jingfeng Zhang, Yu Liu	ICE-Bench: A Unified and Comprehensive Benchmark for Image Creating and Editing	17 pages	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Image generation has witnessed significant advancements in the past few years. However, evaluating the performance of image generation models remains a formidable challenge. In this paper, we propose ICE-Bench, a unified and comprehensive benchmark designed to rigorously assess image generation models. Its comprehensiveness could be summarized in the following key features: (1) Coarse-to-Fine Tasks: We systematically deconstruct image generation into four task categories: No-ref/Ref Image Creating/Editing, based on the presence or absence of source images and reference images. And further decompose them into 31 fine-grained tasks covering a broad spectrum of image generation requirements, culminating in a comprehensive benchmark. (2) Multi-dimensional Metrics: The evaluation framework assesses image generation capabilities across 6 dimensions: aesthetic quality, imaging quality, prompt following, source consistency, reference consistency, and controllability. 11 metrics are introduced to support the multi-dimensional evaluation. Notably, we introduce VLLM-QA, an innovative metric designed to assess the success of image editing by leveraging large models. (3) Hybrid Data: The data comes from real scenes and virtual generation, which effectively improves data diversity and alleviates the bias problem in model evaluation. Through ICE-Bench, we conduct a thorough analysis of existing generation models, revealing both the challenging nature of our benchmark and the gap between current model capabilities and real-world generation requirements. To foster further advancements in the field, we will open-source ICE-Bench, including its dataset, evaluation code, and models, thereby providing a valuable resource for the research community.	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:53:29 GMT" } ]	2025-03-19T00:00:00	[ [ "Pan", "Yulin", "" ], [ "He", "Xiangteng", "" ], [ "Mao", "Chaojie", "" ], [ "Han", "Zhen", "" ], [ "Jiang", "Zeyinzi", "" ], [ "Zhang", "Jingfeng", "" ], [ "Liu", "Yu", "" ] ]	TITLE: ICE-Bench: A Unified and Comprehensive Benchmark for Image Creating and Editing ABSTRACT: Image generation has witnessed significant advancements in the past few years. However, evaluating the performance of image generation models remains a formidable challenge. In this paper, we propose ICE-Bench, a unified and comprehensive benchmark designed to rigorously assess image generation models. Its comprehensiveness could be summarized in the following key features: (1) Coarse-to-Fine Tasks: We systematically deconstruct image generation into four task categories: No-ref/Ref Image Creating/Editing, based on the presence or absence of source images and reference images. And further decompose them into 31 fine-grained tasks covering a broad spectrum of image generation requirements, culminating in a comprehensive benchmark. (2) Multi-dimensional Metrics: The evaluation framework assesses image generation capabilities across 6 dimensions: aesthetic quality, imaging quality, prompt following, source consistency, reference consistency, and controllability. 11 metrics are introduced to support the multi-dimensional evaluation. Notably, we introduce VLLM-QA, an innovative metric designed to assess the success of image editing by leveraging large models. (3) Hybrid Data: The data comes from real scenes and virtual generation, which effectively improves data diversity and alleviates the bias problem in model evaluation. Through ICE-Bench, we conduct a thorough analysis of existing generation models, revealing both the challenging nature of our benchmark and the gap between current model capabilities and real-world generation requirements. To foster further advancements in the field, we will open-source ICE-Bench, including its dataset, evaluation code, and models, thereby providing a valuable resource for the research community.
2503.14483	Haoyu Guo	Haoyu Guo, He Zhu, Sida Peng, Haotong Lin, Yunzhi Yan, Tao Xie, Wenguan Wang, Xiaowei Zhou, Hujun Bao	Multi-view Reconstruction via SfM-guided Monocular Depth Estimation	CVPR 2025. Project page: https://zju3dv.github.io/murre/	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In this paper, we present a new method for multi-view geometric reconstruction. In recent years, large vision models have rapidly developed, performing excellently across various tasks and demonstrating remarkable generalization capabilities. Some works use large vision models for monocular depth estimation, which have been applied to facilitate multi-view reconstruction tasks in an indirect manner. Due to the ambiguity of the monocular depth estimation task, the estimated depth values are usually not accurate enough, limiting their utility in aiding multi-view reconstruction. We propose to incorporate SfM information, a strong multi-view prior, into the depth estimation process, thus enhancing the quality of depth prediction and enabling their direct application in multi-view geometric reconstruction. Experimental results on public real-world datasets show that our method significantly improves the quality of depth estimation compared to previous monocular depth estimation works. Additionally, we evaluate the reconstruction quality of our approach in various types of scenes including indoor, streetscape, and aerial views, surpassing state-of-the-art MVS methods. The code and supplementary materials are available at https://zju3dv.github.io/murre/ .	[ { "version": "v1", "created": "Tue, 18 Mar 2025 17:54:06 GMT" } ]	2025-03-19T00:00:00	[ [ "Guo", "Haoyu", "" ], [ "Zhu", "He", "" ], [ "Peng", "Sida", "" ], [ "Lin", "Haotong", "" ], [ "Yan", "Yunzhi", "" ], [ "Xie", "Tao", "" ], [ "Wang", "Wenguan", "" ], [ "Zhou", "Xiaowei", "" ], [ "Bao", "Hujun", "" ] ]	TITLE: Multi-view Reconstruction via SfM-guided Monocular Depth Estimation ABSTRACT: In this paper, we present a new method for multi-view geometric reconstruction. In recent years, large vision models have rapidly developed, performing excellently across various tasks and demonstrating remarkable generalization capabilities. Some works use large vision models for monocular depth estimation, which have been applied to facilitate multi-view reconstruction tasks in an indirect manner. Due to the ambiguity of the monocular depth estimation task, the estimated depth values are usually not accurate enough, limiting their utility in aiding multi-view reconstruction. We propose to incorporate SfM information, a strong multi-view prior, into the depth estimation process, thus enhancing the quality of depth prediction and enabling their direct application in multi-view geometric reconstruction. Experimental results on public real-world datasets show that our method significantly improves the quality of depth estimation compared to previous monocular depth estimation works. Additionally, we evaluate the reconstruction quality of our approach in various types of scenes including indoor, streetscape, and aerial views, surpassing state-of-the-art MVS methods. The code and supplementary materials are available at https://zju3dv.github.io/murre/ .
1902.02595	Antonino Sabetta	Serena E. Ponta, Henrik Plate, Antonino Sabetta, Michele Bezzi, C\'edric Dangremont	A Manually-Curated Dataset of Fixes to Vulnerabilities of Open-Source Software	This is a pre-print version of the paper that appears in the proceedings of The 16th International Conference on Mining Software Repositories (MSR), Data Showcase track	Proceedings of The 16th International Conference on Mining Software Repositories (Data Showcase track), 2019	10.1109/MSR.2019.0006	null	cs.SE cs.CR cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Advancing our understanding of software vulnerabilities, automating their identification, the analysis of their impact, and ultimately their mitigation is necessary to enable the development of software that is more secure. While operating a vulnerability assessment tool that we developed and that is currently used by hundreds of development units at SAP, we manually collected and curated a dataset of vulnerabilities of open-source software and the commits fixing them. The data was obtained both from the National Vulnerability Database (NVD) and from project-specific Web resources that we monitor on a continuous basis. From that data, we extracted a dataset that maps 624 publicly disclosed vulnerabilities affecting 205 distinct open-source Java projects, used in SAP products or internal tools, onto the 1282 commits that fix them. Out of 624 vulnerabilities, 29 do not have a CVE identifier at all and 46, which do have a CVE identifier assigned by a numbering authority, are not available in the NVD yet. The dataset is released under an open-source license, together with supporting scripts that allow researchers to automatically retrieve the actual content of the commits from the corresponding repositories and to augment the attributes available for each instance. Also, these scripts allow to complement the dataset with additional instances that are not security fixes (which is useful, for example, in machine learning applications). Our dataset has been successfully used to train classifiers that could automatically identify security-relevant commits in code repositories. The release of this dataset and the supporting code as open-source will allow future research to be based on data of industrial relevance; also, it represents a concrete step towards making the maintenance of this dataset a shared effort involving open-source communities, academia, and the industry.	[ { "version": "v1", "created": "Thu, 7 Feb 2019 12:47:13 GMT" }, { "version": "v2", "created": "Fri, 8 Feb 2019 09:06:58 GMT" }, { "version": "v3", "created": "Tue, 19 Mar 2019 10:33:34 GMT" } ]	2025-03-18T00:00:00	[ [ "Ponta", "Serena E.", "" ], [ "Plate", "Henrik", "" ], [ "Sabetta", "Antonino", "" ], [ "Bezzi", "Michele", "" ], [ "Dangremont", "Cédric", "" ] ]	TITLE: A Manually-Curated Dataset of Fixes to Vulnerabilities of Open-Source Software ABSTRACT: Advancing our understanding of software vulnerabilities, automating their identification, the analysis of their impact, and ultimately their mitigation is necessary to enable the development of software that is more secure. While operating a vulnerability assessment tool that we developed and that is currently used by hundreds of development units at SAP, we manually collected and curated a dataset of vulnerabilities of open-source software and the commits fixing them. The data was obtained both from the National Vulnerability Database (NVD) and from project-specific Web resources that we monitor on a continuous basis. From that data, we extracted a dataset that maps 624 publicly disclosed vulnerabilities affecting 205 distinct open-source Java projects, used in SAP products or internal tools, onto the 1282 commits that fix them. Out of 624 vulnerabilities, 29 do not have a CVE identifier at all and 46, which do have a CVE identifier assigned by a numbering authority, are not available in the NVD yet. The dataset is released under an open-source license, together with supporting scripts that allow researchers to automatically retrieve the actual content of the commits from the corresponding repositories and to augment the attributes available for each instance. Also, these scripts allow to complement the dataset with additional instances that are not security fixes (which is useful, for example, in machine learning applications). Our dataset has been successfully used to train classifiers that could automatically identify security-relevant commits in code repositories. The release of this dataset and the supporting code as open-source will allow future research to be based on data of industrial relevance; also, it represents a concrete step towards making the maintenance of this dataset a shared effort involving open-source communities, academia, and the industry.
1906.08256	Bala Krishnamoorthy	Dustin L. Arendt, Matthew Broussard, Bala Krishnamoorthy, Nathaniel Saul, Amber Thrall	Steinhaus Filtration and Stable Paths in the Mapper	Proof of stability added; to appear in SoCG 2025	null	null	null	cs.LG cs.CG math.AT	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We define a new filtration called the Steinhaus filtration built from a single cover based on a generalized Steinhaus distance, a generalization of Jaccard distance. The homology persistence module of a Steinhaus filtration with infinitely many cover elements may not be $q$-tame, even when the covers are in a totally bounded space. While this may pose a challenge to derive stability results, we show that the Steinhaus filtration is stable when the cover is finite. We show that while the \v{C}ech and Steinhaus filtrations are not isomorphic in general, they are isomorphic for a finite point set in dimension one. Furthermore, the VR filtration completely determines the $1$-skeleton of the Steinhaus filtration in arbitrary dimension. We then develop a language and theory for stable paths within the Steinhaus filtration. We demonstrate how the framework can be applied to several applications where a standard metric may not be defined but a cover is readily available. We introduce a new perspective for modeling recommendation system datasets. As an example, we look at a movies dataset and we find the stable paths identified in our framework represent a sequence of movies constituting a gentle transition and ordering from one genre to another. For explainable machine learning, we apply the Mapper algorithm for model induction by building a filtration from a single Mapper complex, and provide explanations in the form of stable paths between subpopulations. For illustration, we build a Mapper complex from a supervised machine learning model trained on the FashionMNIST dataset. Stable paths in the Steinhaus filtration provide improved explanations of relationships between subpopulations of images.	[ { "version": "v1", "created": "Wed, 19 Jun 2019 05:02:42 GMT" }, { "version": "v2", "created": "Mon, 14 Sep 2020 09:23:47 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 18:18:17 GMT" } ]	2025-03-18T00:00:00	[ [ "Arendt", "Dustin L.", "" ], [ "Broussard", "Matthew", "" ], [ "Krishnamoorthy", "Bala", "" ], [ "Saul", "Nathaniel", "" ], [ "Thrall", "Amber", "" ] ]	TITLE: Steinhaus Filtration and Stable Paths in the Mapper ABSTRACT: We define a new filtration called the Steinhaus filtration built from a single cover based on a generalized Steinhaus distance, a generalization of Jaccard distance. The homology persistence module of a Steinhaus filtration with infinitely many cover elements may not be $q$-tame, even when the covers are in a totally bounded space. While this may pose a challenge to derive stability results, we show that the Steinhaus filtration is stable when the cover is finite. We show that while the \v{C}ech and Steinhaus filtrations are not isomorphic in general, they are isomorphic for a finite point set in dimension one. Furthermore, the VR filtration completely determines the $1$-skeleton of the Steinhaus filtration in arbitrary dimension. We then develop a language and theory for stable paths within the Steinhaus filtration. We demonstrate how the framework can be applied to several applications where a standard metric may not be defined but a cover is readily available. We introduce a new perspective for modeling recommendation system datasets. As an example, we look at a movies dataset and we find the stable paths identified in our framework represent a sequence of movies constituting a gentle transition and ordering from one genre to another. For explainable machine learning, we apply the Mapper algorithm for model induction by building a filtration from a single Mapper complex, and provide explanations in the form of stable paths between subpopulations. For illustration, we build a Mapper complex from a supervised machine learning model trained on the FashionMNIST dataset. Stable paths in the Steinhaus filtration provide improved explanations of relationships between subpopulations of images.
