id
stringlengths 9
16
| submitter
stringlengths 3
64
⌀ | authors
stringlengths 5
6.63k
| title
stringlengths 7
245
| comments
stringlengths 1
482
⌀ | journal-ref
stringlengths 4
382
⌀ | doi
stringlengths 9
151
⌀ | report-no
stringclasses 984
values | categories
stringlengths 5
108
| license
stringclasses 9
values | abstract
stringlengths 83
3.41k
| versions
listlengths 1
20
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-11 00:00:00
| authors_parsed
listlengths 1
427
| prompt
stringlengths 166
3.49k
| label
stringclasses 2
values | prob
float64 0.5
0.98
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2503.10604 | Yingshuang Zou | Yingshuang Zou, Yikang Ding, Chuanrui Zhang, Jiazhe Guo, Bohan Li,
Xiaoyang Lyu, Feiyang Tan, Xiaojuan Qi, Haoqian Wang | MuDG: Taming Multi-modal Diffusion with Gaussian Splatting for Urban
Scene Reconstruction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent breakthroughs in radiance fields have significantly advanced 3D scene
reconstruction and novel view synthesis (NVS) in autonomous driving.
Nevertheless, critical limitations persist: reconstruction-based methods
exhibit substantial performance deterioration under significant viewpoint
deviations from training trajectories, while generation-based techniques
struggle with temporal coherence and precise scene controllability. To overcome
these challenges, we present MuDG, an innovative framework that integrates
Multi-modal Diffusion model with Gaussian Splatting (GS) for Urban Scene
Reconstruction. MuDG leverages aggregated LiDAR point clouds with RGB and
geometric priors to condition a multi-modal video diffusion model, synthesizing
photorealistic RGB, depth, and semantic outputs for novel viewpoints. This
synthesis pipeline enables feed-forward NVS without computationally intensive
per-scene optimization, providing comprehensive supervision signals to refine
3DGS representations for rendering robustness enhancement under extreme
viewpoint changes. Experiments on the Open Waymo Dataset demonstrate that MuDG
outperforms existing methods in both reconstruction and synthesis quality.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:48:41 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Zou",
"Yingshuang",
""
],
[
"Ding",
"Yikang",
""
],
[
"Zhang",
"Chuanrui",
""
],
[
"Guo",
"Jiazhe",
""
],
[
"Li",
"Bohan",
""
],
[
"Lyu",
"Xiaoyang",
""
],
[
"Tan",
"Feiyang",
""
],
[
"Qi",
"Xiaojuan",
""
],
[
"Wang",
"Haoqian",
""
]
]
| TITLE: MuDG: Taming Multi-modal Diffusion with Gaussian Splatting for Urban
Scene Reconstruction
ABSTRACT: Recent breakthroughs in radiance fields have significantly advanced 3D scene
reconstruction and novel view synthesis (NVS) in autonomous driving.
Nevertheless, critical limitations persist: reconstruction-based methods
exhibit substantial performance deterioration under significant viewpoint
deviations from training trajectories, while generation-based techniques
struggle with temporal coherence and precise scene controllability. To overcome
these challenges, we present MuDG, an innovative framework that integrates
Multi-modal Diffusion model with Gaussian Splatting (GS) for Urban Scene
Reconstruction. MuDG leverages aggregated LiDAR point clouds with RGB and
geometric priors to condition a multi-modal video diffusion model, synthesizing
photorealistic RGB, depth, and semantic outputs for novel viewpoints. This
synthesis pipeline enables feed-forward NVS without computationally intensive
per-scene optimization, providing comprehensive supervision signals to refine
3DGS representations for rendering robustness enhancement under extreme
viewpoint changes. Experiments on the Open Waymo Dataset demonstrate that MuDG
outperforms existing methods in both reconstruction and synthesis quality.
| no_new_dataset | 0.945751 |
2503.10605 | Alexey Nekrasov | Severin Heidrich, Till Beemelmanns, Alexey Nekrasov, Bastian Leibe,
Lutz Eckstein | OCCUQ: Exploring Efficient Uncertainty Quantification for 3D Occupancy
Prediction | Accepted for publication at ICRA 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Autonomous driving has the potential to significantly enhance productivity
and provide numerous societal benefits. Ensuring robustness in these
safety-critical systems is essential, particularly when vehicles must navigate
adverse weather conditions and sensor corruptions that may not have been
encountered during training. Current methods often overlook uncertainties
arising from adversarial conditions or distributional shifts, limiting their
real-world applicability. We propose an efficient adaptation of an uncertainty
estimation technique for 3D occupancy prediction. Our method dynamically
calibrates model confidence using epistemic uncertainty estimates. Our
evaluation under various camera corruption scenarios, such as fog or missing
cameras, demonstrates that our approach effectively quantifies epistemic
uncertainty by assigning higher uncertainty values to unseen data. We introduce
region-specific corruptions to simulate defects affecting only a single camera
and validate our findings through both scene-level and region-level
assessments. Our results show superior performance in Out-of-Distribution (OoD)
detection and confidence calibration compared to common baselines such as Deep
Ensembles and MC-Dropout. Our approach consistently demonstrates reliable
uncertainty measures, indicating its potential for enhancing the robustness of
autonomous driving systems in real-world scenarios. Code and dataset are
available at https://github.com/ika-rwth-aachen/OCCUQ .
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:50:07 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Heidrich",
"Severin",
""
],
[
"Beemelmanns",
"Till",
""
],
[
"Nekrasov",
"Alexey",
""
],
[
"Leibe",
"Bastian",
""
],
[
"Eckstein",
"Lutz",
""
]
]
| TITLE: OCCUQ: Exploring Efficient Uncertainty Quantification for 3D Occupancy
Prediction
ABSTRACT: Autonomous driving has the potential to significantly enhance productivity
and provide numerous societal benefits. Ensuring robustness in these
safety-critical systems is essential, particularly when vehicles must navigate
adverse weather conditions and sensor corruptions that may not have been
encountered during training. Current methods often overlook uncertainties
arising from adversarial conditions or distributional shifts, limiting their
real-world applicability. We propose an efficient adaptation of an uncertainty
estimation technique for 3D occupancy prediction. Our method dynamically
calibrates model confidence using epistemic uncertainty estimates. Our
evaluation under various camera corruption scenarios, such as fog or missing
cameras, demonstrates that our approach effectively quantifies epistemic
uncertainty by assigning higher uncertainty values to unseen data. We introduce
region-specific corruptions to simulate defects affecting only a single camera
and validate our findings through both scene-level and region-level
assessments. Our results show superior performance in Out-of-Distribution (OoD)
detection and confidence calibration compared to common baselines such as Deep
Ensembles and MC-Dropout. Our approach consistently demonstrates reliable
uncertainty measures, indicating its potential for enhancing the robustness of
autonomous driving systems in real-world scenarios. Code and dataset are
available at https://github.com/ika-rwth-aachen/OCCUQ .