1906.10969	Mirco Schoenfeld	Mirco Schoenfeld, Steffen Eckhard, Ronny Patz, Hilde van Meegdenburg, Antonio Pires	The UN Security Council debates 1992-2023	The UN Security Council Debates corpus is available online at https://doi.org/10.7910/DVN/KGVSYH	null	null	null	cs.DL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper presents an updated dataset containing 106,302 speeches held in the public meetings of the UN Security Council (UNSC) between 1992 and 2023. The dataset is based on publicly available meeting transcripts with the S/PV document symbol and includes the full substance of individual speeches as well as automatically extracted and manually corrected metadata on the speaker, the position of the speech in the sequence of speeches of a meeting, and the date of the speech. After contextualizing the dataset in recent research on the UNSC, the paper presents descriptive statistics on UNSC meetings and speeches that characterize the period covered by the dataset. Data highlight the extensive presence of the UN bureaucracy in UNSC meetings as well as an emerging trend towards more lengthy open UNSC debates. These open debates cover key issues that have emerged only during the period that is covered by the dataset, for example the debates relating to Women, Peace and Security or Climate-related Disasters. The corpus is available online: https://doi.org/10.7910/DVN/KGVSYH	[ { "version": "v1", "created": "Wed, 26 Jun 2019 10:57:34 GMT" }, { "version": "v2", "created": "Fri, 4 Oct 2019 12:30:38 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 10:07:09 GMT" } ]	2025-03-18T00:00:00	[ [ "Schoenfeld", "Mirco", "" ], [ "Eckhard", "Steffen", "" ], [ "Patz", "Ronny", "" ], [ "van Meegdenburg", "Hilde", "" ], [ "Pires", "Antonio", "" ] ]	TITLE: The UN Security Council debates 1992-2023 ABSTRACT: This paper presents an updated dataset containing 106,302 speeches held in the public meetings of the UN Security Council (UNSC) between 1992 and 2023. The dataset is based on publicly available meeting transcripts with the S/PV document symbol and includes the full substance of individual speeches as well as automatically extracted and manually corrected metadata on the speaker, the position of the speech in the sequence of speeches of a meeting, and the date of the speech. After contextualizing the dataset in recent research on the UNSC, the paper presents descriptive statistics on UNSC meetings and speeches that characterize the period covered by the dataset. Data highlight the extensive presence of the UN bureaucracy in UNSC meetings as well as an emerging trend towards more lengthy open UNSC debates. These open debates cover key issues that have emerged only during the period that is covered by the dataset, for example the debates relating to Women, Peace and Security or Climate-related Disasters. The corpus is available online: https://doi.org/10.7910/DVN/KGVSYH
1911.07308	Tsu-Jui Fu	Tsu-Jui Fu, Xin Eric Wang, Matthew Peterson, Scott Grafton, Miguel Eckstein, William Yang Wang	Counterfactual Vision-and-Language Navigation via Adversarial Path Sampling	ECCV'20 (Spotlight)	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Vision-and-Language Navigation (VLN) is a task where agents must decide how to move through a 3D environment to reach a goal by grounding natural language instructions to the visual surroundings. One of the problems of the VLN task is data scarcity since it is difficult to collect enough navigation paths with human-annotated instructions for interactive environments. In this paper, we explore the use of counterfactual thinking as a human-inspired data augmentation method that results in robust models. Counterfactual thinking is a concept that describes the human propensity to create possible alternatives to life events that have already occurred. We propose an adversarial-driven counterfactual reasoning model that can consider effective conditions instead of low-quality augmented data. In particular, we present a model-agnostic adversarial path sampler (APS) that learns to sample challenging paths that force the navigator to improve based on the navigation performance. APS also serves to do pre-exploration of unseen environments to strengthen the model's ability to generalize. We evaluate the influence of APS on the performance of different VLN baseline models using the room-to-room dataset (R2R). The results show that the adversarial training process with our proposed APS benefits VLN models under both seen and unseen environments. And the pre-exploration process can further gain additional improvements under unseen environments.	[ { "version": "v1", "created": "Sun, 17 Nov 2019 18:02:51 GMT" }, { "version": "v2", "created": "Tue, 7 Jul 2020 15:46:58 GMT" }, { "version": "v3", "created": "Fri, 17 Jul 2020 00:18:45 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 16:53:11 GMT" } ]	2025-03-18T00:00:00	[ [ "Fu", "Tsu-Jui", "" ], [ "Wang", "Xin Eric", "" ], [ "Peterson", "Matthew", "" ], [ "Grafton", "Scott", "" ], [ "Eckstein", "Miguel", "" ], [ "Wang", "William Yang", "" ] ]	TITLE: Counterfactual Vision-and-Language Navigation via Adversarial Path Sampling ABSTRACT: Vision-and-Language Navigation (VLN) is a task where agents must decide how to move through a 3D environment to reach a goal by grounding natural language instructions to the visual surroundings. One of the problems of the VLN task is data scarcity since it is difficult to collect enough navigation paths with human-annotated instructions for interactive environments. In this paper, we explore the use of counterfactual thinking as a human-inspired data augmentation method that results in robust models. Counterfactual thinking is a concept that describes the human propensity to create possible alternatives to life events that have already occurred. We propose an adversarial-driven counterfactual reasoning model that can consider effective conditions instead of low-quality augmented data. In particular, we present a model-agnostic adversarial path sampler (APS) that learns to sample challenging paths that force the navigator to improve based on the navigation performance. APS also serves to do pre-exploration of unseen environments to strengthen the model's ability to generalize. We evaluate the influence of APS on the performance of different VLN baseline models using the room-to-room dataset (R2R). The results show that the adversarial training process with our proposed APS benefits VLN models under both seen and unseen environments. And the pre-exploration process can further gain additional improvements under unseen environments.
2009.09566	Tsu-Jui Fu	Tsu-Jui Fu, Xin Eric Wang, Scott Grafton, Miguel Eckstein, William Yang Wang	SSCR: Iterative Language-Based Image Editing via Self-Supervised Counterfactual Reasoning	EMNLP'20 (Oral)	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Iterative Language-Based Image Editing (IL-BIE) tasks follow iterative instructions to edit images step by step. Data scarcity is a significant issue for ILBIE as it is challenging to collect large-scale examples of images before and after instruction-based changes. However, humans still accomplish these editing tasks even when presented with an unfamiliar image-instruction pair. Such ability results from counterfactual thinking and the ability to think about alternatives to events that have happened already. In this paper, we introduce a Self-Supervised Counterfactual Reasoning (SSCR) framework that incorporates counterfactual thinking to overcome data scarcity. SSCR allows the model to consider out-of-distribution instructions paired with previous images. With the help of cross-task consistency (CTC), we train these counterfactual instructions in a self-supervised scenario. Extensive results show that SSCR improves the correctness of ILBIE in terms of both object identity and position, establishing a new state of the art (SOTA) on two IBLIE datasets (i-CLEVR and CoDraw). Even with only 50% of the training data, SSCR achieves a comparable result to using complete data.	[ { "version": "v1", "created": "Mon, 21 Sep 2020 01:45:58 GMT" }, { "version": "v2", "created": "Tue, 29 Sep 2020 00:24:25 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 16:55:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Fu", "Tsu-Jui", "" ], [ "Wang", "Xin Eric", "" ], [ "Grafton", "Scott", "" ], [ "Eckstein", "Miguel", "" ], [ "Wang", "William Yang", "" ] ]	TITLE: SSCR: Iterative Language-Based Image Editing via Self-Supervised Counterfactual Reasoning ABSTRACT: Iterative Language-Based Image Editing (IL-BIE) tasks follow iterative instructions to edit images step by step. Data scarcity is a significant issue for ILBIE as it is challenging to collect large-scale examples of images before and after instruction-based changes. However, humans still accomplish these editing tasks even when presented with an unfamiliar image-instruction pair. Such ability results from counterfactual thinking and the ability to think about alternatives to events that have happened already. In this paper, we introduce a Self-Supervised Counterfactual Reasoning (SSCR) framework that incorporates counterfactual thinking to overcome data scarcity. SSCR allows the model to consider out-of-distribution instructions paired with previous images. With the help of cross-task consistency (CTC), we train these counterfactual instructions in a self-supervised scenario. Extensive results show that SSCR improves the correctness of ILBIE in terms of both object identity and position, establishing a new state of the art (SOTA) on two IBLIE datasets (i-CLEVR and CoDraw). Even with only 50% of the training data, SSCR achieves a comparable result to using complete data.
2111.09304	Archismita Dalal	Archismita Dalal, Mohsen Bagherimehrab and Barry C. Sanders	Quantum-Assisted Support Vector Regression	15 pages, 5 figures	Quantum Inf Process 24, 82 (2025)	10.1007/s11128-025-04674-0	null	quant-ph cs.LG	http://creativecommons.org/licenses/by/4.0/	A popular machine-learning model for regression tasks, including stock-market prediction, weather forecasting and real-estate pricing, is the classical support vector regression (SVR). However, a practically realisable quantum SVR remains to be formulated. We devise annealing-based algorithms, namely simulated and quantum-classical hybrid, for training two SVR models and compare their empirical performances against the SVR implementation of Python's scikit-learn package for facial-landmark detection (FLD), a particular use case for SVR. Our method is to derive a quadratic-unconstrained-binary formulation for the optimisation problem used for training a SVR model and solve this problem using annealing. Using D-Wave's hybrid solver, we construct a quantum-assisted SVR model, thereby demonstrating a slight advantage over classical models regarding FLD accuracy. Furthermore, we observe that annealing-based SVR models predict landmarks with lower variances compared to the SVR models trained by gradient-based methods. Our work is a proof-of-concept example for applying quantum-assisted SVR to a supervised-learning task with a small training dataset.	[ { "version": "v1", "created": "Wed, 17 Nov 2021 18:57:10 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 19:33:03 GMT" } ]	2025-03-18T00:00:00	[ [ "Dalal", "Archismita", "" ], [ "Bagherimehrab", "Mohsen", "" ], [ "Sanders", "Barry C.", "" ] ]	TITLE: Quantum-Assisted Support Vector Regression ABSTRACT: A popular machine-learning model for regression tasks, including stock-market prediction, weather forecasting and real-estate pricing, is the classical support vector regression (SVR). However, a practically realisable quantum SVR remains to be formulated. We devise annealing-based algorithms, namely simulated and quantum-classical hybrid, for training two SVR models and compare their empirical performances against the SVR implementation of Python's scikit-learn package for facial-landmark detection (FLD), a particular use case for SVR. Our method is to derive a quadratic-unconstrained-binary formulation for the optimisation problem used for training a SVR model and solve this problem using annealing. Using D-Wave's hybrid solver, we construct a quantum-assisted SVR model, thereby demonstrating a slight advantage over classical models regarding FLD accuracy. Furthermore, we observe that annealing-based SVR models predict landmarks with lower variances compared to the SVR models trained by gradient-based methods. Our work is a proof-of-concept example for applying quantum-assisted SVR to a supervised-learning task with a small training dataset.
2202.01602	Shahin Jabbari	Satyapriya Krishna, Tessa Han, Alex Gu, Steven Wu, Shahin Jabbari, Himabindu Lakkaraju	The Disagreement Problem in Explainable Machine Learning: A Practitioner's Perspective	Published in Transactions on Machine Learning Research (TMLR). Added a misplaced reference	null	null	null	cs.LG cs.AI	http://creativecommons.org/licenses/by/4.0/	As various post hoc explanation methods are increasingly being leveraged to explain complex models in high-stakes settings, it becomes critical to develop a deeper understanding of if and when the explanations output by these methods disagree with each other, and how such disagreements are resolved in practice. However, there is little to no research that provides answers to these critical questions. In this work, we introduce and study the disagreement problem in explainable machine learning. More specifically, we formalize the notion of disagreement between explanations, analyze how often such disagreements occur in practice, and how practitioners resolve these disagreements. We first conduct interviews with data scientists to understand what constitutes disagreement between explanations generated by different methods for the same model prediction and introduce a novel quantitative framework to formalize this understanding. We then leverage this framework to carry out a rigorous empirical analysis with four real-world datasets, six state-of-the-art post hoc explanation methods, and six different predictive models, to measure the extent of disagreement between the explanations generated by various popular explanation methods. In addition, we carry out an online user study with data scientists to understand how they resolve the aforementioned disagreements. Our results indicate that (1) state-of-the-art explanation methods often disagree in terms of the explanations they output, and (2) machine learning practitioners often employ ad hoc heuristics when resolving such disagreements. These findings suggest that practitioners may be relying on misleading explanations when making consequential decisions. They also underscore the importance of developing principled frameworks for effectively evaluating and comparing explanations output by various explanation techniques.	[ { "version": "v1", "created": "Thu, 3 Feb 2022 14:19:23 GMT" }, { "version": "v2", "created": "Fri, 4 Feb 2022 01:46:00 GMT" }, { "version": "v3", "created": "Tue, 8 Feb 2022 07:24:09 GMT" }, { "version": "v4", "created": "Mon, 8 Jul 2024 12:11:38 GMT" }, { "version": "v5", "created": "Mon, 17 Mar 2025 15:00:52 GMT" } ]	2025-03-18T00:00:00	[ [ "Krishna", "Satyapriya", "" ], [ "Han", "Tessa", "" ], [ "Gu", "Alex", "" ], [ "Wu", "Steven", "" ], [ "Jabbari", "Shahin", "" ], [ "Lakkaraju", "Himabindu", "" ] ]	TITLE: The Disagreement Problem in Explainable Machine Learning: A Practitioner's Perspective ABSTRACT: As various post hoc explanation methods are increasingly being leveraged to explain complex models in high-stakes settings, it becomes critical to develop a deeper understanding of if and when the explanations output by these methods disagree with each other, and how such disagreements are resolved in practice. However, there is little to no research that provides answers to these critical questions. In this work, we introduce and study the disagreement problem in explainable machine learning. More specifically, we formalize the notion of disagreement between explanations, analyze how often such disagreements occur in practice, and how practitioners resolve these disagreements. We first conduct interviews with data scientists to understand what constitutes disagreement between explanations generated by different methods for the same model prediction and introduce a novel quantitative framework to formalize this understanding. We then leverage this framework to carry out a rigorous empirical analysis with four real-world datasets, six state-of-the-art post hoc explanation methods, and six different predictive models, to measure the extent of disagreement between the explanations generated by various popular explanation methods. In addition, we carry out an online user study with data scientists to understand how they resolve the aforementioned disagreements. Our results indicate that (1) state-of-the-art explanation methods often disagree in terms of the explanations they output, and (2) machine learning practitioners often employ ad hoc heuristics when resolving such disagreements. These findings suggest that practitioners may be relying on misleading explanations when making consequential decisions. They also underscore the importance of developing principled frameworks for effectively evaluating and comparing explanations output by various explanation techniques.
2211.03983	Chengchun Shi	Liyuan Hu and Mengbing Li and Chengchun Shi and Zhenke Wu and Piotr Fryzlewicz	Doubly Inhomogeneous Reinforcement Learning	null	null	null	null	stat.ML cs.AI cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	This paper studies reinforcement learning (RL) in doubly inhomogeneous environments under temporal non-stationarity and subject heterogeneity. In a number of applications, it is commonplace to encounter datasets generated by system dynamics that may change over time and population, challenging high-quality sequential decision making. Nonetheless, most existing RL solutions require either temporal stationarity or subject homogeneity, which would result in sub-optimal policies if both assumptions were violated. To address both challenges simultaneously, we propose an original algorithm to determine the ``best data chunks" that display similar dynamics over time and across individuals for policy learning, which alternates between most recent change point detection and cluster identification. Our method is general, and works with a wide range of clustering and change point detection algorithms. It is multiply robust in the sense that it takes multiple initial estimators as input and only requires one of them to be consistent. Moreover, by borrowing information over time and population, it allows us to detect weaker signals and has better convergence properties when compared to applying the clustering algorithm per time or the change point detection algorithm per subject. Empirically, we demonstrate the usefulness of our method through extensive simulations and a real data application.	[ { "version": "v1", "created": "Tue, 8 Nov 2022 03:41:14 GMT" }, { "version": "v2", "created": "Sat, 12 Nov 2022 09:35:42 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 17:25:47 GMT" } ]	2025-03-18T00:00:00	[ [ "Hu", "Liyuan", "" ], [ "Li", "Mengbing", "" ], [ "Shi", "Chengchun", "" ], [ "Wu", "Zhenke", "" ], [ "Fryzlewicz", "Piotr", "" ] ]	TITLE: Doubly Inhomogeneous Reinforcement Learning ABSTRACT: This paper studies reinforcement learning (RL) in doubly inhomogeneous environments under temporal non-stationarity and subject heterogeneity. In a number of applications, it is commonplace to encounter datasets generated by system dynamics that may change over time and population, challenging high-quality sequential decision making. Nonetheless, most existing RL solutions require either temporal stationarity or subject homogeneity, which would result in sub-optimal policies if both assumptions were violated. To address both challenges simultaneously, we propose an original algorithm to determine the ``best data chunks" that display similar dynamics over time and across individuals for policy learning, which alternates between most recent change point detection and cluster identification. Our method is general, and works with a wide range of clustering and change point detection algorithms. It is multiply robust in the sense that it takes multiple initial estimators as input and only requires one of them to be consistent. Moreover, by borrowing information over time and population, it allows us to detect weaker signals and has better convergence properties when compared to applying the clustering algorithm per time or the change point detection algorithm per subject. Empirically, we demonstrate the usefulness of our method through extensive simulations and a real data application.
2211.10724	Youwei Huang	Youwei Huang, Sen Fang, Jianwen Li, Jiachun Tao, Bin Hu, and Tao Zhang	Deep Smart Contract Intent Detection	12 pages, 8 figures, conference	null	10.1109/SANER64311.2025.00020	null	cs.SE cs.LG	http://creativecommons.org/licenses/by/4.0/	In recent years, research in software security has concentrated on identifying vulnerabilities in smart contracts to prevent significant losses of crypto assets on blockchains. Despite early successes in this area, detecting developers' intents in smart contracts has become a more pressing issue, as malicious intents have caused substantial financial losses. Unfortunately, existing research lacks effective methods for detecting development intents in smart contracts. To address this gap, we propose \textsc{SmartIntentNN} (Smart Contract Intent Neural Network), a deep learning model designed to automatically detect development intents in smart contracts. \textsc{SmartIntentNN} leverages a pre-trained sentence encoder to generate contextual representations of smart contracts, employs a K-means clustering model to identify and highlight prominent intent features, and utilizes a bidirectional LSTM-based deep neural network for multi-label classification. We trained and evaluated \textsc{SmartIntentNN} on a dataset containing over 40,000 real-world smart contracts, employing self-comparison baselines in our experimental setup. The results show that \textsc{SmartIntentNN} achieves an F1-score of 0.8633 in identifying intents across 10 distinct categories, outperforming all baselines and addressing the gap in smart contract detection by incorporating intent analysis.	[ { "version": "v1", "created": "Sat, 19 Nov 2022 15:40:26 GMT" }, { "version": "v2", "created": "Thu, 17 Oct 2024 02:48:51 GMT" }, { "version": "v3", "created": "Thu, 26 Dec 2024 13:10:25 GMT" } ]	2025-03-18T00:00:00	[ [ "Huang", "Youwei", "" ], [ "Fang", "Sen", "" ], [ "Li", "Jianwen", "" ], [ "Tao", "Jiachun", "" ], [ "Hu", "Bin", "" ], [ "Zhang", "Tao", "" ] ]	TITLE: Deep Smart Contract Intent Detection ABSTRACT: In recent years, research in software security has concentrated on identifying vulnerabilities in smart contracts to prevent significant losses of crypto assets on blockchains. Despite early successes in this area, detecting developers' intents in smart contracts has become a more pressing issue, as malicious intents have caused substantial financial losses. Unfortunately, existing research lacks effective methods for detecting development intents in smart contracts. To address this gap, we propose \textsc{SmartIntentNN} (Smart Contract Intent Neural Network), a deep learning model designed to automatically detect development intents in smart contracts. \textsc{SmartIntentNN} leverages a pre-trained sentence encoder to generate contextual representations of smart contracts, employs a K-means clustering model to identify and highlight prominent intent features, and utilizes a bidirectional LSTM-based deep neural network for multi-label classification. We trained and evaluated \textsc{SmartIntentNN} on a dataset containing over 40,000 real-world smart contracts, employing self-comparison baselines in our experimental setup. The results show that \textsc{SmartIntentNN} achieves an F1-score of 0.8633 in identifying intents across 10 distinct categories, outperforming all baselines and addressing the gap in smart contract detection by incorporating intent analysis.
2211.10760	Javier Mar\'in	Javier Marin	Evaluating Synthetic Tabular Data Generated To Augment Small Sample Datasets	null	null	null	null	cs.LG stat.ML	http://creativecommons.org/licenses/by-sa/4.0/	This work proposes a method to evaluate synthetic tabular data generated to augment small sample datasets. While data augmentation techniques can increase sample counts for machine learning applications, traditional validation approaches fail when applied to extremely limited sample sizes. Our experiments across four datasets reveal significant inconsistencies between global metrics and topological measures, with statistical tests producing unreliable significance values due to insufficient sample sizes. We demonstrate that common metrics like propensity scoring and MMD often suggest similarity where fundamental topological differences exist. Our proposed normalized Bottleneck distance based metric provides complementary insights but suffers from high variability across experimental runs and occasional values exceeding theoretical bounds, showing inherent instability in topological approaches for very small datasets. These findings highlight the critical need for multi-faceted evaluation methodologies when validating synthetic data generated from limited samples, as no single metric reliably captures both distributional and structural similarity.	[ { "version": "v1", "created": "Sat, 19 Nov 2022 18:18:52 GMT" }, { "version": "v2", "created": "Thu, 15 Dec 2022 15:00:12 GMT" }, { "version": "v3", "created": "Sat, 21 Jan 2023 09:50:45 GMT" }, { "version": "v4", "created": "Mon, 11 Nov 2024 11:04:06 GMT" }, { "version": "v5", "created": "Fri, 14 Mar 2025 18:08:54 GMT" } ]	2025-03-18T00:00:00	[ [ "Marin", "Javier", "" ] ]	TITLE: Evaluating Synthetic Tabular Data Generated To Augment Small Sample Datasets ABSTRACT: This work proposes a method to evaluate synthetic tabular data generated to augment small sample datasets. While data augmentation techniques can increase sample counts for machine learning applications, traditional validation approaches fail when applied to extremely limited sample sizes. Our experiments across four datasets reveal significant inconsistencies between global metrics and topological measures, with statistical tests producing unreliable significance values due to insufficient sample sizes. We demonstrate that common metrics like propensity scoring and MMD often suggest similarity where fundamental topological differences exist. Our proposed normalized Bottleneck distance based metric provides complementary insights but suffers from high variability across experimental runs and occasional values exceeding theoretical bounds, showing inherent instability in topological approaches for very small datasets. These findings highlight the critical need for multi-faceted evaluation methodologies when validating synthetic data generated from limited samples, as no single metric reliably captures both distributional and structural similarity.
2302.02150	Dimitris Iakovidis	Dimitrios E. Diamantis, Panagiota Gatoula, Anastasios Koulaouzidis, and Dimitris K. Iakovidis	This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation	10 pages	IEEE Access, 12, 25668-25683 (2024)	10.1109/ACCESS.2024.3366801	null	cs.CV cs.LG eess.IV	http://creativecommons.org/licenses/by-nc-nd/4.0/	Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems. To this end, Generative Adversarial Networks (GANs) have been mainly applied to support the algorithm training process by generating synthetic images for data augmentation. However, in the field of Wireless Capsule Endoscopy (WCE), the limited content diversity and size of existing publicly available annotated datasets, adversely affect both the training stability and synthesis performance of GANs. Aiming to a viable solution for WCE image synthesis, a novel Variational Autoencoder architecture is proposed, namely "This Intestine Does not Exist" (TIDE). The proposed architecture comprises multiscale feature extraction convolutional blocks and residual connections, which enable the generation of high-quality and diverse datasets even with a limited number of training images. Contrary to the current approaches, which are oriented towards the augmentation of the available datasets, this study demonstrates that using TIDE, real WCE datasets can be fully substituted by artificially generated ones, without compromising classification performance. Furthermore, qualitative and user evaluation studies by experienced WCE specialists, validate from a medical viewpoint that both the normal and abnormal WCE images synthesized by TIDE are sufficiently realistic.	[ { "version": "v1", "created": "Sat, 4 Feb 2023 11:49:38 GMT" }, { "version": "v2", "created": "Tue, 7 Feb 2023 03:50:25 GMT" } ]	2025-03-18T00:00:00	[ [ "Diamantis", "Dimitrios E.", "" ], [ "Gatoula", "Panagiota", "" ], [ "Koulaouzidis", "Anastasios", "" ], [ "Iakovidis", "Dimitris K.", "" ] ]	TITLE: This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation ABSTRACT: Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems. To this end, Generative Adversarial Networks (GANs) have been mainly applied to support the algorithm training process by generating synthetic images for data augmentation. However, in the field of Wireless Capsule Endoscopy (WCE), the limited content diversity and size of existing publicly available annotated datasets, adversely affect both the training stability and synthesis performance of GANs. Aiming to a viable solution for WCE image synthesis, a novel Variational Autoencoder architecture is proposed, namely "This Intestine Does not Exist" (TIDE). The proposed architecture comprises multiscale feature extraction convolutional blocks and residual connections, which enable the generation of high-quality and diverse datasets even with a limited number of training images. Contrary to the current approaches, which are oriented towards the augmentation of the available datasets, this study demonstrates that using TIDE, real WCE datasets can be fully substituted by artificially generated ones, without compromising classification performance. Furthermore, qualitative and user evaluation studies by experienced WCE specialists, validate from a medical viewpoint that both the normal and abnormal WCE images synthesized by TIDE are sufficiently realistic.
2305.00767	Cong Cao	Huanjing Yue, Cong Cao, Lei Liao, and Jingyu Yang	RViDeformer: Efficient Raw Video Denoising Transformer with a Larger Benchmark Dataset	Accepted by TCSVT 2025	null	null	null	cs.CV cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In recent years, raw video denoising has garnered increased attention due to the consistency with the imaging process and well-studied noise modeling in the raw domain. However, two problems still hinder the denoising performance. Firstly, there is no large dataset with realistic motions for supervised raw video denoising, as capturing noisy and clean frames for real dynamic scenes is difficult. To address this, we propose recapturing existing high-resolution videos displayed on a 4K screen with high-low ISO settings to construct noisy-clean paired frames. In this way, we construct a video denoising dataset (named as ReCRVD) with 120 groups of noisy-clean videos, whose ISO values ranging from 1600 to 25600. Secondly, while non-local temporal-spatial attention is beneficial for denoising, it often leads to heavy computation costs. We propose an efficient raw video denoising transformer network (RViDeformer) that explores both short and long-distance correlations. Specifically, we propose multi-branch spatial and temporal attention modules, which explore the patch correlations from local window, local low-resolution window, global downsampled window, and neighbor-involved window, and then they are fused together. We employ reparameterization to reduce computation costs. Our network is trained in both supervised and unsupervised manners, achieving the best performance compared with state-of-the-art methods. Additionally, the model trained with our proposed dataset (ReCRVD) outperforms the model trained with previous benchmark dataset (CRVD) when evaluated on the real-world outdoor noisy videos. Our code and dataset are available at https://github.com/cao-cong/RViDeformer.	[ { "version": "v1", "created": "Mon, 1 May 2023 11:06:58 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 10:07:37 GMT" } ]	2025-03-18T00:00:00	[ [ "Yue", "Huanjing", "" ], [ "Cao", "Cong", "" ], [ "Liao", "Lei", "" ], [ "Yang", "Jingyu", "" ] ]	TITLE: RViDeformer: Efficient Raw Video Denoising Transformer with a Larger Benchmark Dataset ABSTRACT: In recent years, raw video denoising has garnered increased attention due to the consistency with the imaging process and well-studied noise modeling in the raw domain. However, two problems still hinder the denoising performance. Firstly, there is no large dataset with realistic motions for supervised raw video denoising, as capturing noisy and clean frames for real dynamic scenes is difficult. To address this, we propose recapturing existing high-resolution videos displayed on a 4K screen with high-low ISO settings to construct noisy-clean paired frames. In this way, we construct a video denoising dataset (named as ReCRVD) with 120 groups of noisy-clean videos, whose ISO values ranging from 1600 to 25600. Secondly, while non-local temporal-spatial attention is beneficial for denoising, it often leads to heavy computation costs. We propose an efficient raw video denoising transformer network (RViDeformer) that explores both short and long-distance correlations. Specifically, we propose multi-branch spatial and temporal attention modules, which explore the patch correlations from local window, local low-resolution window, global downsampled window, and neighbor-involved window, and then they are fused together. We employ reparameterization to reduce computation costs. Our network is trained in both supervised and unsupervised manners, achieving the best performance compared with state-of-the-art methods. Additionally, the model trained with our proposed dataset (ReCRVD) outperforms the model trained with previous benchmark dataset (CRVD) when evaluated on the real-world outdoor noisy videos. Our code and dataset are available at https://github.com/cao-cong/RViDeformer.
2305.03944	Deyi Ji	Deyi Ji, Haoran Wang, Mingyuan Tao, Jianqiang Huang, Xian-Sheng Hua, Hongtao Lu	Structural and Statistical Texture Knowledge Distillation for Semantic Segmentation	Accepted to CVPR 2022. Extended TPAMI 2025 Journal Version: arXiv:2503.08043	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Existing knowledge distillation works for semantic segmentation mainly focus on transferring high-level contextual knowledge from teacher to student. However, low-level texture knowledge is also of vital importance for characterizing the local structural pattern and global statistical property, such as boundary, smoothness, regularity and color contrast, which may not be well addressed by high-level deep features. In this paper, we are intended to take full advantage of both structural and statistical texture knowledge and propose a novel Structural and Statistical Texture Knowledge Distillation (SSTKD) framework for semantic segmentation. Specifically, for structural texture knowledge, we introduce a Contourlet Decomposition Module (CDM) that decomposes low-level features with iterative Laplacian pyramid and directional filter bank to mine the structural texture knowledge. For statistical knowledge, we propose a Denoised Texture Intensity Equalization Module (DTIEM) to adaptively extract and enhance statistical texture knowledge through heuristics iterative quantization and denoised operation. Finally, each knowledge learning is supervised by an individual loss function, forcing the student network to mimic the teacher better from a broader perspective. Experiments show that the proposed method achieves state-of-the-art performance on Cityscapes, Pascal VOC 2012 and ADE20K datasets.	[ { "version": "v1", "created": "Sat, 6 May 2023 06:01:11 GMT" }, { "version": "v2", "created": "Thu, 6 Jul 2023 02:43:50 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 11:07:28 GMT" } ]	2025-03-18T00:00:00	[ [ "Ji", "Deyi", "" ], [ "Wang", "Haoran", "" ], [ "Tao", "Mingyuan", "" ], [ "Huang", "Jianqiang", "" ], [ "Hua", "Xian-Sheng", "" ], [ "Lu", "Hongtao", "" ] ]	TITLE: Structural and Statistical Texture Knowledge Distillation for Semantic Segmentation ABSTRACT: Existing knowledge distillation works for semantic segmentation mainly focus on transferring high-level contextual knowledge from teacher to student. However, low-level texture knowledge is also of vital importance for characterizing the local structural pattern and global statistical property, such as boundary, smoothness, regularity and color contrast, which may not be well addressed by high-level deep features. In this paper, we are intended to take full advantage of both structural and statistical texture knowledge and propose a novel Structural and Statistical Texture Knowledge Distillation (SSTKD) framework for semantic segmentation. Specifically, for structural texture knowledge, we introduce a Contourlet Decomposition Module (CDM) that decomposes low-level features with iterative Laplacian pyramid and directional filter bank to mine the structural texture knowledge. For statistical knowledge, we propose a Denoised Texture Intensity Equalization Module (DTIEM) to adaptively extract and enhance statistical texture knowledge through heuristics iterative quantization and denoised operation. Finally, each knowledge learning is supervised by an individual loss function, forcing the student network to mimic the teacher better from a broader perspective. Experiments show that the proposed method achieves state-of-the-art performance on Cityscapes, Pascal VOC 2012 and ADE20K datasets.
2305.13584	Pan Li	Li Pan, Lv Peizhuo, Chen Kai, Zhang Shengzhi, Cai Yuling, Xiang Fan	A Model Stealing Attack Against Multi-Exit Networks	null	null	null	null	cs.CR cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Compared to traditional neural networks with a single output channel, a multi-exit network has multiple exits that allow for early outputs from the model's intermediate layers, thus significantly improving computational efficiency while maintaining similar main task accuracy. Existing model stealing attacks can only steal the model's utility while failing to capture its output strategy, i.e., a set of thresholds used to determine from which exit to output. This leads to a significant decrease in computational efficiency for the extracted model, thereby losing the advantage of multi-exit networks. In this paper, we propose the first model stealing attack against multi-exit networks to extract both the model utility and the output strategy. We employ Kernel Density Estimation to analyze the target model's output strategy and use performance loss and strategy loss to guide the training of the extracted model. Furthermore, we design a novel output strategy search algorithm to maximize the consistency between the victim model and the extracted model's output behaviors. In experiments across multiple multi-exit networks and benchmark datasets, our method always achieves accuracy and efficiency closest to the victim models.	[ { "version": "v1", "created": "Tue, 23 May 2023 01:24:39 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 00:56:01 GMT" } ]	2025-03-18T00:00:00	[ [ "Pan", "Li", "" ], [ "Peizhuo", "Lv", "" ], [ "Kai", "Chen", "" ], [ "Shengzhi", "Zhang", "" ], [ "Yuling", "Cai", "" ], [ "Fan", "Xiang", "" ] ]	TITLE: A Model Stealing Attack Against Multi-Exit Networks ABSTRACT: Compared to traditional neural networks with a single output channel, a multi-exit network has multiple exits that allow for early outputs from the model's intermediate layers, thus significantly improving computational efficiency while maintaining similar main task accuracy. Existing model stealing attacks can only steal the model's utility while failing to capture its output strategy, i.e., a set of thresholds used to determine from which exit to output. This leads to a significant decrease in computational efficiency for the extracted model, thereby losing the advantage of multi-exit networks. In this paper, we propose the first model stealing attack against multi-exit networks to extract both the model utility and the output strategy. We employ Kernel Density Estimation to analyze the target model's output strategy and use performance loss and strategy loss to guide the training of the extracted model. Furthermore, we design a novel output strategy search algorithm to maximize the consistency between the victim model and the extracted model's output behaviors. In experiments across multiple multi-exit networks and benchmark datasets, our method always achieves accuracy and efficiency closest to the victim models.