| no_new_dataset | 0.881819 |
2503.10621 | Ayesha Ishaq Ms | Ayesha Ishaq, Jean Lahoud, Ketan More, Omkar Thawakar, Ritesh Thawkar,
Dinura Dissanayake, Noor Ahsan, Yuhao Li, Fahad Shahbaz Khan, Hisham
Cholakkal, Ivan Laptev, Rao Muhammad Anwer, Salman Khan | DriveLMM-o1: A Step-by-Step Reasoning Dataset and Large Multimodal Model
for Driving Scenario Understanding | 8 pages, 4 figures, 3 tables, github:
https://github.com/ayesha-ishaq/DriveLMM-o1 | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | While large multimodal models (LMMs) have demonstrated strong performance
across various Visual Question Answering (VQA) tasks, certain challenges
require complex multi-step reasoning to reach accurate answers. One
particularly challenging task is autonomous driving, which demands thorough
cognitive processing before decisions can be made. In this domain, a sequential
and interpretive understanding of visual cues is essential for effective
perception, prediction, and planning. Nevertheless, common VQA benchmarks often
focus on the accuracy of the final answer while overlooking the reasoning
process that enables the generation of accurate responses. Moreover, existing
methods lack a comprehensive framework for evaluating step-by-step reasoning in
realistic driving scenarios. To address this gap, we propose DriveLMM-o1, a new
dataset and benchmark specifically designed to advance step-wise visual
reasoning for autonomous driving. Our benchmark features over 18k VQA examples
in the training set and more than 4k in the test set, covering diverse
questions on perception, prediction, and planning, each enriched with
step-by-step reasoning to ensure logical inference in autonomous driving
scenarios. We further introduce a large multimodal model that is fine-tuned on
our reasoning dataset, demonstrating robust performance in complex driving
scenarios. In addition, we benchmark various open-source and closed-source
methods on our proposed dataset, systematically comparing their reasoning
capabilities for autonomous driving tasks. Our model achieves a +7.49% gain in
final answer accuracy, along with a 3.62% improvement in reasoning score over
the previous best open-source model. Our framework, dataset, and model are
available at https://github.com/ayesha-ishaq/DriveLMM-o1.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:01 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Ishaq",
"Ayesha",
""
],
[
"Lahoud",
"Jean",
""
],
[
"More",
"Ketan",
""
],
[
"Thawakar",
"Omkar",
""
],
[
"Thawkar",
"Ritesh",
""
],
[
"Dissanayake",
"Dinura",
""
],
[
"Ahsan",
"Noor",
""
],
[
"Li",
"Yuhao",
""
],
[
"Khan",
"Fahad Shahbaz",
""
],
[
"Cholakkal",
"Hisham",
""
],
[
"Laptev",
"Ivan",
""
],
[
"Anwer",
"Rao Muhammad",
""
],
[
"Khan",
"Salman",
""
]
]
| TITLE: DriveLMM-o1: A Step-by-Step Reasoning Dataset and Large Multimodal Model
for Driving Scenario Understanding
ABSTRACT: While large multimodal models (LMMs) have demonstrated strong performance
across various Visual Question Answering (VQA) tasks, certain challenges
require complex multi-step reasoning to reach accurate answers. One
particularly challenging task is autonomous driving, which demands thorough
cognitive processing before decisions can be made. In this domain, a sequential
and interpretive understanding of visual cues is essential for effective
perception, prediction, and planning. Nevertheless, common VQA benchmarks often
focus on the accuracy of the final answer while overlooking the reasoning
process that enables the generation of accurate responses. Moreover, existing
methods lack a comprehensive framework for evaluating step-by-step reasoning in
realistic driving scenarios. To address this gap, we propose DriveLMM-o1, a new
dataset and benchmark specifically designed to advance step-wise visual
reasoning for autonomous driving. Our benchmark features over 18k VQA examples
in the training set and more than 4k in the test set, covering diverse
questions on perception, prediction, and planning, each enriched with
step-by-step reasoning to ensure logical inference in autonomous driving
scenarios. We further introduce a large multimodal model that is fine-tuned on
our reasoning dataset, demonstrating robust performance in complex driving
scenarios. In addition, we benchmark various open-source and closed-source
methods on our proposed dataset, systematically comparing their reasoning
capabilities for autonomous driving tasks. Our model achieves a +7.49% gain in
final answer accuracy, along with a 3.62% improvement in reasoning score over
the previous best open-source model. Our framework, dataset, and model are
available at https://github.com/ayesha-ishaq/DriveLMM-o1.
| new_dataset | 0.963472 |
2503.10629 | Hashmat Shadab Malik | Hashmat Shadab Malik, Shahina Kunhimon, Muzammal Naseer, Fahad Shahbaz
Khan, Salman Khan | Hierarchical Self-Supervised Adversarial Training for Robust Vision
Models in Histopathology | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Adversarial attacks pose significant challenges for vision models in critical
fields like healthcare, where reliability is essential. Although adversarial
training has been well studied in natural images, its application to biomedical
and microscopy data remains limited. Existing self-supervised adversarial
training methods overlook the hierarchical structure of histopathology images,
where patient-slide-patch relationships provide valuable discriminative
signals. To address this, we propose Hierarchical Self-Supervised Adversarial
Training (HSAT), which exploits these properties to craft adversarial examples
using multi-level contrastive learning and integrate it into adversarial
training for enhanced robustness. We evaluate HSAT on multiclass histopathology
dataset OpenSRH and the results show that HSAT outperforms existing methods
from both biomedical and natural image domains. HSAT enhances robustness,
achieving an average gain of 54.31% in the white-box setting and reducing
performance drops to 3-4% in the black-box setting, compared to 25-30% for the
baseline. These results set a new benchmark for adversarial training in this
domain, paving the way for more robust models. Our Code for training and
evaluation is available at https://github.com/HashmatShadab/HSAT.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:47 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Malik",
"Hashmat Shadab",
""
],
[
"Kunhimon",
"Shahina",
""
],
[
"Naseer",
"Muzammal",
""
],
[
"Khan",
"Fahad Shahbaz",
""
],
[
"Khan",
"Salman",
""
]
]
| TITLE: Hierarchical Self-Supervised Adversarial Training for Robust Vision
Models in Histopathology
ABSTRACT: Adversarial attacks pose significant challenges for vision models in critical
fields like healthcare, where reliability is essential. Although adversarial
training has been well studied in natural images, its application to biomedical
and microscopy data remains limited. Existing self-supervised adversarial
training methods overlook the hierarchical structure of histopathology images,
where patient-slide-patch relationships provide valuable discriminative
signals. To address this, we propose Hierarchical Self-Supervised Adversarial
Training (HSAT), which exploits these properties to craft adversarial examples
using multi-level contrastive learning and integrate it into adversarial
training for enhanced robustness. We evaluate HSAT on multiclass histopathology
dataset OpenSRH and the results show that HSAT outperforms existing methods
from both biomedical and natural image domains. HSAT enhances robustness,
achieving an average gain of 54.31% in the white-box setting and reducing
performance drops to 3-4% in the black-box setting, compared to 25-30% for the
baseline. These results set a new benchmark for adversarial training in this
domain, paving the way for more robust models. Our Code for training and
evaluation is available at https://github.com/HashmatShadab/HSAT.