2305.17473	Farhad Mortezapour Shiri	Farhad Mortezapour Shiri, Thinagaran Perumal, Norwati Mustapha, Raihani Mohamed	A Comprehensive Overview and Comparative Analysis on Deep Learning Models: CNN, RNN, LSTM, GRU	62 pages, 37 figures	Journal on Artificial Intelligence 2024 Vol. 6 Issue 1 Pages 301-360	10.32604/jai.2024.054314	null	cs.LG cs.AI	http://creativecommons.org/licenses/by-nc-nd/4.0/	Deep learning (DL) has emerged as a powerful subset of machine learning (ML) and artificial intelligence (AI), outperforming traditional ML methods, especially in handling unstructured and large datasets. Its impact spans across various domains, including speech recognition, healthcare, autonomous vehicles, cybersecurity, predictive analytics, and more. However, the complexity and dynamic nature of real-world problems present challenges in designing effective deep learning models. Consequently, several deep learning models have been developed to address different problems and applications. In this article, we conduct a comprehensive survey of various deep learning models, including Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Temporal Convolutional Networks (TCN), Transformer, Kolmogorov-Arnold networks (KAN), Generative Models, Deep Reinforcement Learning (DRL), and Deep Transfer Learning. We examine the structure, applications, benefits, and limitations of each model. Furthermore, we perform an analysis using three publicly available datasets: IMDB, ARAS, and Fruit-360. We compared the performance of six renowned deep learning models: CNN, RNN, Long Short-Term Memory (LSTM), Bidirectional LSTM, Gated Recurrent Unit (GRU), and Bidirectional GRU alongside two newer models, TCN and Transformer, using the IMDB and ARAS datasets. Additionally, we evaluated the performance of eight CNN-based models, including VGG (Visual Geometry Group), Inception, ResNet (Residual Network), InceptionResNet, Xception (Extreme Inception), MobileNet, DenseNet (Dense Convolutional Network), and NASNet (Neural Architecture Search Network), for image classification tasks using the Fruit-360 dataset.	[ { "version": "v1", "created": "Sat, 27 May 2023 13:23:21 GMT" }, { "version": "v2", "created": "Thu, 1 Jun 2023 16:53:28 GMT" }, { "version": "v3", "created": "Thu, 24 Oct 2024 17:41:58 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 10:18:52 GMT" } ]	2025-03-18T00:00:00	[ [ "Shiri", "Farhad Mortezapour", "" ], [ "Perumal", "Thinagaran", "" ], [ "Mustapha", "Norwati", "" ], [ "Mohamed", "Raihani", "" ] ]	TITLE: A Comprehensive Overview and Comparative Analysis on Deep Learning Models: CNN, RNN, LSTM, GRU ABSTRACT: Deep learning (DL) has emerged as a powerful subset of machine learning (ML) and artificial intelligence (AI), outperforming traditional ML methods, especially in handling unstructured and large datasets. Its impact spans across various domains, including speech recognition, healthcare, autonomous vehicles, cybersecurity, predictive analytics, and more. However, the complexity and dynamic nature of real-world problems present challenges in designing effective deep learning models. Consequently, several deep learning models have been developed to address different problems and applications. In this article, we conduct a comprehensive survey of various deep learning models, including Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Temporal Convolutional Networks (TCN), Transformer, Kolmogorov-Arnold networks (KAN), Generative Models, Deep Reinforcement Learning (DRL), and Deep Transfer Learning. We examine the structure, applications, benefits, and limitations of each model. Furthermore, we perform an analysis using three publicly available datasets: IMDB, ARAS, and Fruit-360. We compared the performance of six renowned deep learning models: CNN, RNN, Long Short-Term Memory (LSTM), Bidirectional LSTM, Gated Recurrent Unit (GRU), and Bidirectional GRU alongside two newer models, TCN and Transformer, using the IMDB and ARAS datasets. Additionally, we evaluated the performance of eight CNN-based models, including VGG (Visual Geometry Group), Inception, ResNet (Residual Network), InceptionResNet, Xception (Extreme Inception), MobileNet, DenseNet (Dense Convolutional Network), and NASNet (Neural Architecture Search Network), for image classification tasks using the Fruit-360 dataset.
2306.07207	Ziwang Zhao	Ruipu Luo, Ziwang Zhao, Min Yang, Zheming Yang, Minghui Qiu, Tao Wang, Zhongyu Wei, Yanhao Wang, Cen Chen	Valley: Video Assistant with Large Language model Enhanced abilitY	null	null	null	null	cs.CV cs.AI cs.CL	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Large Language Models (LLMs), with remarkable conversational capability, have emerged as AI assistants that can handle both visual and textual modalities. However, their effectiveness in joint video and language understanding has not been extensively explored. In the paper, we introduce Valley, a multi-modal foundation model that is designed to enable enhanced video comprehension and instruction-following capabilities. To this end, we construct two datasets, namely Valley-702k and Valley-instruct-73k, to cover a diverse range of video-text alignment and video-based instruction tasks, such as multi-shot captions, long video descriptions, action recognition, causal inference, etc. Then, we adopt ViT-L/14 as the vision encoder and explore three different temporal modeling modules to learn multifaceted features for enhanced video understanding. In addition, we implement a two-phase training approach for Valley: the first phase focuses solely on training the projection module to facilitate the LLM's capacity to understand visual input, and the second phase jointly trains the projection module and the LLM to improve their instruction following ability. Extensive experiments demonstrate that Valley has the potential to serve as an effective video assistant, simplifying complex video-understanding scenarios. Our code and data are published anonymously at https://github.com/valley-vl/Valley.	[ { "version": "v1", "created": "Mon, 12 Jun 2023 16:11:10 GMT" }, { "version": "v2", "created": "Sun, 8 Oct 2023 09:49:53 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 13:51:51 GMT" } ]	2025-03-18T00:00:00	[ [ "Luo", "Ruipu", "" ], [ "Zhao", "Ziwang", "" ], [ "Yang", "Min", "" ], [ "Yang", "Zheming", "" ], [ "Qiu", "Minghui", "" ], [ "Wang", "Tao", "" ], [ "Wei", "Zhongyu", "" ], [ "Wang", "Yanhao", "" ], [ "Chen", "Cen", "" ] ]	TITLE: Valley: Video Assistant with Large Language model Enhanced abilitY ABSTRACT: Large Language Models (LLMs), with remarkable conversational capability, have emerged as AI assistants that can handle both visual and textual modalities. However, their effectiveness in joint video and language understanding has not been extensively explored. In the paper, we introduce Valley, a multi-modal foundation model that is designed to enable enhanced video comprehension and instruction-following capabilities. To this end, we construct two datasets, namely Valley-702k and Valley-instruct-73k, to cover a diverse range of video-text alignment and video-based instruction tasks, such as multi-shot captions, long video descriptions, action recognition, causal inference, etc. Then, we adopt ViT-L/14 as the vision encoder and explore three different temporal modeling modules to learn multifaceted features for enhanced video understanding. In addition, we implement a two-phase training approach for Valley: the first phase focuses solely on training the projection module to facilitate the LLM's capacity to understand visual input, and the second phase jointly trains the projection module and the LLM to improve their instruction following ability. Extensive experiments demonstrate that Valley has the potential to serve as an effective video assistant, simplifying complex video-understanding scenarios. Our code and data are published anonymously at https://github.com/valley-vl/Valley.
2306.14858	Dominik Peters	Nikhil Chandak, Shashwat Goel, Dominik Peters	Proportional Aggregation of Preferences for Sequential Decision Making	Updated version with improved exposition. Axioms were renamed to better fit the literature	null	null	null	cs.GT cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We study the problem of fair sequential decision making given voter preferences. In each round, a decision rule must choose a decision from a set of alternatives where each voter reports which of these alternatives they approve. Instead of going with the most popular choice in each round, we aim for proportional representation across rounds, using axioms inspired by the multi-winner voting literature. The axioms require that every group of $\alpha\%$ of the voters that agrees in every round (i.e., approves a common alternative), must approve at least $\alpha\%$ of the decisions. A stronger version of the axioms requires that every group of $\alpha\%$ of the voters that agrees in a $\beta$ fraction of rounds must approve $\beta\cdot\alpha\%$ of the decisions. We show that three attractive voting rules satisfy axioms of this style. One of them (Sequential Phragm\'en) makes its decisions online, and the other two satisfy strengthened versions of the axioms but make decisions semi-online (Method of Equal Shares) or fully offline (Proportional Approval Voting). We present empirical results for these rules based on synthetic data and U.S. political elections. We also run experiments using the moral machine dataset about ethical dilemmas: We train preference models on user responses from different countries and let the models cast votes. We find that aggregating these votes using our rules leads to a more equal utility distribution across demographics than making decisions using a single global preference model.	[ { "version": "v1", "created": "Mon, 26 Jun 2023 17:10:10 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 14:25:48 GMT" } ]	2025-03-18T00:00:00	[ [ "Chandak", "Nikhil", "" ], [ "Goel", "Shashwat", "" ], [ "Peters", "Dominik", "" ] ]	TITLE: Proportional Aggregation of Preferences for Sequential Decision Making ABSTRACT: We study the problem of fair sequential decision making given voter preferences. In each round, a decision rule must choose a decision from a set of alternatives where each voter reports which of these alternatives they approve. Instead of going with the most popular choice in each round, we aim for proportional representation across rounds, using axioms inspired by the multi-winner voting literature. The axioms require that every group of $\alpha\%$ of the voters that agrees in every round (i.e., approves a common alternative), must approve at least $\alpha\%$ of the decisions. A stronger version of the axioms requires that every group of $\alpha\%$ of the voters that agrees in a $\beta$ fraction of rounds must approve $\beta\cdot\alpha\%$ of the decisions. We show that three attractive voting rules satisfy axioms of this style. One of them (Sequential Phragm\'en) makes its decisions online, and the other two satisfy strengthened versions of the axioms but make decisions semi-online (Method of Equal Shares) or fully offline (Proportional Approval Voting). We present empirical results for these rules based on synthetic data and U.S. political elections. We also run experiments using the moral machine dataset about ethical dilemmas: We train preference models on user responses from different countries and let the models cast votes. We find that aggregating these votes using our rules leads to a more equal utility distribution across demographics than making decisions using a single global preference model.
2307.03601	Shilong Zhang	Shilong Zhang, Peize Sun, Shoufa Chen, Min Xiao, Wenqi Shao, Wenwei Zhang, Yu Liu, Kai Chen, Ping Luo	GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interest	Code has been released at https://github.com/jshilong/GPT4RoI	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Visual instruction tuning large language model(LLM) on image-text pairs has achieved general-purpose vision-language abilities. However, the lack of region-text pairs limits their advancements to fine-grained multimodal understanding. In this paper, we propose spatial instruction tuning, which introduces the reference to the region-of-interest(RoI) in the instruction. Before sending to LLM, the reference is replaced by RoI features and interleaved with language embeddings as a sequence. Our model GPT4RoI, trained on 7 region-text pair datasets, brings an unprecedented interactive and conversational experience compared to previous image-level models. (1) Interaction beyond language: Users can interact with our model by both language and drawing bounding boxes to flexibly adjust the referring granularity. (2) Versatile multimodal abilities: A variety of attribute information within each RoI can be mined by GPT4RoI, e.g., color, shape, material, action, etc. Furthermore, it can reason about multiple RoIs based on common sense. On the Visual Commonsense Reasoning(VCR) dataset, GPT4RoI achieves a remarkable accuracy of 81.6%, surpassing all existing models by a significant margin (the second place is 75.6%) and almost reaching human-level performance of 85.0%. The code and model can be found at https://github.com/jshilong/GPT4RoI.	[ { "version": "v1", "created": "Fri, 7 Jul 2023 13:43:44 GMT" }, { "version": "v2", "created": "Fri, 13 Oct 2023 03:25:34 GMT" }, { "version": "v3", "created": "Sat, 1 Jun 2024 08:50:14 GMT" }, { "version": "v4", "created": "Sun, 16 Mar 2025 02:50:51 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Shilong", "" ], [ "Sun", "Peize", "" ], [ "Chen", "Shoufa", "" ], [ "Xiao", "Min", "" ], [ "Shao", "Wenqi", "" ], [ "Zhang", "Wenwei", "" ], [ "Liu", "Yu", "" ], [ "Chen", "Kai", "" ], [ "Luo", "Ping", "" ] ]	TITLE: GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interest ABSTRACT: Visual instruction tuning large language model(LLM) on image-text pairs has achieved general-purpose vision-language abilities. However, the lack of region-text pairs limits their advancements to fine-grained multimodal understanding. In this paper, we propose spatial instruction tuning, which introduces the reference to the region-of-interest(RoI) in the instruction. Before sending to LLM, the reference is replaced by RoI features and interleaved with language embeddings as a sequence. Our model GPT4RoI, trained on 7 region-text pair datasets, brings an unprecedented interactive and conversational experience compared to previous image-level models. (1) Interaction beyond language: Users can interact with our model by both language and drawing bounding boxes to flexibly adjust the referring granularity. (2) Versatile multimodal abilities: A variety of attribute information within each RoI can be mined by GPT4RoI, e.g., color, shape, material, action, etc. Furthermore, it can reason about multiple RoIs based on common sense. On the Visual Commonsense Reasoning(VCR) dataset, GPT4RoI achieves a remarkable accuracy of 81.6%, surpassing all existing models by a significant margin (the second place is 75.6%) and almost reaching human-level performance of 85.0%. The code and model can be found at https://github.com/jshilong/GPT4RoI.
2307.03972	Fanyi Qu	Fanyi Qu, Chenming Tang and Yunfang Wu	Evaluating the Capability of Large-scale Language Models on Chinese Grammatical Error Correction Task	null	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	Large-scale language models (LLMs) has shown remarkable capability in various of Natural Language Processing (NLP) tasks and attracted lots of attention recently. However, some studies indicated that large language models fail to achieve promising result beyond the state-of-the-art models in English grammatical error correction (GEC) tasks. In this report, we aim to explore the how large language models perform on Chinese grammatical error correction tasks and provide guidance for future work. We conduct experiments with 3 different LLMs of different model scale on 4 Chinese GEC dataset. Our experimental results indicate that the performances of LLMs on automatic evaluation metrics falls short of the previous sota models because of the problem of over-correction. Furthermore, we also discover notable variations in the performance of LLMs when evaluated on different data distributions. Our findings demonstrates that further investigation is required for the application of LLMs on Chinese GEC task.	[ { "version": "v1", "created": "Sat, 8 Jul 2023 13:10:59 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 11:21:32 GMT" } ]	2025-03-18T00:00:00	[ [ "Qu", "Fanyi", "" ], [ "Tang", "Chenming", "" ], [ "Wu", "Yunfang", "" ] ]	TITLE: Evaluating the Capability of Large-scale Language Models on Chinese Grammatical Error Correction Task ABSTRACT: Large-scale language models (LLMs) has shown remarkable capability in various of Natural Language Processing (NLP) tasks and attracted lots of attention recently. However, some studies indicated that large language models fail to achieve promising result beyond the state-of-the-art models in English grammatical error correction (GEC) tasks. In this report, we aim to explore the how large language models perform on Chinese grammatical error correction tasks and provide guidance for future work. We conduct experiments with 3 different LLMs of different model scale on 4 Chinese GEC dataset. Our experimental results indicate that the performances of LLMs on automatic evaluation metrics falls short of the previous sota models because of the problem of over-correction. Furthermore, we also discover notable variations in the performance of LLMs when evaluated on different data distributions. Our findings demonstrates that further investigation is required for the application of LLMs on Chinese GEC task.