| no_new_dataset | 0.950503 |
2503.10632 | Subhajit Maity | Subhajit Maity, Killian Hitsman, Xin Li, Aritra Dutta | Kolmogorov-Arnold Attention: Is Learnable Attention Better For Vision
Transformers? | Preprint, Appendix included | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Kolmogorov-Arnold networks (KANs) are a remarkable innovation consisting of
learnable activation functions with the potential to capture more complex
relationships from data. Although KANs are useful in finding symbolic
representations and continual learning of one-dimensional functions, their
effectiveness in diverse machine learning (ML) tasks, such as vision, remains
questionable. Presently, KANs are deployed by replacing multilayer perceptrons
(MLPs) in deep network architectures, including advanced architectures such as
vision Transformers (ViTs). In this paper, we are the first to design a general
learnable Kolmogorov-Arnold Attention (KArAt) for vanilla ViTs that can operate
on any choice of basis. However, the computing and memory costs of training
them motivated us to propose a more modular version, and we designed particular
learnable attention, called Fourier-KArAt. Fourier-KArAt and its variants
either outperform their ViT counterparts or show comparable performance on
CIFAR-10, CIFAR-100, and ImageNet-1K datasets. We dissect these architectures'
performance and generalization capacity by analyzing their loss landscapes,
weight distributions, optimizer path, attention visualization, and spectral
behavior, and contrast them with vanilla ViTs. The goal of this paper is not to
produce parameter- and compute-efficient attention, but to encourage the
community to explore KANs in conjunction with more advanced architectures that
require a careful understanding of learnable activations. Our open-source code
and implementation details are available on: https://subhajitmaity.me/KArAt
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:52 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Maity",
"Subhajit",
""
],
[
"Hitsman",
"Killian",
""
],
[
"Li",
"Xin",
""
],
[
"Dutta",
"Aritra",
""
]
]
| TITLE: Kolmogorov-Arnold Attention: Is Learnable Attention Better For Vision
Transformers?
ABSTRACT: Kolmogorov-Arnold networks (KANs) are a remarkable innovation consisting of
learnable activation functions with the potential to capture more complex
relationships from data. Although KANs are useful in finding symbolic
representations and continual learning of one-dimensional functions, their
effectiveness in diverse machine learning (ML) tasks, such as vision, remains
questionable. Presently, KANs are deployed by replacing multilayer perceptrons
(MLPs) in deep network architectures, including advanced architectures such as
vision Transformers (ViTs). In this paper, we are the first to design a general
learnable Kolmogorov-Arnold Attention (KArAt) for vanilla ViTs that can operate
on any choice of basis. However, the computing and memory costs of training
them motivated us to propose a more modular version, and we designed particular
learnable attention, called Fourier-KArAt. Fourier-KArAt and its variants
either outperform their ViT counterparts or show comparable performance on
CIFAR-10, CIFAR-100, and ImageNet-1K datasets. We dissect these architectures'
performance and generalization capacity by analyzing their loss landscapes,
weight distributions, optimizer path, attention visualization, and spectral
behavior, and contrast them with vanilla ViTs. The goal of this paper is not to
produce parameter- and compute-efficient attention, but to encourage the
community to explore KANs in conjunction with more advanced architectures that
require a careful understanding of learnable activations. Our open-source code
and implementation details are available on: https://subhajitmaity.me/KArAt
| no_new_dataset | 0.945248 |
2503.10633 | Eliahu Horwitz | Eliahu Horwitz, Nitzan Kurer, Jonathan Kahana, Liel Amar, Yedid Hoshen | Charting and Navigating Hugging Face's Model Atlas | null | null | null | null | cs.LG cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As there are now millions of publicly available neural networks, searching
and analyzing large model repositories becomes increasingly important.
Navigating so many models requires an atlas, but as most models are poorly
documented charting such an atlas is challenging. To explore the hidden
potential of model repositories, we chart a preliminary atlas representing the
documented fraction of Hugging Face. It provides stunning visualizations of the
model landscape and evolution. We demonstrate several applications of this
atlas including predicting model attributes (e.g., accuracy), and analyzing
trends in computer vision models. However, as the current atlas remains
incomplete, we propose a method for charting undocumented regions.
Specifically, we identify high-confidence structural priors based on dominant
real-world model training practices. Leveraging these priors, our approach
enables accurate mapping of previously undocumented areas of the atlas. We
publicly release our datasets, code, and interactive atlas.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:53 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Horwitz",
"Eliahu",
""
],
[
"Kurer",
"Nitzan",
""
],
[
"Kahana",
"Jonathan",
""
],
[
"Amar",
"Liel",
""
],
[
"Hoshen",
"Yedid",
""
]
]
| TITLE: Charting and Navigating Hugging Face's Model Atlas
ABSTRACT: As there are now millions of publicly available neural networks, searching
and analyzing large model repositories becomes increasingly important.
Navigating so many models requires an atlas, but as most models are poorly
documented charting such an atlas is challenging. To explore the hidden
potential of model repositories, we chart a preliminary atlas representing the
documented fraction of Hugging Face. It provides stunning visualizations of the
model landscape and evolution. We demonstrate several applications of this
atlas including predicting model attributes (e.g., accuracy), and analyzing
trends in computer vision models. However, as the current atlas remains
incomplete, we propose a method for charting undocumented regions.
Specifically, we identify high-confidence structural priors based on dominant
real-world model training practices. Leveraging these priors, our approach
enables accurate mapping of previously undocumented areas of the atlas. We
publicly release our datasets, code, and interactive atlas.
| new_dataset | 0.953751 |
2503.10635 | Zhiqiang Shen | Zhaoyi Li and Xiaohan Zhao and Dong-Dong Wu and Jiacheng Cui and
Zhiqiang Shen | A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90%
Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1 | Code at: https://github.com/VILA-Lab/M-Attack | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Despite promising performance on open-source large vision-language models
(LVLMs), transfer-based targeted attacks often fail against black-box
commercial LVLMs. Analyzing failed adversarial perturbations reveals that the
learned perturbations typically originate from a uniform distribution and lack
clear semantic details, resulting in unintended responses. This critical
absence of semantic information leads commercial LVLMs to either ignore the
perturbation entirely or misinterpret its embedded semantics, thereby causing
the attack to fail. To overcome these issues, we notice that identifying core
semantic objects is a key objective for models trained with various datasets
and methodologies. This insight motivates our approach that refines semantic
clarity by encoding explicit semantic details within local regions, thus
ensuring interoperability and capturing finer-grained features, and by
concentrating modifications on semantically rich areas rather than applying
them uniformly. To achieve this, we propose a simple yet highly effective
solution: at each optimization step, the adversarial image is cropped randomly
by a controlled aspect ratio and scale, resized, and then aligned with the
target image in the embedding space. Experimental results confirm our
hypothesis. Our adversarial examples crafted with local-aggregated
perturbations focused on crucial regions exhibit surprisingly good
transferability to commercial LVLMs, including GPT-4.5, GPT-4o,
Gemini-2.0-flash, Claude-3.5-sonnet, Claude-3.7-sonnet, and even reasoning
models like o1, Claude-3.7-thinking and Gemini-2.0-flash-thinking. Our approach
achieves success rates exceeding 90% on GPT-4.5, 4o, and o1, significantly
outperforming all prior state-of-the-art attack methods. Our optimized
adversarial examples under different configurations and training code are
available at https://github.com/VILA-Lab/M-Attack.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:55 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Li",
"Zhaoyi",
""
],
[
"Zhao",
"Xiaohan",
""
],
[
"Wu",
"Dong-Dong",
""
],
[
"Cui",
"Jiacheng",
""
],
[
"Shen",
"Zhiqiang",
""
]
]
| TITLE: A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90%
Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1
ABSTRACT: Despite promising performance on open-source large vision-language models
(LVLMs), transfer-based targeted attacks often fail against black-box
commercial LVLMs. Analyzing failed adversarial perturbations reveals that the
learned perturbations typically originate from a uniform distribution and lack
clear semantic details, resulting in unintended responses. This critical
absence of semantic information leads commercial LVLMs to either ignore the
perturbation entirely or misinterpret its embedded semantics, thereby causing
the attack to fail. To overcome these issues, we notice that identifying core
semantic objects is a key objective for models trained with various datasets
and methodologies. This insight motivates our approach that refines semantic
clarity by encoding explicit semantic details within local regions, thus
ensuring interoperability and capturing finer-grained features, and by
concentrating modifications on semantically rich areas rather than applying
them uniformly. To achieve this, we propose a simple yet highly effective
solution: at each optimization step, the adversarial image is cropped randomly
by a controlled aspect ratio and scale, resized, and then aligned with the
target image in the embedding space. Experimental results confirm our
hypothesis. Our adversarial examples crafted with local-aggregated
perturbations focused on crucial regions exhibit surprisingly good
transferability to commercial LVLMs, including GPT-4.5, GPT-4o,
Gemini-2.0-flash, Claude-3.5-sonnet, Claude-3.7-sonnet, and even reasoning
models like o1, Claude-3.7-thinking and Gemini-2.0-flash-thinking. Our approach
achieves success rates exceeding 90% on GPT-4.5, 4o, and o1, significantly
outperforming all prior state-of-the-art attack methods. Our optimized
adversarial examples under different configurations and training code are
available at https://github.com/VILA-Lab/M-Attack.