2307.08789	Ranjan Sapkota	Ranjan Sapkota, Manoj Karkee	Generative AI in Agriculture: Creating Image Datasets Using DALL.E's Advanced Large Language Model Capabilities	9 Figures, 1 table, 19 pages	null	null	null	cs.CV	http://creativecommons.org/licenses/by/4.0/	This research investigated the role of artificial intelligence (AI), specifically the DALL.E model by OpenAI, in advancing data generation and visualization techniques in agriculture. DALL.E, an advanced AI image generator, works alongside ChatGPT's language processing to transform text descriptions and image clues into realistic visual representations of the content. The study used both approaches of image generation: text-to-image and image-to-image (variation). Six types of datasets depicting fruit crop environment were generated. These AI-generated images were then compared against ground truth images captured by sensors in real agricultural fields. The comparison was based on Peak Signal-to-Noise Ratio (PSNR) and Feature Similarity Index (FSIM) metrics. The image-to-image generation exhibited a 5.78% increase in average PSNR over text-to-image methods, signifying superior image clarity and quality. However, this method also resulted in a 10.23% decrease in average FSIM, indicating a diminished structural and textural similarity to the original images. Similar to these measures, human evaluation also showed that images generated using image-to-image-based method were more realistic compared to those generated with text-to-image approach. The results highlighted DALL.E's potential in generating realistic agricultural image datasets and thus accelerating the development and adoption of imaging-based precision agricultural solutions. In future, DALL.E along with other alternative LLM based image generation models such as MidJourney, Stable Diffusion, Craiyon, Imagen, Parti, DreamStudio, Make-A-Scene, DeepDream, and VQ-GAN + CLIP could demonstrate further significant potential for enhancing image clarity, quality, and realism in depicting agricultural environments, which could revolutionize precision farming practices.	[ { "version": "v1", "created": "Mon, 17 Jul 2023 19:17:10 GMT" }, { "version": "v2", "created": "Sun, 10 Mar 2024 17:47:38 GMT" }, { "version": "v3", "created": "Sat, 16 Mar 2024 18:38:18 GMT" }, { "version": "v4", "created": "Tue, 27 Aug 2024 16:43:17 GMT" }, { "version": "v5", "created": "Sat, 15 Mar 2025 17:13:08 GMT" } ]	2025-03-18T00:00:00	[ [ "Sapkota", "Ranjan", "" ], [ "Karkee", "Manoj", "" ] ]	TITLE: Generative AI in Agriculture: Creating Image Datasets Using DALL.E's Advanced Large Language Model Capabilities ABSTRACT: This research investigated the role of artificial intelligence (AI), specifically the DALL.E model by OpenAI, in advancing data generation and visualization techniques in agriculture. DALL.E, an advanced AI image generator, works alongside ChatGPT's language processing to transform text descriptions and image clues into realistic visual representations of the content. The study used both approaches of image generation: text-to-image and image-to-image (variation). Six types of datasets depicting fruit crop environment were generated. These AI-generated images were then compared against ground truth images captured by sensors in real agricultural fields. The comparison was based on Peak Signal-to-Noise Ratio (PSNR) and Feature Similarity Index (FSIM) metrics. The image-to-image generation exhibited a 5.78% increase in average PSNR over text-to-image methods, signifying superior image clarity and quality. However, this method also resulted in a 10.23% decrease in average FSIM, indicating a diminished structural and textural similarity to the original images. Similar to these measures, human evaluation also showed that images generated using image-to-image-based method were more realistic compared to those generated with text-to-image approach. The results highlighted DALL.E's potential in generating realistic agricultural image datasets and thus accelerating the development and adoption of imaging-based precision agricultural solutions. In future, DALL.E along with other alternative LLM based image generation models such as MidJourney, Stable Diffusion, Craiyon, Imagen, Parti, DreamStudio, Make-A-Scene, DeepDream, and VQ-GAN + CLIP could demonstrate further significant potential for enhancing image clarity, quality, and realism in depicting agricultural environments, which could revolutionize precision farming practices.
2308.06743	Baolin Liu	Baolin Liu and Zongyuan Yang and Pengfei Wang and Junjie Zhou and Ziqi Liu and Ziyi Song and Yan Liu and Yongping Xiong	TextDiff: Mask-Guided Residual Diffusion Models for Scene Text Image Super-Resolution	Correct and update some data	Pattern Recognition (2025)	10.1016/j.patcog.2025.111513	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The goal of scene text image super-resolution is to reconstruct high-resolution text-line images from unrecognizable low-resolution inputs. The existing methods relying on the optimization of pixel-level loss tend to yield text edges that exhibit a notable degree of blurring, thereby exerting a substantial impact on both the readability and recognizability of the text. To address these issues, we propose TextDiff, the first diffusion-based framework tailored for scene text image super-resolution. It contains two modules: the Text Enhancement Module (TEM) and the Mask-Guided Residual Diffusion Module (MRD). The TEM generates an initial deblurred text image and a mask that encodes the spatial location of the text. The MRD is responsible for effectively sharpening the text edge by modeling the residuals between the ground-truth images and the initial deblurred images. Extensive experiments demonstrate that our TextDiff achieves state-of-the-art (SOTA) performance on public benchmark datasets and can improve the readability of scene text images. Moreover, our proposed MRD module is plug-and-play that effectively sharpens the text edges produced by SOTA methods. This enhancement not only improves the readability and recognizability of the results generated by SOTA methods but also does not require any additional joint training. Available Codes:https://github.com/Lenubolim/TextDiff.	[ { "version": "v1", "created": "Sun, 13 Aug 2023 11:02:16 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 08:22:00 GMT" } ]	2025-03-18T00:00:00	[ [ "Liu", "Baolin", "" ], [ "Yang", "Zongyuan", "" ], [ "Wang", "Pengfei", "" ], [ "Zhou", "Junjie", "" ], [ "Liu", "Ziqi", "" ], [ "Song", "Ziyi", "" ], [ "Liu", "Yan", "" ], [ "Xiong", "Yongping", "" ] ]	TITLE: TextDiff: Mask-Guided Residual Diffusion Models for Scene Text Image Super-Resolution ABSTRACT: The goal of scene text image super-resolution is to reconstruct high-resolution text-line images from unrecognizable low-resolution inputs. The existing methods relying on the optimization of pixel-level loss tend to yield text edges that exhibit a notable degree of blurring, thereby exerting a substantial impact on both the readability and recognizability of the text. To address these issues, we propose TextDiff, the first diffusion-based framework tailored for scene text image super-resolution. It contains two modules: the Text Enhancement Module (TEM) and the Mask-Guided Residual Diffusion Module (MRD). The TEM generates an initial deblurred text image and a mask that encodes the spatial location of the text. The MRD is responsible for effectively sharpening the text edge by modeling the residuals between the ground-truth images and the initial deblurred images. Extensive experiments demonstrate that our TextDiff achieves state-of-the-art (SOTA) performance on public benchmark datasets and can improve the readability of scene text images. Moreover, our proposed MRD module is plug-and-play that effectively sharpens the text edges produced by SOTA methods. This enhancement not only improves the readability and recognizability of the results generated by SOTA methods but also does not require any additional joint training. Available Codes:https://github.com/Lenubolim/TextDiff.
2308.15568	Singh Akansha	Singh Akansha	Over-Squashing in Graph Neural Networks: A Comprehensive survey	18 pages	null	null	null	cs.AI	http://creativecommons.org/licenses/by/4.0/	Graph Neural Networks (GNNs) revolutionize machine learning for graph-structured data, effectively capturing complex relationships. They disseminate information through interconnected nodes, but long-range interactions face challenges known as "over-squashing". This survey delves into the challenge of over-squashing in Graph Neural Networks (GNNs), where long-range information dissemination is hindered, impacting tasks reliant on intricate long-distance interactions. It comprehensively explores the causes, consequences, and mitigation strategies for over-squashing. Various methodologies are reviewed, including graph rewiring, novel normalization, spectral analysis, and curvature-based strategies, with a focus on their trade-offs and effectiveness. The survey also discusses the interplay between over-squashing and other GNN limitations, such as over-smoothing, and provides a taxonomy of models designed to address these issues in node and graph-level tasks. Benchmark datasets for performance evaluation are also detailed, making this survey a valuable resource for researchers and practitioners in the GNN field.	[ { "version": "v1", "created": "Tue, 29 Aug 2023 18:46:15 GMT" }, { "version": "v2", "created": "Mon, 4 Sep 2023 11:54:33 GMT" }, { "version": "v3", "created": "Sun, 17 Sep 2023 13:06:01 GMT" }, { "version": "v4", "created": "Sat, 21 Oct 2023 09:39:48 GMT" }, { "version": "v5", "created": "Tue, 28 Nov 2023 11:03:06 GMT" }, { "version": "v6", "created": "Mon, 29 Apr 2024 14:15:42 GMT" }, { "version": "v7", "created": "Fri, 14 Mar 2025 20:10:31 GMT" } ]	2025-03-18T00:00:00	[ [ "Akansha", "Singh", "" ] ]	TITLE: Over-Squashing in Graph Neural Networks: A Comprehensive survey ABSTRACT: Graph Neural Networks (GNNs) revolutionize machine learning for graph-structured data, effectively capturing complex relationships. They disseminate information through interconnected nodes, but long-range interactions face challenges known as "over-squashing". This survey delves into the challenge of over-squashing in Graph Neural Networks (GNNs), where long-range information dissemination is hindered, impacting tasks reliant on intricate long-distance interactions. It comprehensively explores the causes, consequences, and mitigation strategies for over-squashing. Various methodologies are reviewed, including graph rewiring, novel normalization, spectral analysis, and curvature-based strategies, with a focus on their trade-offs and effectiveness. The survey also discusses the interplay between over-squashing and other GNN limitations, such as over-smoothing, and provides a taxonomy of models designed to address these issues in node and graph-level tasks. Benchmark datasets for performance evaluation are also detailed, making this survey a valuable resource for researchers and practitioners in the GNN field.
2310.07049	Yiming Chen	Yiming Chen, Chi Chen, Inhui Hwang, Michael J. Davis, Wanli Yang, Chengjun Sun, Gi-Hyeok Lee, Dylan McReynolds, Daniel Allen, Juan Marulanda Arias, Shyue Ping Ong and Maria K.Y. Chan	Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization	null	Chemistry of Materials 36.5 (2024): 2304-2313	10.1021/acs.chemmater.3c02584	null	physics.comp-ph	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	X-ray absorption spectroscopy (XAS) is a commonly-employed technique for characterizing functional materials. In particular, x-ray absorption near edge spectra (XANES) encodes local coordination and electronic information and machine learning approaches to extract this information is of significant interest. To date, most ML approaches for XANES have primarily focused on using the raw spectral intensities as input, overlooking the potential benefits of incorporating spectral transformations and dimensionality reduction techniques into ML predictions. In this work, we focused on systematically comparing the impact of different featurization methods on the performance of ML models for XAS analysis. We evaluated the classification and regression capabilities of these models on computed datasets and validated their performance on previously unseen experimental datasets. Our analysis revealed an intriguing discovery: the cumulative distribution function (CDF) feature achieves both high prediction accuracy and exceptional transferability. This remarkably robust performance can be attributed to its tolerance to horizontal shifts in spectra, which is crucial when validating models using experimental data. While this work exclusively focuses on XANES analysis, we anticipate that the methodology presented here will hold promise as a versatile asset to the broader spectroscopy community.	[ { "version": "v1", "created": "Tue, 10 Oct 2023 22:23:36 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 23:05:41 GMT" } ]	2025-03-18T00:00:00	[ [ "Chen", "Yiming", "" ], [ "Chen", "Chi", "" ], [ "Hwang", "Inhui", "" ], [ "Davis", "Michael J.", "" ], [ "Yang", "Wanli", "" ], [ "Sun", "Chengjun", "" ], [ "Lee", "Gi-Hyeok", "" ], [ "McReynolds", "Dylan", "" ], [ "Allen", "Daniel", "" ], [ "Arias", "Juan Marulanda", "" ], [ "Ong", "Shyue Ping", "" ], [ "Chan", "Maria K. Y.", "" ] ]	TITLE: Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization ABSTRACT: X-ray absorption spectroscopy (XAS) is a commonly-employed technique for characterizing functional materials. In particular, x-ray absorption near edge spectra (XANES) encodes local coordination and electronic information and machine learning approaches to extract this information is of significant interest. To date, most ML approaches for XANES have primarily focused on using the raw spectral intensities as input, overlooking the potential benefits of incorporating spectral transformations and dimensionality reduction techniques into ML predictions. In this work, we focused on systematically comparing the impact of different featurization methods on the performance of ML models for XAS analysis. We evaluated the classification and regression capabilities of these models on computed datasets and validated their performance on previously unseen experimental datasets. Our analysis revealed an intriguing discovery: the cumulative distribution function (CDF) feature achieves both high prediction accuracy and exceptional transferability. This remarkably robust performance can be attributed to its tolerance to horizontal shifts in spectra, which is crucial when validating models using experimental data. While this work exclusively focuses on XANES analysis, we anticipate that the methodology presented here will hold promise as a versatile asset to the broader spectroscopy community.
2311.05589	Yida Yin	Yida Yin, Zhiqiu Xu, Zhiyuan Li, Trevor Darrell, Zhuang Liu	A Coefficient Makes SVRG Effective	Published in ICLR 2025	null	null	null	cs.LG math.OC stat.ML	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Stochastic Variance Reduced Gradient (SVRG), introduced by Johnson & Zhang (2013), is a theoretically compelling optimization method. However, as Defazio & Bottou (2019) highlight, its effectiveness in deep learning is yet to be proven. In this work, we demonstrate the potential of SVRG in optimizing real-world neural networks. Our empirical analysis finds that, for deeper neural networks, the strength of the variance reduction term in SVRG should be smaller and decrease as training progresses. Inspired by this, we introduce a multiplicative coefficient $\alpha$ to control the strength and adjust it through a linear decay schedule. We name our method $\alpha$-SVRG. Our results show $\alpha$-SVRG better optimizes models, consistently reducing training loss compared to the baseline and standard SVRG across various model architectures and multiple image classification datasets. We hope our findings encourage further exploration into variance reduction techniques in deep learning. Code is available at github.com/davidyyd/alpha-SVRG.	[ { "version": "v1", "created": "Thu, 9 Nov 2023 18:47:44 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 11:14:58 GMT" } ]	2025-03-18T00:00:00	[ [ "Yin", "Yida", "" ], [ "Xu", "Zhiqiu", "" ], [ "Li", "Zhiyuan", "" ], [ "Darrell", "Trevor", "" ], [ "Liu", "Zhuang", "" ] ]	TITLE: A Coefficient Makes SVRG Effective ABSTRACT: Stochastic Variance Reduced Gradient (SVRG), introduced by Johnson & Zhang (2013), is a theoretically compelling optimization method. However, as Defazio & Bottou (2019) highlight, its effectiveness in deep learning is yet to be proven. In this work, we demonstrate the potential of SVRG in optimizing real-world neural networks. Our empirical analysis finds that, for deeper neural networks, the strength of the variance reduction term in SVRG should be smaller and decrease as training progresses. Inspired by this, we introduce a multiplicative coefficient $\alpha$ to control the strength and adjust it through a linear decay schedule. We name our method $\alpha$-SVRG. Our results show $\alpha$-SVRG better optimizes models, consistently reducing training loss compared to the baseline and standard SVRG across various model architectures and multiple image classification datasets. We hope our findings encourage further exploration into variance reduction techniques in deep learning. Code is available at github.com/davidyyd/alpha-SVRG.
2312.10766	Xiaoyu Zhang	Xiaoyu Zhang, Cen Zhang, Tianlin Li, Yihao Huang, Xiaojun Jia, Ming Hu, Jie Zhang, Yang Liu, Shiqing Ma, Chao Shen	JailGuard: A Universal Detection Framework for LLM Prompt-based Attacks	40 pages, 12 figures	null	null	null	cs.CR	http://creativecommons.org/licenses/by/4.0/	The systems and software powered by Large Language Models (LLMs) and Multi-Modal LLMs (MLLMs) have played a critical role in numerous scenarios. However, current LLM systems are vulnerable to prompt-based attacks, with jailbreaking attacks enabling the LLM system to generate harmful content, while hijacking attacks manipulate the LLM system to perform attacker-desired tasks, underscoring the necessity for detection tools. Unfortunately, existing detecting approaches are usually tailored to specific attacks, resulting in poor generalization in detecting various attacks across different modalities. To address it, we propose JailGuard, a universal detection framework deployed on top of LLM systems for prompt-based attacks across text and image modalities. JailGuard operates on the principle that attacks are inherently less robust than benign ones. Specifically, JailGuard mutates untrusted inputs to generate variants and leverages the discrepancy of the variants' responses on the target model to distinguish attack samples from benign samples. We implement 18 mutators for text and image inputs and design a mutator combination policy to further improve detection generalization. The evaluation on the dataset containing 15 known attack types suggests that JailGuard achieves the best detection accuracy of 86.14%/82.90% on text and image inputs, outperforming state-of-the-art methods by 11.81%-25.73% and 12.20%-21.40%.	[ { "version": "v1", "created": "Sun, 17 Dec 2023 17:02:14 GMT" }, { "version": "v2", "created": "Sat, 23 Dec 2023 14:17:31 GMT" }, { "version": "v3", "created": "Tue, 18 Jun 2024 02:21:02 GMT" }, { "version": "v4", "created": "Sat, 15 Mar 2025 00:49:45 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Xiaoyu", "" ], [ "Zhang", "Cen", "" ], [ "Li", "Tianlin", "" ], [ "Huang", "Yihao", "" ], [ "Jia", "Xiaojun", "" ], [ "Hu", "Ming", "" ], [ "Zhang", "Jie", "" ], [ "Liu", "Yang", "" ], [ "Ma", "Shiqing", "" ], [ "Shen", "Chao", "" ] ]	TITLE: JailGuard: A Universal Detection Framework for LLM Prompt-based Attacks ABSTRACT: The systems and software powered by Large Language Models (LLMs) and Multi-Modal LLMs (MLLMs) have played a critical role in numerous scenarios. However, current LLM systems are vulnerable to prompt-based attacks, with jailbreaking attacks enabling the LLM system to generate harmful content, while hijacking attacks manipulate the LLM system to perform attacker-desired tasks, underscoring the necessity for detection tools. Unfortunately, existing detecting approaches are usually tailored to specific attacks, resulting in poor generalization in detecting various attacks across different modalities. To address it, we propose JailGuard, a universal detection framework deployed on top of LLM systems for prompt-based attacks across text and image modalities. JailGuard operates on the principle that attacks are inherently less robust than benign ones. Specifically, JailGuard mutates untrusted inputs to generate variants and leverages the discrepancy of the variants' responses on the target model to distinguish attack samples from benign samples. We implement 18 mutators for text and image inputs and design a mutator combination policy to further improve detection generalization. The evaluation on the dataset containing 15 known attack types suggests that JailGuard achieves the best detection accuracy of 86.14%/82.90% on text and image inputs, outperforming state-of-the-art methods by 11.81%-25.73% and 12.20%-21.40%.
2312.12634	Payam Jome Yazdian	Payam Jome Yazdian, Rachel Lagasse, Hamid Mohammadi, Eric Liu, Li Cheng, Angelica Lim	MotionScript: Natural Language Descriptions for Expressive 3D Human Motions	Project webpage: https://pjyazdian.github.io/MotionScript	null	null	null	cs.CV cs.AI cs.CL cs.RO	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	We introduce MotionScript, a novel framework for generating highly detailed, natural language descriptions of 3D human motions. Unlike existing motion datasets that rely on broad action labels or generic captions, MotionScript provides fine-grained, structured descriptions that capture the full complexity of human movement including expressive actions (e.g., emotions, stylistic walking) and interactions beyond standard motion capture datasets. MotionScript serves as both a descriptive tool and a training resource for text-to-motion models, enabling the synthesis of highly realistic and diverse human motions from text. By augmenting motion datasets with MotionScript captions, we demonstrate significant improvements in out-of-distribution motion generation, allowing large language models (LLMs) to generate motions that extend beyond existing data. Additionally, MotionScript opens new applications in animation, virtual human simulation, and robotics, providing an interpretable bridge between intuitive descriptions and motion synthesis. To the best of our knowledge, this is the first attempt to systematically translate 3D motion into structured natural language without requiring training data.	[ { "version": "v1", "created": "Tue, 19 Dec 2023 22:33:17 GMT" }, { "version": "v2", "created": "Sun, 29 Sep 2024 20:24:27 GMT" }, { "version": "v3", "created": "Wed, 12 Mar 2025 21:16:45 GMT" }, { "version": "v4", "created": "Sun, 16 Mar 2025 17:50:27 GMT" } ]	2025-03-18T00:00:00	[ [ "Yazdian", "Payam Jome", "" ], [ "Lagasse", "Rachel", "" ], [ "Mohammadi", "Hamid", "" ], [ "Liu", "Eric", "" ], [ "Cheng", "Li", "" ], [ "Lim", "Angelica", "" ] ]	TITLE: MotionScript: Natural Language Descriptions for Expressive 3D Human Motions ABSTRACT: We introduce MotionScript, a novel framework for generating highly detailed, natural language descriptions of 3D human motions. Unlike existing motion datasets that rely on broad action labels or generic captions, MotionScript provides fine-grained, structured descriptions that capture the full complexity of human movement including expressive actions (e.g., emotions, stylistic walking) and interactions beyond standard motion capture datasets. MotionScript serves as both a descriptive tool and a training resource for text-to-motion models, enabling the synthesis of highly realistic and diverse human motions from text. By augmenting motion datasets with MotionScript captions, we demonstrate significant improvements in out-of-distribution motion generation, allowing large language models (LLMs) to generate motions that extend beyond existing data. Additionally, MotionScript opens new applications in animation, virtual human simulation, and robotics, providing an interpretable bridge between intuitive descriptions and motion synthesis. To the best of our knowledge, this is the first attempt to systematically translate 3D motion into structured natural language without requiring training data.
2401.05787	Md Rizwan Parvez	Md Rizwan Parvez	Chain of Evidences and Evidence to Generate: Prompting for Context Grounded and Retrieval Augmented Reasoning	Accepted at NAACL KnowledgeNLP 2025	null	null	null	cs.CL	http://creativecommons.org/licenses/by/4.0/	While chain-of-thoughts (CoT) prompting has revolutionized how LLMs perform reasoning tasks, its current methods and variations (e.g, Self-consistency, ReACT, Reflexion, Tree-of-Thoughts (ToT), Cumulative Reasoning (CR) etc.,) suffer from limitations like limited context grounding, hallucination/inconsistent output generation, and iterative sluggishness. To overcome these challenges, we introduce a novel mono/dual-step zero-shot prompting framework built upon two unique strategies Chain of Evidences (CoE)} and Evidence to Generate (E2G). Instead of unverified reasoning claims, our innovative approaches leverage the power of "evidence for decision making" by first focusing exclusively on the thought sequences explicitly mentioned in the context which then serve as extracted evidence, guiding the LLM's output generation process with greater precision and efficiency. This simple yet potent approach unlocks the full potential of chain-of-thoughts prompting, facilitating faster, more reliable, and contextually aware reasoning in LLMs. Our framework consistently achieves remarkable results across various knowledge-intensive reasoning and generation tasks, surpassing baseline approaches with state-of-the-art LLMs. For instance, (i) on the LogiQA benchmark using GPT-4, CoE achieves a new state-of-the-art accuracy of 53.8%, surpassing CoT by 18%, ToT by 11%, and CR by 9%; (ii) CoE with PaLM-2 outperforms the variable-shot performance of Gemini Ultra by 0.9 F1 points, achieving an F1 score of 83.3 on DROP. We release our prompts and outputs on these benchmarks as a new instruction tuning dataset for future research at https://huggingface.co/datasets/kagnlp/Chain-of-Evidences/.	[ { "version": "v1", "created": "Thu, 11 Jan 2024 09:49:15 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 10:35:11 GMT" } ]	2025-03-18T00:00:00	[ [ "Parvez", "Md Rizwan", "" ] ]	TITLE: Chain of Evidences and Evidence to Generate: Prompting for Context Grounded and Retrieval Augmented Reasoning ABSTRACT: While chain-of-thoughts (CoT) prompting has revolutionized how LLMs perform reasoning tasks, its current methods and variations (e.g, Self-consistency, ReACT, Reflexion, Tree-of-Thoughts (ToT), Cumulative Reasoning (CR) etc.,) suffer from limitations like limited context grounding, hallucination/inconsistent output generation, and iterative sluggishness. To overcome these challenges, we introduce a novel mono/dual-step zero-shot prompting framework built upon two unique strategies Chain of Evidences (CoE)} and Evidence to Generate (E2G). Instead of unverified reasoning claims, our innovative approaches leverage the power of "evidence for decision making" by first focusing exclusively on the thought sequences explicitly mentioned in the context which then serve as extracted evidence, guiding the LLM's output generation process with greater precision and efficiency. This simple yet potent approach unlocks the full potential of chain-of-thoughts prompting, facilitating faster, more reliable, and contextually aware reasoning in LLMs. Our framework consistently achieves remarkable results across various knowledge-intensive reasoning and generation tasks, surpassing baseline approaches with state-of-the-art LLMs. For instance, (i) on the LogiQA benchmark using GPT-4, CoE achieves a new state-of-the-art accuracy of 53.8%, surpassing CoT by 18%, ToT by 11%, and CR by 9%; (ii) CoE with PaLM-2 outperforms the variable-shot performance of Gemini Ultra by 0.9 F1 points, achieving an F1 score of 83.3 on DROP. We release our prompts and outputs on these benchmarks as a new instruction tuning dataset for future research at https://huggingface.co/datasets/kagnlp/Chain-of-Evidences/.
2401.08957	Kun Wu	Kun Wu, Ning Liu, Zhen Zhao, Di Qiu, Jinming Li, Zhengping Che, Zhiyuan Xu, Jian Tang	Learning from Imperfect Demonstrations with Self-Supervision for Robotic Manipulation	8 pages, 4 figures	null	null	null	cs.RO cs.AI	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Improving data utilization, especially for imperfect data from task failures, is crucial for robotic manipulation due to the challenging, time-consuming, and expensive data collection process in the real world. Current imitation learning (IL) typically discards imperfect data, focusing solely on successful expert data. While reinforcement learning (RL) can learn from explorations and failures, the sim2real gap and its reliance on dense reward and online exploration make it difficult to apply effectively in real-world scenarios. In this work, we aim to conquer the challenge of leveraging imperfect data without the need for reward information to improve the model performance for robotic manipulation in an offline manner. Specifically, we introduce a Self-Supervised Data Filtering framework (SSDF) that combines expert and imperfect data to compute quality scores for failed trajectory segments. High-quality segments from the failed data are used to expand the training dataset. Then, the enhanced dataset can be used with any downstream policy learning method for robotic manipulation tasks. Extensive experiments on the ManiSkill2 benchmark built on the high-fidelity Sapien simulator and real-world robotic manipulation tasks using the Franka robot arm demonstrated that the SSDF can accurately expand the training dataset with high-quality imperfect data and improve the success rates for all robotic manipulation tasks.	[ { "version": "v1", "created": "Wed, 17 Jan 2024 04:15:56 GMT" }, { "version": "v2", "created": "Mon, 17 Feb 2025 06:41:03 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 06:17:11 GMT" } ]	2025-03-18T00:00:00	[ [ "Wu", "Kun", "" ], [ "Liu", "Ning", "" ], [ "Zhao", "Zhen", "" ], [ "Qiu", "Di", "" ], [ "Li", "Jinming", "" ], [ "Che", "Zhengping", "" ], [ "Xu", "Zhiyuan", "" ], [ "Tang", "Jian", "" ] ]	TITLE: Learning from Imperfect Demonstrations with Self-Supervision for Robotic Manipulation ABSTRACT: Improving data utilization, especially for imperfect data from task failures, is crucial for robotic manipulation due to the challenging, time-consuming, and expensive data collection process in the real world. Current imitation learning (IL) typically discards imperfect data, focusing solely on successful expert data. While reinforcement learning (RL) can learn from explorations and failures, the sim2real gap and its reliance on dense reward and online exploration make it difficult to apply effectively in real-world scenarios. In this work, we aim to conquer the challenge of leveraging imperfect data without the need for reward information to improve the model performance for robotic manipulation in an offline manner. Specifically, we introduce a Self-Supervised Data Filtering framework (SSDF) that combines expert and imperfect data to compute quality scores for failed trajectory segments. High-quality segments from the failed data are used to expand the training dataset. Then, the enhanced dataset can be used with any downstream policy learning method for robotic manipulation tasks. Extensive experiments on the ManiSkill2 benchmark built on the high-fidelity Sapien simulator and real-world robotic manipulation tasks using the Franka robot arm demonstrated that the SSDF can accurately expand the training dataset with high-quality imperfect data and improve the success rates for all robotic manipulation tasks.
2401.16515	Sean Lam	Sean Lam, Ahmed Khaled, Simon Bilodeau, Bicky A. Marquez, Paul R. Prucnal, Lukas Chrostowski, Bhavin J. Shastri, Sudip Shekhar	Neuromorphic Photonic Computing with an Electro-Optic Analog Memory	null	null	null	null	cs.ET cs.SY eess.SP eess.SY physics.optics	http://creativecommons.org/licenses/by/4.0/	Artificial intelligence (AI) has seen remarkable advancements across various domains, including natural language processing, computer vision, autonomous vehicles, and biology. However, the rapid expansion of AI technologies has escalated the demand for more powerful computing resources. As digital computing approaches fundamental limits, neuromorphic photonics emerges as a promising platform to complement existing digital systems. In neuromorphic photonic computing, photonic devices are controlled using analog signals. This necessitates the use of digital-to-analog converters (DAC) and analog-to-digital converters (ADC) for interfacing with these devices during inference and training. However, data movement between memory and these converters in conventional von Neumann computing architectures consumes energy. To address this, analog memory co-located with photonic computing devices is proposed. This approach aims to reduce the reliance on DACs and minimize data movement to enhance compute efficiency. This paper demonstrates a monolithically integrated neuromorphic photonic circuit with co-located capacitive analog memory and analyzes analog memory specifications for neuromorphic photonic computing using the MNIST dataset as a benchmark.	[ { "version": "v1", "created": "Mon, 29 Jan 2024 19:37:50 GMT" }, { "version": "v2", "created": "Tue, 10 Sep 2024 23:55:57 GMT" }, { "version": "v3", "created": "Sun, 16 Mar 2025 21:58:12 GMT" } ]	2025-03-18T00:00:00	[ [ "Lam", "Sean", "" ], [ "Khaled", "Ahmed", "" ], [ "Bilodeau", "Simon", "" ], [ "Marquez", "Bicky A.", "" ], [ "Prucnal", "Paul R.", "" ], [ "Chrostowski", "Lukas", "" ], [ "Shastri", "Bhavin J.", "" ], [ "Shekhar", "Sudip", "" ] ]	TITLE: Neuromorphic Photonic Computing with an Electro-Optic Analog Memory ABSTRACT: Artificial intelligence (AI) has seen remarkable advancements across various domains, including natural language processing, computer vision, autonomous vehicles, and biology. However, the rapid expansion of AI technologies has escalated the demand for more powerful computing resources. As digital computing approaches fundamental limits, neuromorphic photonics emerges as a promising platform to complement existing digital systems. In neuromorphic photonic computing, photonic devices are controlled using analog signals. This necessitates the use of digital-to-analog converters (DAC) and analog-to-digital converters (ADC) for interfacing with these devices during inference and training. However, data movement between memory and these converters in conventional von Neumann computing architectures consumes energy. To address this, analog memory co-located with photonic computing devices is proposed. This approach aims to reduce the reliance on DACs and minimize data movement to enhance compute efficiency. This paper demonstrates a monolithically integrated neuromorphic photonic circuit with co-located capacitive analog memory and analyzes analog memory specifications for neuromorphic photonic computing using the MNIST dataset as a benchmark.