| no_new_dataset | 0.947866 |
2503.10638 | Xiaoming Zhao | Xiaoming Zhao, Alexander G. Schwing | Studying Classifier(-Free) Guidance From a Classifier-Centric
Perspective | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Classifier-free guidance has become a staple for conditional generation with
denoising diffusion models. However, a comprehensive understanding of
classifier-free guidance is still missing. In this work, we carry out an
empirical study to provide a fresh perspective on classifier-free guidance.
Concretely, instead of solely focusing on classifier-free guidance, we trace
back to the root, i.e., classifier guidance, pinpoint the key assumption for
the derivation, and conduct a systematic study to understand the role of the
classifier. We find that both classifier guidance and classifier-free guidance
achieve conditional generation by pushing the denoising diffusion trajectories
away from decision boundaries, i.e., areas where conditional information is
usually entangled and is hard to learn. Based on this classifier-centric
understanding, we propose a generic postprocessing step built upon
flow-matching to shrink the gap between the learned distribution for a
pre-trained denoising diffusion model and the real data distribution, majorly
around the decision boundaries. Experiments on various datasets verify the
effectiveness of the proposed approach.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:59 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Zhao",
"Xiaoming",
""
],
[
"Schwing",
"Alexander G.",
""
]
]
| TITLE: Studying Classifier(-Free) Guidance From a Classifier-Centric
Perspective
ABSTRACT: Classifier-free guidance has become a staple for conditional generation with
denoising diffusion models. However, a comprehensive understanding of
classifier-free guidance is still missing. In this work, we carry out an
empirical study to provide a fresh perspective on classifier-free guidance.
Concretely, instead of solely focusing on classifier-free guidance, we trace
back to the root, i.e., classifier guidance, pinpoint the key assumption for
the derivation, and conduct a systematic study to understand the role of the
classifier. We find that both classifier guidance and classifier-free guidance
achieve conditional generation by pushing the denoising diffusion trajectories
away from decision boundaries, i.e., areas where conditional information is
usually entangled and is hard to learn. Based on this classifier-centric
understanding, we propose a generic postprocessing step built upon
flow-matching to shrink the gap between the learned distribution for a
pre-trained denoising diffusion model and the real data distribution, majorly
around the decision boundaries. Experiments on various datasets verify the
effectiveness of the proposed approach.
| no_new_dataset | 0.949902 |
2503.10639 | Rongyao Fang | Rongyao Fang, Chengqi Duan, Kun Wang, Linjiang Huang, Hao Li, Shilin
Yan, Hao Tian, Xingyu Zeng, Rui Zhao, Jifeng Dai, Xihui Liu, Hongsheng Li | GoT: Unleashing Reasoning Capability of Multimodal Large Language Model
for Visual Generation and Editing | Dataset and models are released in https://github.com/rongyaofang/GoT | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Current image generation and editing methods primarily process textual
prompts as direct inputs without reasoning about visual composition and
explicit operations. We present Generation Chain-of-Thought (GoT), a novel
paradigm that enables generation and editing through an explicit language
reasoning process before outputting images. This approach transforms
conventional text-to-image generation and editing into a reasoning-guided
framework that analyzes semantic relationships and spatial arrangements. We
define the formulation of GoT and construct large-scale GoT datasets containing
over 9M samples with detailed reasoning chains capturing semantic-spatial
relationships. To leverage the advantages of GoT, we implement a unified
framework that integrates Qwen2.5-VL for reasoning chain generation with an
end-to-end diffusion model enhanced by our novel Semantic-Spatial Guidance
Module. Experiments show our GoT framework achieves excellent performance on
both generation and editing tasks, with significant improvements over
baselines. Additionally, our approach enables interactive visual generation,
allowing users to explicitly modify reasoning steps for precise image
adjustments. GoT pioneers a new direction for reasoning-driven visual
generation and editing, producing images that better align with human intent.
To facilitate future research, we make our datasets, code, and pretrained
models publicly available at https://github.com/rongyaofang/GoT.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 17:59:59 GMT"
}
]
| 2025-03-14T00:00:00 | [
[
"Fang",
"Rongyao",
""
],
[
"Duan",
"Chengqi",
""
],
[
"Wang",
"Kun",
""
],
[
"Huang",
"Linjiang",
""
],
[
"Li",
"Hao",
""
],
[
"Yan",
"Shilin",
""
],
[
"Tian",
"Hao",
""
],
[
"Zeng",
"Xingyu",
""
],
[
"Zhao",
"Rui",
""
],
[
"Dai",
"Jifeng",
""
],
[
"Liu",
"Xihui",
""
],
[
"Li",
"Hongsheng",
""
]
]
| TITLE: GoT: Unleashing Reasoning Capability of Multimodal Large Language Model
for Visual Generation and Editing
ABSTRACT: Current image generation and editing methods primarily process textual
prompts as direct inputs without reasoning about visual composition and
explicit operations. We present Generation Chain-of-Thought (GoT), a novel
paradigm that enables generation and editing through an explicit language
reasoning process before outputting images. This approach transforms
conventional text-to-image generation and editing into a reasoning-guided
framework that analyzes semantic relationships and spatial arrangements. We
define the formulation of GoT and construct large-scale GoT datasets containing
over 9M samples with detailed reasoning chains capturing semantic-spatial
relationships. To leverage the advantages of GoT, we implement a unified
framework that integrates Qwen2.5-VL for reasoning chain generation with an
end-to-end diffusion model enhanced by our novel Semantic-Spatial Guidance
Module. Experiments show our GoT framework achieves excellent performance on
both generation and editing tasks, with significant improvements over
baselines. Additionally, our approach enables interactive visual generation,
allowing users to explicitly modify reasoning steps for precise image
adjustments. GoT pioneers a new direction for reasoning-driven visual
generation and editing, producing images that better align with human intent.