2402.04398	Sujay Nagaraj	Sujay Nagaraj, Walter Gerych, Sana Tonekaboni, Anna Goldenberg, Berk Ustun, Thomas Hartvigsen	Learning under Temporal Label Noise	The Thirteenth International Conference on Learning Representations (ICLR 2025)	null	null	null	cs.LG cs.AI stat.ML	http://creativecommons.org/licenses/by/4.0/	Many time series classification tasks, where labels vary over time, are affected by label noise that also varies over time. Such noise can cause label quality to improve, worsen, or periodically change over time. We first propose and formalize temporal label noise, an unstudied problem for sequential classification of time series. In this setting, multiple labels are recorded over time while being corrupted by a time-dependent noise function. We first demonstrate the importance of modeling the temporal nature of the label noise function and how existing methods will consistently underperform. We then propose methods to train noise-tolerant classifiers by estimating the temporal label noise function directly from data. We show that our methods lead to state-of-the-art performance under diverse types of temporal label noise on real-world datasets	[ { "version": "v1", "created": "Tue, 6 Feb 2024 20:56:31 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 09:14:36 GMT" } ]	2025-03-18T00:00:00	[ [ "Nagaraj", "Sujay", "" ], [ "Gerych", "Walter", "" ], [ "Tonekaboni", "Sana", "" ], [ "Goldenberg", "Anna", "" ], [ "Ustun", "Berk", "" ], [ "Hartvigsen", "Thomas", "" ] ]	TITLE: Learning under Temporal Label Noise ABSTRACT: Many time series classification tasks, where labels vary over time, are affected by label noise that also varies over time. Such noise can cause label quality to improve, worsen, or periodically change over time. We first propose and formalize temporal label noise, an unstudied problem for sequential classification of time series. In this setting, multiple labels are recorded over time while being corrupted by a time-dependent noise function. We first demonstrate the importance of modeling the temporal nature of the label noise function and how existing methods will consistently underperform. We then propose methods to train noise-tolerant classifiers by estimating the temporal label noise function directly from data. We show that our methods lead to state-of-the-art performance under diverse types of temporal label noise on real-world datasets
2402.06423	Erkang Cheng	Yifeng Bai, Zhirong Chen, Pengpeng Liang, Bo Song, Erkang Cheng	CurveFormer++: 3D Lane Detection by Curve Propagation with Temporal Curve Queries and Attention	arXiv admin note: text overlap with arXiv:2209.07989	null	null	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	In autonomous driving, accurate 3D lane detection using monocular cameras is important for downstream tasks. Recent CNN and Transformer approaches usually apply a two-stage model design. The first stage transforms the image feature from a front image into a bird's-eye-view (BEV) representation. Subsequently, a sub-network processes the BEV feature to generate the 3D detection results. However, these approaches heavily rely on a challenging image feature transformation module from a perspective view to a BEV representation. In our work, we present CurveFormer++, a single-stage Transformer-based method that does not require the view transform module and directly infers 3D lane results from the perspective image features. Specifically, our approach models the 3D lane detection task as a curve propagation problem, where each lane is represented by a curve query with a dynamic and ordered anchor point set. By employing a Transformer decoder, the model can iteratively refine the 3D lane results. A curve cross-attention module is introduced to calculate similarities between image features and curve queries. To handle varying lane lengths, we employ context sampling and anchor point restriction techniques to compute more relevant image features. Furthermore, we apply a temporal fusion module that incorporates selected informative sparse curve queries and their corresponding anchor point sets to leverage historical information. In the experiments, we evaluate our approach on two publicly real-world datasets. The results demonstrate that our method provides outstanding performance compared with both CNN and Transformer based methods. We also conduct ablation studies to analyze the impact of each component.	[ { "version": "v1", "created": "Fri, 9 Feb 2024 14:13:40 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 14:40:20 GMT" } ]	2025-03-18T00:00:00	[ [ "Bai", "Yifeng", "" ], [ "Chen", "Zhirong", "" ], [ "Liang", "Pengpeng", "" ], [ "Song", "Bo", "" ], [ "Cheng", "Erkang", "" ] ]	TITLE: CurveFormer++: 3D Lane Detection by Curve Propagation with Temporal Curve Queries and Attention ABSTRACT: In autonomous driving, accurate 3D lane detection using monocular cameras is important for downstream tasks. Recent CNN and Transformer approaches usually apply a two-stage model design. The first stage transforms the image feature from a front image into a bird's-eye-view (BEV) representation. Subsequently, a sub-network processes the BEV feature to generate the 3D detection results. However, these approaches heavily rely on a challenging image feature transformation module from a perspective view to a BEV representation. In our work, we present CurveFormer++, a single-stage Transformer-based method that does not require the view transform module and directly infers 3D lane results from the perspective image features. Specifically, our approach models the 3D lane detection task as a curve propagation problem, where each lane is represented by a curve query with a dynamic and ordered anchor point set. By employing a Transformer decoder, the model can iteratively refine the 3D lane results. A curve cross-attention module is introduced to calculate similarities between image features and curve queries. To handle varying lane lengths, we employ context sampling and anchor point restriction techniques to compute more relevant image features. Furthermore, we apply a temporal fusion module that incorporates selected informative sparse curve queries and their corresponding anchor point sets to leverage historical information. In the experiments, we evaluate our approach on two publicly real-world datasets. The results demonstrate that our method provides outstanding performance compared with both CNN and Transformer based methods. We also conduct ablation studies to analyze the impact of each component.
2402.07927	Pranab Sahoo	Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha	A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications	12 pages, 2 figures	null	null	null	cs.AI cs.CL cs.HC	http://creativecommons.org/licenses/by/4.0/	Prompt engineering has emerged as an indispensable technique for extending the capabilities of large language models (LLMs) and vision-language models (VLMs). This approach leverages task-specific instructions, known as prompts, to enhance model efficacy without modifying the core model parameters. Rather than updating the model parameters, prompts allow seamless integration of pre-trained models into downstream tasks by eliciting desired model behaviors solely based on the given prompt. Prompts can be natural language instructions that provide context to guide the model or learned vector representations that activate relevant knowledge. This burgeoning field has enabled success across various applications, from question-answering to commonsense reasoning. However, there remains a lack of systematic organization and understanding of the diverse prompt engineering methods and techniques. This survey paper addresses the gap by providing a structured overview of recent advancements in prompt engineering, categorized by application area. For each prompting approach, we provide a summary detailing the prompting methodology, its applications, the models involved, and the datasets utilized. We also delve into the strengths and limitations of each approach and include a taxonomy diagram and table summarizing datasets, models, and critical points of each prompting technique. This systematic analysis enables a better understanding of this rapidly developing field and facilitates future research by illuminating open challenges and opportunities for prompt engineering.	[ { "version": "v1", "created": "Mon, 5 Feb 2024 19:49:13 GMT" }, { "version": "v2", "created": "Sun, 16 Mar 2025 06:23:34 GMT" } ]	2025-03-18T00:00:00	[ [ "Sahoo", "Pranab", "" ], [ "Singh", "Ayush Kumar", "" ], [ "Saha", "Sriparna", "" ], [ "Jain", "Vinija", "" ], [ "Mondal", "Samrat", "" ], [ "Chadha", "Aman", "" ] ]	TITLE: A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications ABSTRACT: Prompt engineering has emerged as an indispensable technique for extending the capabilities of large language models (LLMs) and vision-language models (VLMs). This approach leverages task-specific instructions, known as prompts, to enhance model efficacy without modifying the core model parameters. Rather than updating the model parameters, prompts allow seamless integration of pre-trained models into downstream tasks by eliciting desired model behaviors solely based on the given prompt. Prompts can be natural language instructions that provide context to guide the model or learned vector representations that activate relevant knowledge. This burgeoning field has enabled success across various applications, from question-answering to commonsense reasoning. However, there remains a lack of systematic organization and understanding of the diverse prompt engineering methods and techniques. This survey paper addresses the gap by providing a structured overview of recent advancements in prompt engineering, categorized by application area. For each prompting approach, we provide a summary detailing the prompting methodology, its applications, the models involved, and the datasets utilized. We also delve into the strengths and limitations of each approach and include a taxonomy diagram and table summarizing datasets, models, and critical points of each prompting technique. This systematic analysis enables a better understanding of this rapidly developing field and facilitates future research by illuminating open challenges and opportunities for prompt engineering.
2402.12265	Christophe Roux	Christophe Roux, Max Zimmer, Sebastian Pokutta	On the Byzantine-Resilience of Distillation-Based Federated Learning	null	null	null	null	cs.LG cs.AI cs.DC	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Federated Learning (FL) algorithms using Knowledge Distillation (KD) have received increasing attention due to their favorable properties with respect to privacy, non-i.i.d. data and communication cost. These methods depart from transmitting model parameters and instead communicate information about a learning task by sharing predictions on a public dataset. In this work, we study the performance of such approaches in the byzantine setting, where a subset of the clients act in an adversarial manner aiming to disrupt the learning process. We show that KD-based FL algorithms are remarkably resilient and analyze how byzantine clients can influence the learning process. Based on these insights, we introduce two new byzantine attacks and demonstrate their ability to break existing byzantine-resilient methods. Additionally, we propose a novel defence method which enhances the byzantine resilience of KD-based FL algorithms. Finally, we provide a general framework to obfuscate attacks, making them significantly harder to detect, thereby improving their effectiveness. Our findings serve as an important building block in the analysis of byzantine FL, contributing through the development of new attacks and new defence mechanisms, further advancing the robustness of KD-based FL algorithms.	[ { "version": "v1", "created": "Mon, 19 Feb 2024 16:26:40 GMT" }, { "version": "v2", "created": "Wed, 9 Oct 2024 12:38:26 GMT" }, { "version": "v3", "created": "Mon, 17 Mar 2025 14:08:19 GMT" } ]	2025-03-18T00:00:00	[ [ "Roux", "Christophe", "" ], [ "Zimmer", "Max", "" ], [ "Pokutta", "Sebastian", "" ] ]	TITLE: On the Byzantine-Resilience of Distillation-Based Federated Learning ABSTRACT: Federated Learning (FL) algorithms using Knowledge Distillation (KD) have received increasing attention due to their favorable properties with respect to privacy, non-i.i.d. data and communication cost. These methods depart from transmitting model parameters and instead communicate information about a learning task by sharing predictions on a public dataset. In this work, we study the performance of such approaches in the byzantine setting, where a subset of the clients act in an adversarial manner aiming to disrupt the learning process. We show that KD-based FL algorithms are remarkably resilient and analyze how byzantine clients can influence the learning process. Based on these insights, we introduce two new byzantine attacks and demonstrate their ability to break existing byzantine-resilient methods. Additionally, we propose a novel defence method which enhances the byzantine resilience of KD-based FL algorithms. Finally, we provide a general framework to obfuscate attacks, making them significantly harder to detect, thereby improving their effectiveness. Our findings serve as an important building block in the analysis of byzantine FL, contributing through the development of new attacks and new defence mechanisms, further advancing the robustness of KD-based FL algorithms.
2402.14598	Depin Liang	Jianming Lv, Chengjun Wang, Depin Liang, Qianli Ma, Wei Chen, Xueqi Cheng	EMN: Brain-inspired Elastic Memory Network for Quick Domain Adaptive Feature Mapping	15 pages,15 figures	null	null	null	cs.NE cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Utilizing unlabeled data in the target domain to perform continuous optimization is critical to enhance the generalization ability of neural networks. Most domain adaptation methods focus on time-consuming optimization of deep feature extractors, which limits the deployment on lightweight edge devices. Inspired by the memory mechanism and powerful generalization ability of biological neural networks in human brains, we propose a novel gradient-free Elastic Memory Network, namely EMN, to support quick fine-tuning of the mapping between features and prediction without heavy optimization of deep features. In particular, EMN adopts randomly connected neurons to memorize the association of features and labels, where the signals in the network are propagated as impulses, and the prediction is made by associating the memories stored on neurons based on their confidence. More importantly, EMN supports reinforced memorization of feature mapping based on unlabeled data to quickly adapt to a new domain. Experiments based on four cross-domain real-world datasets show that EMN can achieve up to 10% enhancement of performance while only needing less than 1% timing cost of traditional domain adaptation methods.	[ { "version": "v1", "created": "Sun, 4 Feb 2024 09:58:17 GMT" }, { "version": "v2", "created": "Mon, 17 Mar 2025 08:34:07 GMT" } ]	2025-03-18T00:00:00	[ [ "Lv", "Jianming", "" ], [ "Wang", "Chengjun", "" ], [ "Liang", "Depin", "" ], [ "Ma", "Qianli", "" ], [ "Chen", "Wei", "" ], [ "Cheng", "Xueqi", "" ] ]	TITLE: EMN: Brain-inspired Elastic Memory Network for Quick Domain Adaptive Feature Mapping ABSTRACT: Utilizing unlabeled data in the target domain to perform continuous optimization is critical to enhance the generalization ability of neural networks. Most domain adaptation methods focus on time-consuming optimization of deep feature extractors, which limits the deployment on lightweight edge devices. Inspired by the memory mechanism and powerful generalization ability of biological neural networks in human brains, we propose a novel gradient-free Elastic Memory Network, namely EMN, to support quick fine-tuning of the mapping between features and prediction without heavy optimization of deep features. In particular, EMN adopts randomly connected neurons to memorize the association of features and labels, where the signals in the network are propagated as impulses, and the prediction is made by associating the memories stored on neurons based on their confidence. More importantly, EMN supports reinforced memorization of feature mapping based on unlabeled data to quickly adapt to a new domain. Experiments based on four cross-domain real-world datasets show that EMN can achieve up to 10% enhancement of performance while only needing less than 1% timing cost of traditional domain adaptation methods.
2402.17555	Xinliang Zhang	Xinliang Zhang, Lei Zhu, Hangzhou He, Lujia Jin, Yanye Lu	Scribble Hides Class: Promoting Scribble-Based Weakly-Supervised Semantic Segmentation with Its Class Label	null	null	10.1609/aaai.v38i7.28563	null	cs.CV	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	Scribble-based weakly-supervised semantic segmentation using sparse scribble supervision is gaining traction as it reduces annotation costs when compared to fully annotated alternatives. Existing methods primarily generate pseudo-labels by diffusing labeled pixels to unlabeled ones with local cues for supervision. However, this diffusion process fails to exploit global semantics and class-specific cues, which are important for semantic segmentation. In this study, we propose a class-driven scribble promotion network, which utilizes both scribble annotations and pseudo-labels informed by image-level classes and global semantics for supervision. Directly adopting pseudo-labels might misguide the segmentation model, thus we design a localization rectification module to correct foreground representations in the feature space. To further combine the advantages of both supervisions, we also introduce a distance entropy loss for uncertainty reduction, which adapts per-pixel confidence weights according to the reliable region determined by the scribble and pseudo-label's boundary. Experiments on the ScribbleSup dataset with different qualities of scribble annotations outperform all the previous methods, demonstrating the superiority and robustness of our method.The code is available at https://github.com/Zxl19990529/Class-driven-Scribble-Promotion-Network.	[ { "version": "v1", "created": "Tue, 27 Feb 2024 14:51:56 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhang", "Xinliang", "" ], [ "Zhu", "Lei", "" ], [ "He", "Hangzhou", "" ], [ "Jin", "Lujia", "" ], [ "Lu", "Yanye", "" ] ]	TITLE: Scribble Hides Class: Promoting Scribble-Based Weakly-Supervised Semantic Segmentation with Its Class Label ABSTRACT: Scribble-based weakly-supervised semantic segmentation using sparse scribble supervision is gaining traction as it reduces annotation costs when compared to fully annotated alternatives. Existing methods primarily generate pseudo-labels by diffusing labeled pixels to unlabeled ones with local cues for supervision. However, this diffusion process fails to exploit global semantics and class-specific cues, which are important for semantic segmentation. In this study, we propose a class-driven scribble promotion network, which utilizes both scribble annotations and pseudo-labels informed by image-level classes and global semantics for supervision. Directly adopting pseudo-labels might misguide the segmentation model, thus we design a localization rectification module to correct foreground representations in the feature space. To further combine the advantages of both supervisions, we also introduce a distance entropy loss for uncertainty reduction, which adapts per-pixel confidence weights according to the reliable region determined by the scribble and pseudo-label's boundary. Experiments on the ScribbleSup dataset with different qualities of scribble annotations outperform all the previous methods, demonstrating the superiority and robustness of our method.The code is available at https://github.com/Zxl19990529/Class-driven-Scribble-Promotion-Network.