To facilitate future research, we make our datasets, code, and pretrained
models publicly available at https://github.com/rongyaofang/GoT.
| new_dataset | 0.662387 |
2202.04348 | Yunli Wang | Siguang Huang, Yunli Wang, Lili Mou, Huayue Zhang, Han Zhu, Chuan Yu,
Bo Zheng | MBCT: Tree-Based Feature-Aware Binning for Individual Uncertainty
Calibration | WWW 2022. The new version fixed an error in Eq13 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most machine learning classifiers only concern classification accuracy, while
certain applications (such as medical diagnosis, meteorological forecasting,
and computation advertising) require the model to predict the true probability,
known as a calibrated estimate. In previous work, researchers have developed
several calibration methods to post-process the outputs of a predictor to
obtain calibrated values, such as binning and scaling methods. Compared with
scaling, binning methods are shown to have distribution-free theoretical
guarantees, which motivates us to prefer binning methods for calibration.
However, we notice that existing binning methods have several drawbacks: (a)
the binning scheme only considers the original prediction values, thus limiting
the calibration performance; and (b) the binning approach is non-individual,
mapping multiple samples in a bin to the same value, and thus is not suitable
for order-sensitive applications. In this paper, we propose a feature-aware
binning framework, called Multiple Boosting Calibration Trees (MBCT), along
with a multi-view calibration loss to tackle the above issues. Our MBCT
optimizes the binning scheme by the tree structures of features, and adopts a
linear function in a tree node to achieve individual calibration. Our MBCT is
non-monotonic, and has the potential to improve order accuracy, due to its
learnable binning scheme and the individual calibration. We conduct
comprehensive experiments on three datasets in different fields. Results show
that our method outperforms all competing models in terms of both calibration
error and order accuracy. We also conduct simulation experiments, justifying
that the proposed multi-view calibration loss is a better metric in modeling
calibration error.
| [
{
"version": "v1",
"created": "Wed, 9 Feb 2022 08:59:16 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:15:57 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Huang",
"Siguang",
""
],
[
"Wang",
"Yunli",
""
],
[
"Mou",
"Lili",
""
],
[
"Zhang",
"Huayue",
""
],
[
"Zhu",
"Han",
""
],
[
"Yu",
"Chuan",
""
],
[
"Zheng",
"Bo",
""
]
]
| TITLE: MBCT: Tree-Based Feature-Aware Binning for Individual Uncertainty
Calibration
ABSTRACT: Most machine learning classifiers only concern classification accuracy, while
certain applications (such as medical diagnosis, meteorological forecasting,
and computation advertising) require the model to predict the true probability,
known as a calibrated estimate. In previous work, researchers have developed
several calibration methods to post-process the outputs of a predictor to
obtain calibrated values, such as binning and scaling methods. Compared with
scaling, binning methods are shown to have distribution-free theoretical
guarantees, which motivates us to prefer binning methods for calibration.
However, we notice that existing binning methods have several drawbacks: (a)
the binning scheme only considers the original prediction values, thus limiting
the calibration performance; and (b) the binning approach is non-individual,
mapping multiple samples in a bin to the same value, and thus is not suitable
for order-sensitive applications. In this paper, we propose a feature-aware
binning framework, called Multiple Boosting Calibration Trees (MBCT), along
with a multi-view calibration loss to tackle the above issues. Our MBCT
optimizes the binning scheme by the tree structures of features, and adopts a
linear function in a tree node to achieve individual calibration. Our MBCT is
non-monotonic, and has the potential to improve order accuracy, due to its
learnable binning scheme and the individual calibration. We conduct
comprehensive experiments on three datasets in different fields. Results show
that our method outperforms all competing models in terms of both calibration
error and order accuracy. We also conduct simulation experiments, justifying
that the proposed multi-view calibration loss is a better metric in modeling
calibration error.
| no_new_dataset | 0.943556 |
2208.11636 | Julius Gonsior | Julius Gonsior, Maik Thiele, Wolfgang Lehner | ImitAL: Learned Active Learning Strategy on Synthetic Data | arXiv admin note: text overlap with arXiv:2108.07670 | null | 10.1007/978-3-031-18840-4_4 | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Active Learning (AL) is a well-known standard method for efficiently
obtaining annotated data by first labeling the samples that contain the most
information based on a query strategy. In the past, a large variety of such
query strategies has been proposed, with each generation of new strategies
increasing the runtime and adding more complexity. However, to the best of our
our knowledge, none of these strategies excels consistently over a large number
of datasets from different application domains. Basically, most of the the
existing AL strategies are a combination of the two simple heuristics
informativeness and representativeness, and the big differences lie in the
combination of the often conflicting heuristics. Within this paper, we propose
ImitAL, a domain-independent novel query strategy, which encodes AL as a
learning-to-rank problem and learns an optimal combination between both
heuristics. We train ImitAL on large-scale simulated AL runs on purely
synthetic datasets. To show that ImitAL was successfully trained, we perform an
extensive evaluation comparing our strategy on 13 different datasets, from a
wide range of domains, with 7 other query strategies.
| [
{
"version": "v1",
"created": "Wed, 24 Aug 2022 16:17:53 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Gonsior",
"Julius",
""
],
[
"Thiele",
"Maik",
""
],
[
"Lehner",
"Wolfgang",
""
]
]
| TITLE: ImitAL: Learned Active Learning Strategy on Synthetic Data
ABSTRACT: Active Learning (AL) is a well-known standard method for efficiently
obtaining annotated data by first labeling the samples that contain the most
information based on a query strategy. In the past, a large variety of such
query strategies has been proposed, with each generation of new strategies
increasing the runtime and adding more complexity. However, to the best of our
our knowledge, none of these strategies excels consistently over a large number
of datasets from different application domains. Basically, most of the the
existing AL strategies are a combination of the two simple heuristics
informativeness and representativeness, and the big differences lie in the
combination of the often conflicting heuristics. Within this paper, we propose
ImitAL, a domain-independent novel query strategy, which encodes AL as a
learning-to-rank problem and learns an optimal combination between both
heuristics. We train ImitAL on large-scale simulated AL runs on purely
synthetic datasets. To show that ImitAL was successfully trained, we perform an
extensive evaluation comparing our strategy on 13 different datasets, from a
wide range of domains, with 7 other query strategies.
| no_new_dataset | 0.948106 |
2210.03005 | Julius Gonsior | Julius Gonsior, Christian Falkenberg, Silvio Magino, Anja Reusch, Maik
Thiele, Wolfgang Lehner | To Softmax, or not to Softmax: that is the question when applying Active
Learning for Transformer Models | null | null | 10.1007/978-3-031-42914-9_9 | null | cs.LG cs.AI cs.CL cs.DB | http://creativecommons.org/licenses/by/4.0/ | Despite achieving state-of-the-art results in nearly all Natural Language
Processing applications, fine-tuning Transformer-based language models still
requires a significant amount of labeled data to work. A well known technique
to reduce the amount of human effort in acquiring a labeled dataset is
\textit{Active Learning} (AL): an iterative process in which only the minimal
amount of samples is labeled. AL strategies require access to a quantified
confidence measure of the model predictions. A common choice is the softmax
activation function for the final layer. As the softmax function provides
misleading probabilities, this paper compares eight alternatives on seven
datasets. Our almost paradoxical finding is that most of the methods are too
good at identifying the true most uncertain samples (outliers), and that
labeling therefore exclusively outliers results in worse performance. As a
heuristic we propose to systematically ignore samples, which results in
improvements of various methods compared to the softmax function.
| [
{
"version": "v1",
"created": "Thu, 6 Oct 2022 15:51:39 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Gonsior",
"Julius",
""
],
[
"Falkenberg",
"Christian",
""
],
[
"Magino",
"Silvio",
""
],
[
"Reusch",
"Anja",
""
],
[
"Thiele",
"Maik",
""
],
[
"Lehner",
"Wolfgang",
""
]
]
| TITLE: To Softmax, or not to Softmax: that is the question when applying Active
Learning for Transformer Models
ABSTRACT: Despite achieving state-of-the-art results in nearly all Natural Language
Processing applications, fine-tuning Transformer-based language models still
requires a significant amount of labeled data to work. A well known technique
to reduce the amount of human effort in acquiring a labeled dataset is
\textit{Active Learning} (AL): an iterative process in which only the minimal
amount of samples is labeled. AL strategies require access to a quantified
confidence measure of the model predictions. A common choice is the softmax
activation function for the final layer. As the softmax function provides
misleading probabilities, this paper compares eight alternatives on seven
datasets. Our almost paradoxical finding is that most of the methods are too
good at identifying the true most uncertain samples (outliers), and that
labeling therefore exclusively outliers results in worse performance. As a
heuristic we propose to systematically ignore samples, which results in
improvements of various methods compared to the softmax function.
| no_new_dataset | 0.949153 |
2303.02278 | Chun-Yin Huang | Chun-Yin Huang, Ruinan Jin, Can Zhao, Daguang Xu, and Xiaoxiao Li | Federated Learning on Virtual Heterogeneous Data with Local-global
Distillation | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While Federated Learning (FL) is gaining popularity for training machine
learning models in a decentralized fashion, numerous challenges persist, such
as asynchronization, computational expenses, data heterogeneity, and gradient
and membership privacy attacks. Lately, dataset distillation has emerged as a
promising solution for addressing the aforementioned challenges by generating a
compact synthetic dataset that preserves a model's training efficacy. However,
we discover that using distilled local datasets can amplify the heterogeneity
issue in FL. To address this, we propose Federated Learning on Virtual
Heterogeneous Data with Local-Global Dataset Distillation (FedLGD), where we
seamlessly integrate dataset distillation algorithms into FL pipeline and train
FL using a smaller synthetic dataset (referred as virtual data). Specifically,
to harmonize the domain shifts, we propose iterative distribution matching to
inpaint global information to local virtual data and use federated gradient
matching to distill global virtual data that serve as anchor points to rectify
heterogeneous local training, without compromising data privacy. We experiment
on both benchmark and real-world datasets that contain heterogeneous data from
different sources, and further scale up to an FL scenario that contains a large
number of clients with heterogeneous and class-imbalanced data. Our method
outperforms state-of-the-art heterogeneous FL algorithms under various
settings. Our code is available at https://github.com/ubc-tea/FedLGD.
| [
{
"version": "v1",
"created": "Sat, 4 Mar 2023 00:35:29 GMT"
},
{
"version": "v2",
"created": "Mon, 5 Jun 2023 18:43:26 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 01:01:17 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Huang",
"Chun-Yin",
""
],
[
"Jin",
"Ruinan",
""
],
[
"Zhao",
"Can",
""
],
[
"Xu",
"Daguang",
""
],
[
"Li",
"Xiaoxiao",
""
]
]
| TITLE: Federated Learning on Virtual Heterogeneous Data with Local-global
Distillation
ABSTRACT: While Federated Learning (FL) is gaining popularity for training machine
learning models in a decentralized fashion, numerous challenges persist, such
as asynchronization, computational expenses, data heterogeneity, and gradient
and membership privacy attacks. Lately, dataset distillation has emerged as a
promising solution for addressing the aforementioned challenges by generating a
compact synthetic dataset that preserves a model's training efficacy. However,
we discover that using distilled local datasets can amplify the heterogeneity
issue in FL. To address this, we propose Federated Learning on Virtual
Heterogeneous Data with Local-Global Dataset Distillation (FedLGD), where we
seamlessly integrate dataset distillation algorithms into FL pipeline and train
FL using a smaller synthetic dataset (referred as virtual data). Specifically,
to harmonize the domain shifts, we propose iterative distribution matching to
inpaint global information to local virtual data and use federated gradient
matching to distill global virtual data that serve as anchor points to rectify
heterogeneous local training, without compromising data privacy. We experiment
on both benchmark and real-world datasets that contain heterogeneous data from
different sources, and further scale up to an FL scenario that contains a large
number of clients with heterogeneous and class-imbalanced data. Our method
outperforms state-of-the-art heterogeneous FL algorithms under various
settings. Our code is available at https://github.com/ubc-tea/FedLGD.
| no_new_dataset | 0.946941 |
2308.04371 | Yifan Zhang | Yifan Zhang, Jingqin Yang, Yang Yuan, Andrew Chi-Chih Yao | Cumulative Reasoning with Large Language Models | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in large language models (LLMs) have shown remarkable
progress, yet their ability to solve complex problems remains limited. In this
work, we introduce Cumulative Reasoning (CR), an approach that utilizes LLMs
cumulatively and iteratively, mirroring human thought processes for
problem-solving. CR decomposes tasks into smaller, manageable components and
leverages previous propositions for effective composition, significantly
enhancing problem-solving capabilities. We demonstrate CR's advantage through
several complex reasoning tasks: it outperforms existing methods in logical
inference tasks with up to a 9.3% improvement, achieving 98.04% accuracy on the
curated FOLIO wiki dataset. In the Game of 24, it achieves 98% accuracy,
marking a 24% improvement over the prior state-of-the-art. In solving MATH
problems, CR achieves a 4.2% increase from previous methods and a 43% relative
improvement in the most challenging level 5 problems. When incorporating a code
environment with CR, we further harness LLMs' reasoning capabilities and
outperform the Program of Thought (PoT) method by 38.8%. The code is available
at https://github.com/iiis-ai/cumulative-reasoning.
| [
{
"version": "v1",
"created": "Tue, 8 Aug 2023 16:18:20 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Aug 2023 14:37:37 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Aug 2023 08:24:09 GMT"
},
{
"version": "v4",
"created": "Fri, 25 Aug 2023 02:40:37 GMT"
},
{
"version": "v5",
"created": "Sat, 2 Dec 2023 02:59:12 GMT"
},
{
"version": "v6",
"created": "Tue, 2 Apr 2024 03:37:39 GMT"
},
{
"version": "v7",
"created": "Wed, 12 Mar 2025 02:55:36 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhang",
"Yifan",
""
],
[
"Yang",
"Jingqin",
""
],
[
"Yuan",
"Yang",
""
],
[
"Yao",
"Andrew Chi-Chih",
""
]
]
| TITLE: Cumulative Reasoning with Large Language Models
ABSTRACT: Recent advancements in large language models (LLMs) have shown remarkable
progress, yet their ability to solve complex problems remains limited. In this
work, we introduce Cumulative Reasoning (CR), an approach that utilizes LLMs
cumulatively and iteratively, mirroring human thought processes for
problem-solving. CR decomposes tasks into smaller, manageable components and
leverages previous propositions for effective composition, significantly
enhancing problem-solving capabilities. We demonstrate CR's advantage through
several complex reasoning tasks: it outperforms existing methods in logical
inference tasks with up to a 9.3% improvement, achieving 98.04% accuracy on the
curated FOLIO wiki dataset. In the Game of 24, it achieves 98% accuracy,
marking a 24% improvement over the prior state-of-the-art. In solving MATH
problems, CR achieves a 4.2% increase from previous methods and a 43% relative
improvement in the most challenging level 5 problems. When incorporating a code
environment with CR, we further harness LLMs' reasoning capabilities and
outperform the Program of Thought (PoT) method by 38.8%. The code is available
at https://github.com/iiis-ai/cumulative-reasoning.