2403.02971	Xiaoyi Zhu	Xiaoyi Zhu, Yuxiang Tian, Lingxiao Huang, Zengfeng Huang	Space Complexity of Euclidean Clustering	Accepted by SoCG2024, TIT2025, in IEEE Transactions on Information Theory, 2025	null	10.1109/TIT.2025.3550192	null	cs.CG cs.DS	http://creativecommons.org/licenses/by/4.0/	The $(k, z)$-Clustering problem in Euclidean space $\mathbb{R}^d$ has been extensively studied. Given the scale of data involved, compression methods for the Euclidean $(k, z)$-Clustering problem, such as data compression and dimension reduction, have received significant attention in the literature. However, the space complexity of the clustering problem, specifically, the number of bits required to compress the cost function within a multiplicative error $\varepsilon$, remains unclear in existing literature. This paper initiates the study of space complexity for Euclidean $(k, z)$-Clustering and offers both upper and lower bounds. Our space bounds are nearly tight when $k$ is constant, indicating that storing a coreset, a well-known data compression approach, serves as the optimal compression scheme. Furthermore, our lower bound result for $(k, z)$-Clustering establishes a tight space bound of $\Theta( n d )$ for terminal embedding, where $n$ represents the dataset size. Our technical approach leverages new geometric insights for principal angles and discrepancy methods, which may hold independent interest.	[ { "version": "v1", "created": "Tue, 5 Mar 2024 13:49:32 GMT" }, { "version": "v2", "created": "Wed, 6 Mar 2024 02:05:36 GMT" }, { "version": "v3", "created": "Sat, 15 Mar 2025 01:58:29 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhu", "Xiaoyi", "" ], [ "Tian", "Yuxiang", "" ], [ "Huang", "Lingxiao", "" ], [ "Huang", "Zengfeng", "" ] ]	TITLE: Space Complexity of Euclidean Clustering ABSTRACT: The $(k, z)$-Clustering problem in Euclidean space $\mathbb{R}^d$ has been extensively studied. Given the scale of data involved, compression methods for the Euclidean $(k, z)$-Clustering problem, such as data compression and dimension reduction, have received significant attention in the literature. However, the space complexity of the clustering problem, specifically, the number of bits required to compress the cost function within a multiplicative error $\varepsilon$, remains unclear in existing literature. This paper initiates the study of space complexity for Euclidean $(k, z)$-Clustering and offers both upper and lower bounds. Our space bounds are nearly tight when $k$ is constant, indicating that storing a coreset, a well-known data compression approach, serves as the optimal compression scheme. Furthermore, our lower bound result for $(k, z)$-Clustering establishes a tight space bound of $\Theta( n d )$ for terminal embedding, where $n$ represents the dataset size. Our technical approach leverages new geometric insights for principal angles and discrepancy methods, which may hold independent interest.
2403.06869	Hao Chen	Hao Chen, Zihan Wang, Ran Tao, Hongxin Wei, Xing Xie, Masashi Sugiyama, Bhiksha Raj, Jindong Wang	Impact of Noisy Supervision in Foundation Model Learning	18 pages, 10 figures, 6 tables, preprint. arXiv admin note: substantial text overlap with arXiv:2309.17002	null	null	null	cs.LG cs.AI cs.CL cs.CV	http://creativecommons.org/licenses/by/4.0/	Foundation models are usually pre-trained on large-scale datasets and then adapted to downstream tasks through tuning. However, the large-scale pre-training datasets, often inaccessible or too expensive to handle, can contain label noise that may adversely affect the generalization of the model and pose unexpected risks. This paper stands out as the first work to comprehensively understand and analyze the nature of noise in pre-training datasets and then effectively mitigate its impacts on downstream tasks. Specifically, through extensive experiments of fully-supervised and image-text contrastive pre-training on synthetic noisy ImageNet-1K, YFCC15M, and CC12M datasets, we demonstrate that, while slight noise in pre-training can benefit in-domain (ID) performance, where the training and testing data share a similar distribution, it always deteriorates out-of-domain (OOD) performance, where training and testing distributions are significantly different. These observations are agnostic to scales of pre-training datasets, pre-training noise types, model architectures, pre-training objectives, downstream tuning methods, and downstream applications. We empirically ascertain that the reason behind this is that the pre-training noise shapes the feature space differently. We then propose a tuning method (NMTune) to affine the feature space to mitigate the malignant effect of noise and improve generalization, which is applicable in both parameter-efficient and black-box tuning manners. We additionally conduct extensive experiments on popular vision and language models, including APIs, which are supervised and self-supervised pre-trained on realistic noisy data for evaluation. Our analysis and results demonstrate the importance of this novel and fundamental research direction, which we term as Noisy Model Learning.	[ { "version": "v1", "created": "Mon, 11 Mar 2024 16:22:41 GMT" }, { "version": "v2", "created": "Fri, 14 Mar 2025 22:46:43 GMT" } ]	2025-03-18T00:00:00	[ [ "Chen", "Hao", "" ], [ "Wang", "Zihan", "" ], [ "Tao", "Ran", "" ], [ "Wei", "Hongxin", "" ], [ "Xie", "Xing", "" ], [ "Sugiyama", "Masashi", "" ], [ "Raj", "Bhiksha", "" ], [ "Wang", "Jindong", "" ] ]	TITLE: Impact of Noisy Supervision in Foundation Model Learning ABSTRACT: Foundation models are usually pre-trained on large-scale datasets and then adapted to downstream tasks through tuning. However, the large-scale pre-training datasets, often inaccessible or too expensive to handle, can contain label noise that may adversely affect the generalization of the model and pose unexpected risks. This paper stands out as the first work to comprehensively understand and analyze the nature of noise in pre-training datasets and then effectively mitigate its impacts on downstream tasks. Specifically, through extensive experiments of fully-supervised and image-text contrastive pre-training on synthetic noisy ImageNet-1K, YFCC15M, and CC12M datasets, we demonstrate that, while slight noise in pre-training can benefit in-domain (ID) performance, where the training and testing data share a similar distribution, it always deteriorates out-of-domain (OOD) performance, where training and testing distributions are significantly different. These observations are agnostic to scales of pre-training datasets, pre-training noise types, model architectures, pre-training objectives, downstream tuning methods, and downstream applications. We empirically ascertain that the reason behind this is that the pre-training noise shapes the feature space differently. We then propose a tuning method (NMTune) to affine the feature space to mitigate the malignant effect of noise and improve generalization, which is applicable in both parameter-efficient and black-box tuning manners. We additionally conduct extensive experiments on popular vision and language models, including APIs, which are supervised and self-supervised pre-trained on realistic noisy data for evaluation. Our analysis and results demonstrate the importance of this novel and fundamental research direction, which we term as Noisy Model Learning.
2403.07378	Xin Wang	Xin Wang, Yu Zheng, Zhongwei Wan, Mi Zhang	SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression	ICLR 2025; Code available at: https://github.com/AIoT-MLSys-Lab/SVD-LLM	null	null	null	cs.CL cs.LG	http://arxiv.org/licenses/nonexclusive-distrib/1.0/	The advancements in Large Language Models (LLMs) have been hindered by their substantial sizes, which necessitates LLM compression methods for practical deployment. Singular Value Decomposition (SVD) offers a promising solution for LLM compression. However, state-of-the-art SVD-based LLM compression methods have two key limitations: truncating smaller singular values may lead to higher compression loss, and the lack of update on the compressed weights after SVD truncation. In this work, we propose SVD-LLM, a SVD-based post-training LLM compression method that addresses the limitations of existing methods. SVD-LLM incorporates a truncation-aware data whitening technique to ensure a direct mapping between singular values and compression loss. Moreover, SVD-LLM adopts a parameter update with sequential low-rank approximation to compensate for the accuracy degradation after SVD compression. We evaluate SVD-LLM on 10 datasets and seven models from three different LLM families at three different scales. Our results demonstrate the superiority of SVD-LLM over state-of-the-arts, especially at high model compression ratios. Our code is available at https://github.com/AIoT-MLSys-Lab/SVD-LLM	[ { "version": "v1", "created": "Tue, 12 Mar 2024 07:31:18 GMT" }, { "version": "v2", "created": "Fri, 15 Mar 2024 02:59:10 GMT" }, { "version": "v3", "created": "Mon, 1 Apr 2024 15:04:15 GMT" }, { "version": "v4", "created": "Tue, 28 May 2024 13:41:26 GMT" }, { "version": "v5", "created": "Sun, 16 Mar 2025 03:27:33 GMT" } ]	2025-03-18T00:00:00	[ [ "Wang", "Xin", "" ], [ "Zheng", "Yu", "" ], [ "Wan", "Zhongwei", "" ], [ "Zhang", "Mi", "" ] ]	TITLE: SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression ABSTRACT: The advancements in Large Language Models (LLMs) have been hindered by their substantial sizes, which necessitates LLM compression methods for practical deployment. Singular Value Decomposition (SVD) offers a promising solution for LLM compression. However, state-of-the-art SVD-based LLM compression methods have two key limitations: truncating smaller singular values may lead to higher compression loss, and the lack of update on the compressed weights after SVD truncation. In this work, we propose SVD-LLM, a SVD-based post-training LLM compression method that addresses the limitations of existing methods. SVD-LLM incorporates a truncation-aware data whitening technique to ensure a direct mapping between singular values and compression loss. Moreover, SVD-LLM adopts a parameter update with sequential low-rank approximation to compensate for the accuracy degradation after SVD compression. We evaluate SVD-LLM on 10 datasets and seven models from three different LLM families at three different scales. Our results demonstrate the superiority of SVD-LLM over state-of-the-arts, especially at high model compression ratios. Our code is available at https://github.com/AIoT-MLSys-Lab/SVD-LLM
2403.09964	Zixin Yang	Zixin Yang, Richard Simon, Kelly Merrell, Cristian. A. Linte	Boundary Constraint-free Biomechanical Model-Based Surface Matching for Intraoperative Liver Deformation Correction	null	null	null	null	eess.IV cs.CV	http://creativecommons.org/licenses/by-nc-nd/4.0/	In image-guided liver surgery, 3D-3D non-rigid registration methods play a crucial role in estimating the mapping between the preoperative model and the intraoperative surface represented as point clouds, addressing the challenge of tissue deformation. Typically, these methods incorporate a biomechanical model, represented as a finite element model (FEM), into the strain energy term to regularize a surface matching term. We propose a 3D-3D non-rigid registration method that incorporates a modified FEM into the surface matching term. The modified FEM alleviates the need to specify boundary conditions, which is achieved by modifying the stiffness matrix of a FEM and using diagonal loading for stabilization. As a result, the modified surface matching term does not require the specification of boundary conditions or an additional strain energy term to regularize the surface matching term. Optimization is achieved through an accelerated gradient algorithm, further enhanced by our proposed method for determining the optimal step size. We evaluated our method and compared it to several state-of-the-art methods across various datasets. Our straightforward and effective approach consistently outperformed or achieved comparable performance to the state-of-the-art methods. Our code and datasets are available at https://github.com/zixinyang9109/BCF-FEM.	[ { "version": "v1", "created": "Fri, 15 Mar 2024 02:05:20 GMT" }, { "version": "v2", "created": "Mon, 9 Sep 2024 10:41:31 GMT" }, { "version": "v3", "created": "Tue, 4 Feb 2025 17:29:11 GMT" }, { "version": "v4", "created": "Mon, 17 Mar 2025 15:19:09 GMT" } ]	2025-03-18T00:00:00	[ [ "Yang", "Zixin", "" ], [ "Simon", "Richard", "" ], [ "Merrell", "Kelly", "" ], [ "Linte", "Cristian. A.", "" ] ]	TITLE: Boundary Constraint-free Biomechanical Model-Based Surface Matching for Intraoperative Liver Deformation Correction ABSTRACT: In image-guided liver surgery, 3D-3D non-rigid registration methods play a crucial role in estimating the mapping between the preoperative model and the intraoperative surface represented as point clouds, addressing the challenge of tissue deformation. Typically, these methods incorporate a biomechanical model, represented as a finite element model (FEM), into the strain energy term to regularize a surface matching term. We propose a 3D-3D non-rigid registration method that incorporates a modified FEM into the surface matching term. The modified FEM alleviates the need to specify boundary conditions, which is achieved by modifying the stiffness matrix of a FEM and using diagonal loading for stabilization. As a result, the modified surface matching term does not require the specification of boundary conditions or an additional strain energy term to regularize the surface matching term. Optimization is achieved through an accelerated gradient algorithm, further enhanced by our proposed method for determining the optimal step size. We evaluated our method and compared it to several state-of-the-art methods across various datasets. Our straightforward and effective approach consistently outperformed or achieved comparable performance to the state-of-the-art methods. Our code and datasets are available at https://github.com/zixinyang9109/BCF-FEM.
2403.13683	Chen Zhao	Chen Zhao, Tong Zhang, Zheng Dang, Mathieu Salzmann	DVMNet++: Rethinking Relative Pose Estimation for Unseen Objects	null	null	null	null	cs.CV cs.RO	http://creativecommons.org/licenses/by/4.0/	Determining the relative pose of a previously unseen object between two images is pivotal to the success of generalizable object pose estimation. Existing approaches typically predict 3D translation utilizing the ground-truth object bounding box and approximate 3D rotation with a large number of discrete hypotheses. This strategy makes unrealistic assumptions about the availability of ground truth and incurs a computationally expensive process of scoring each hypothesis at test time. By contrast, we rethink the problem of relative pose estimation for unseen objects by presenting a Deep Voxel Matching Network (DVMNet++). Our method computes the relative object pose in a single pass, eliminating the need for ground-truth object bounding boxes and rotation hypotheses. We achieve open-set object detection by leveraging image feature embedding and natural language understanding as reference. The detection result is then employed to approximate the translation parameters and crop the object from the query image. For rotation estimation, we map the two RGB images, i.e., reference and cropped query, to their respective voxelized 3D representations. The resulting voxels are passed through a rotation estimation module, which aligns the voxels and computes the rotation in an end-to-end fashion by solving a least-squares problem. To enhance robustness, we introduce a weighted closest voxel algorithm capable of mitigating the impact of noisy voxels. We conduct extensive experiments on the CO3D, Objaverse, LINEMOD, and LINEMOD-O datasets, demonstrating that our approach delivers more accurate relative pose estimates for novel objects at a lower computational cost compared to state-of-the-art methods. Our code is released at https://github.com/sailor-z/DVMNet/.	[ { "version": "v1", "created": "Wed, 20 Mar 2024 15:41:32 GMT" }, { "version": "v2", "created": "Sat, 15 Mar 2025 04:15:35 GMT" } ]	2025-03-18T00:00:00	[ [ "Zhao", "Chen", "" ], [ "Zhang", "Tong", "" ], [ "Dang", "Zheng", "" ], [ "Salzmann", "Mathieu", "" ] ]	TITLE: DVMNet++: Rethinking Relative Pose Estimation for Unseen Objects ABSTRACT: Determining the relative pose of a previously unseen object between two images is pivotal to the success of generalizable object pose estimation. Existing approaches typically predict 3D translation utilizing the ground-truth object bounding box and approximate 3D rotation with a large number of discrete hypotheses. This strategy makes unrealistic assumptions about the availability of ground truth and incurs a computationally expensive process of scoring each hypothesis at test time. By contrast, we rethink the problem of relative pose estimation for unseen objects by presenting a Deep Voxel Matching Network (DVMNet++). Our method computes the relative object pose in a single pass, eliminating the need for ground-truth object bounding boxes and rotation hypotheses. We achieve open-set object detection by leveraging image feature embedding and natural language understanding as reference. The detection result is then employed to approximate the translation parameters and crop the object from the query image. For rotation estimation, we map the two RGB images, i.e., reference and cropped query, to their respective voxelized 3D representations. The resulting voxels are passed through a rotation estimation module, which aligns the voxels and computes the rotation in an end-to-end fashion by solving a least-squares problem. To enhance robustness, we introduce a weighted closest voxel algorithm capable of mitigating the impact of noisy voxels. We conduct extensive experiments on the CO3D, Objaverse, LINEMOD, and LINEMOD-O datasets, demonstrating that our approach delivers more accurate relative pose estimates for novel objects at a lower computational cost compared to state-of-the-art methods. Our code is released at https://github.com/sailor-z/DVMNet/.

Subsets and Splits

No community queries yet

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