| new_dataset | 0.954223 |
2309.16460 | Qianyu Zhou | Shaocong Long, Qianyu Zhou, Chenhao Ying, Lizhuang Ma, Yuan Luo | Diverse Target and Contribution Scheduling for Domain Generalization | This work has been submitted to the IEEE for possible publication | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generalization under the distribution shift has been a great challenge in
computer vision. The prevailing practice of directly employing the one-hot
labels as the training targets in domain generalization~(DG) can lead to
gradient conflicts, making it insufficient for capturing the intrinsic class
characteristics and hard to increase the intra-class variation. Besides,
existing methods in DG mostly overlook the distinct contributions of source
(seen) domains, resulting in uneven learning from these domains. To address
these issues, we firstly present a theoretical and empirical analysis of the
existence of gradient conflicts in DG, unveiling the previously unexplored
relationship between distribution shifts and gradient conflicts during the
optimization process. In this paper, we present a novel perspective of DG from
the empirical source domain's risk and propose a new paradigm for DG called
Diverse Target and Contribution Scheduling (DTCS). DTCS comprises two
innovative modules: Diverse Target Supervision (DTS) and Diverse Contribution
Balance (DCB), with the aim of addressing the limitations associated with the
common utilization of one-hot labels and equal contributions for source domains
in DG. In specific, DTS employs distinct soft labels as training targets to
account for various feature distributions across domains and thereby mitigates
the gradient conflicts, and DCB dynamically balances the contributions of
source domains by ensuring a fair decline in losses of different source
domains. Extensive experiments with analysis on four benchmark datasets show
that the proposed method achieves a competitive performance in comparison with
the state-of-the-art approaches, demonstrating the effectiveness and advantages
of the proposed DTCS.
| [
{
"version": "v1",
"created": "Thu, 28 Sep 2023 14:10:25 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:24:26 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Long",
"Shaocong",
""
],
[
"Zhou",
"Qianyu",
""
],
[
"Ying",
"Chenhao",
""
],
[
"Ma",
"Lizhuang",
""
],
[
"Luo",
"Yuan",
""
]
]
| TITLE: Diverse Target and Contribution Scheduling for Domain Generalization
ABSTRACT: Generalization under the distribution shift has been a great challenge in
computer vision. The prevailing practice of directly employing the one-hot
labels as the training targets in domain generalization~(DG) can lead to
gradient conflicts, making it insufficient for capturing the intrinsic class
characteristics and hard to increase the intra-class variation. Besides,
existing methods in DG mostly overlook the distinct contributions of source
(seen) domains, resulting in uneven learning from these domains. To address
these issues, we firstly present a theoretical and empirical analysis of the
existence of gradient conflicts in DG, unveiling the previously unexplored
relationship between distribution shifts and gradient conflicts during the
optimization process. In this paper, we present a novel perspective of DG from
the empirical source domain's risk and propose a new paradigm for DG called
Diverse Target and Contribution Scheduling (DTCS). DTCS comprises two
innovative modules: Diverse Target Supervision (DTS) and Diverse Contribution
Balance (DCB), with the aim of addressing the limitations associated with the
common utilization of one-hot labels and equal contributions for source domains
in DG. In specific, DTS employs distinct soft labels as training targets to
account for various feature distributions across domains and thereby mitigates
the gradient conflicts, and DCB dynamically balances the contributions of
source domains by ensuring a fair decline in losses of different source
domains. Extensive experiments with analysis on four benchmark datasets show
that the proposed method achieves a competitive performance in comparison with
the state-of-the-art approaches, demonstrating the effectiveness and advantages
of the proposed DTCS.
| no_new_dataset | 0.949342 |
2310.07259 | Haoyu Zhang | Haoyu Zhang, Meng Liu, Yisen Feng, Yaowei Wang, Weili Guan, Liqiang
Nie | Uncovering Hidden Connections: Iterative Search and Reasoning for
Video-grounded Dialog | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In contrast to conventional visual question answering, video-grounded dialog
necessitates a profound understanding of both dialog history and video content
for accurate response generation. Despite commendable progress made by existing
approaches, they still face the challenges of incrementally understanding
complex dialog history and assimilating video information. In response to these
challenges, we present an iterative search and reasoning framework, which
consists of a textual encoder, a visual encoder, and a generator. Specifically,
we devise a path search and aggregation strategy in the textual encoder, mining
core cues from dialog history that are pivotal to understanding the posed
questions. Concurrently, our visual encoder harnesses an iterative reasoning
network to extract and emphasize critical visual markers from videos, enhancing
the depth of visual comprehension. Finally, we utilize the pre-trained GPT-2
model as our answer generator to decode the mined hidden clues into coherent
and contextualized answers. Extensive experiments on three public datasets
demonstrate the effectiveness and generalizability of our proposed framework.
| [
{
"version": "v1",
"created": "Wed, 11 Oct 2023 07:37:13 GMT"
},
{
"version": "v2",
"created": "Wed, 22 May 2024 11:58:12 GMT"
},
{
"version": "v3",
"created": "Mon, 18 Nov 2024 02:18:14 GMT"
},
{
"version": "v4",
"created": "Wed, 12 Mar 2025 05:09:37 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhang",
"Haoyu",
""
],
[
"Liu",
"Meng",
""
],
[
"Feng",
"Yisen",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Guan",
"Weili",
""
],
[
"Nie",
"Liqiang",
""
]
]
| TITLE: Uncovering Hidden Connections: Iterative Search and Reasoning for
Video-grounded Dialog
ABSTRACT: In contrast to conventional visual question answering, video-grounded dialog
necessitates a profound understanding of both dialog history and video content
for accurate response generation. Despite commendable progress made by existing
approaches, they still face the challenges of incrementally understanding
complex dialog history and assimilating video information. In response to these
challenges, we present an iterative search and reasoning framework, which
consists of a textual encoder, a visual encoder, and a generator. Specifically,
we devise a path search and aggregation strategy in the textual encoder, mining
core cues from dialog history that are pivotal to understanding the posed
questions. Concurrently, our visual encoder harnesses an iterative reasoning
network to extract and emphasize critical visual markers from videos, enhancing
the depth of visual comprehension. Finally, we utilize the pre-trained GPT-2
model as our answer generator to decode the mined hidden clues into coherent
and contextualized answers. Extensive experiments on three public datasets
demonstrate the effectiveness and generalizability of our proposed framework.
| no_new_dataset | 0.933734 |
2310.14687 | Yihan Cao | Yihan Cao, Shuyi Chen, Ryan Liu, Zhiruo Wang, Daniel Fried | API-Assisted Code Generation for Question Answering on Varied Table
Structures | EMNLP 2023 camera ready, 13 pages, 11 figures | Proceedings of the Conference on Empirical Methods in Natural
Language Processing, Association for Computational Linguistics, 2023, pages
14536-14548, Singapore | 10.18653/v1/2023.emnlp-main.897 | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | A persistent challenge to table question answering (TableQA) by generating
executable programs has been adapting to varied table structures, typically
requiring domain-specific logical forms. In response, this paper introduces a
unified TableQA framework that: (1) provides a unified representation for
structured tables as multi-index Pandas data frames, (2) uses Python as a
powerful querying language, and (3) uses few-shot prompting to translate NL
questions into Python programs, which are executable on Pandas data frames.
Furthermore, to answer complex relational questions with extended program
functionality and external knowledge, our framework allows customized APIs that
Python programs can call. We experiment with four TableQA datasets that involve
tables of different structures -- relational, multi-table, and hierarchical
matrix shapes -- and achieve prominent improvements over past state-of-the-art
systems. In ablation studies, we (1) show benefits from our multi-index
representation and APIs over baselines that use only an LLM, and (2)
demonstrate that our approach is modular and can incorporate additional APIs.
| [
{
"version": "v1",
"created": "Mon, 23 Oct 2023 08:26:28 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Cao",
"Yihan",
""
],
[
"Chen",
"Shuyi",
""
],
[
"Liu",
"Ryan",
""
],
[
"Wang",
"Zhiruo",
""
],
[
"Fried",
"Daniel",
""
]
]
| TITLE: API-Assisted Code Generation for Question Answering on Varied Table
Structures
ABSTRACT: A persistent challenge to table question answering (TableQA) by generating
executable programs has been adapting to varied table structures, typically
requiring domain-specific logical forms. In response, this paper introduces a
unified TableQA framework that: (1) provides a unified representation for
structured tables as multi-index Pandas data frames, (2) uses Python as a
powerful querying language, and (3) uses few-shot prompting to translate NL
questions into Python programs, which are executable on Pandas data frames.
Furthermore, to answer complex relational questions with extended program
functionality and external knowledge, our framework allows customized APIs that
Python programs can call. We experiment with four TableQA datasets that involve
tables of different structures -- relational, multi-table, and hierarchical
matrix shapes -- and achieve prominent improvements over past state-of-the-art
systems. In ablation studies, we (1) show benefits from our multi-index
representation and APIs over baselines that use only an LLM, and (2)
demonstrate that our approach is modular and can incorporate additional APIs.
| no_new_dataset | 0.91782 |
2312.04539 | Osman \"Ulger | Osman \"Ulger, Maksymilian Kulicki, Yuki Asano, Martin R. Oswald | Auto-Vocabulary Semantic Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Open-Vocabulary Segmentation (OVS) methods are capable of performing semantic
segmentation without relying on a fixed vocabulary, and in some cases, without
training or fine-tuning. However, OVS methods typically require a human in the
loop to specify the vocabulary based on the task or dataset at hand. In this
paper, we introduce Auto-Vocabulary Semantic Segmentation (AVS), advancing
open-ended image understanding by eliminating the necessity to predefine object
categories for segmentation. Our approach, AutoSeg, presents a framework that
autonomously identifies relevant class names using semantically enhanced BLIP
embeddings and segments them afterwards. Given that open-ended object category
predictions cannot be directly compared with a fixed ground truth, we develop a
Large Language Model-based Auto-Vocabulary Evaluator (LAVE) to efficiently
evaluate the automatically generated classes and their corresponding segments.
With AVS, our method sets new benchmarks on datasets PASCAL VOC, Context,
ADE20K, and Cityscapes, while showing competitive performance to OVS methods
that require specified class names.
| [
{
"version": "v1",
"created": "Thu, 7 Dec 2023 18:55:52 GMT"
},
{
"version": "v2",
"created": "Wed, 20 Mar 2024 16:11:22 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 12:39:35 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Ülger",
"Osman",
""
],
[
"Kulicki",
"Maksymilian",
""
],
[
"Asano",
"Yuki",
""
],
[
"Oswald",
"Martin R.",
""
]
]
| TITLE: Auto-Vocabulary Semantic Segmentation
ABSTRACT: Open-Vocabulary Segmentation (OVS) methods are capable of performing semantic
segmentation without relying on a fixed vocabulary, and in some cases, without
training or fine-tuning. However, OVS methods typically require a human in the
loop to specify the vocabulary based on the task or dataset at hand. In this
paper, we introduce Auto-Vocabulary Semantic Segmentation (AVS), advancing
open-ended image understanding by eliminating the necessity to predefine object
categories for segmentation. Our approach, AutoSeg, presents a framework that
autonomously identifies relevant class names using semantically enhanced BLIP
embeddings and segments them afterwards. Given that open-ended object category
predictions cannot be directly compared with a fixed ground truth, we develop a
Large Language Model-based Auto-Vocabulary Evaluator (LAVE) to efficiently
evaluate the automatically generated classes and their corresponding segments.
With AVS, our method sets new benchmarks on datasets PASCAL VOC, Context,
ADE20K, and Cityscapes, while showing competitive performance to OVS methods
that require specified class names.
| no_new_dataset | 0.949153 |
2402.03166 | Jos\'e Morano | Jos\'e Morano and Guilherme Aresta and Hrvoje Bogunovi\'c | RRWNet: Recursive Refinement Network for effective retinal artery/vein
segmentation and classification | null | Expert Systems with Applications, 2024 | 10.1016/j.eswa.2024.124970 | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The caliber and configuration of retinal blood vessels serve as important
biomarkers for various diseases and medical conditions. A thorough analysis of
the retinal vasculature requires the segmentation of the blood vessels and
their classification into arteries and veins, typically performed on color
fundus images obtained by retinography. However, manually performing these
tasks is labor-intensive and prone to human error. While several automated
methods have been proposed to address this task, the current state of art faces
challenges due to manifest classification errors affecting the topological
consistency of segmentation maps. In this work, we introduce RRWNet, a novel
end-to-end deep learning framework that addresses this limitation. The
framework consists of a fully convolutional neural network that recursively
refines semantic segmentation maps, correcting manifest classification errors
and thus improving topological consistency. In particular, RRWNet is composed
of two specialized subnetworks: a Base subnetwork that generates base
segmentation maps from the input images, and a Recursive Refinement subnetwork
that iteratively and recursively improves these maps. Evaluation on three
different public datasets demonstrates the state-of-the-art performance of the
proposed method, yielding more topologically consistent segmentation maps with
fewer manifest classification errors than existing approaches. In addition, the
Recursive Refinement module within RRWNet proves effective in post-processing
segmentation maps from other methods, further demonstrating its potential. The
model code, weights, and predictions will be publicly available at
https://github.com/j-morano/rrwnet.
| [
{
"version": "v1",
"created": "Mon, 5 Feb 2024 16:35:29 GMT"
},
{
"version": "v2",
"created": "Wed, 13 Mar 2024 12:52:26 GMT"
},
{
"version": "v3",
"created": "Wed, 3 Apr 2024 07:10:22 GMT"
},
{
"version": "v4",
"created": "Thu, 8 Aug 2024 13:32:21 GMT"
},
{
"version": "v5",
"created": "Wed, 12 Mar 2025 17:04:36 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Morano",
"José",
""
],
[
"Aresta",
"Guilherme",
""
],
[
"Bogunović",
"Hrvoje",
""
]
]
| TITLE: RRWNet: Recursive Refinement Network for effective retinal artery/vein
segmentation and classification
ABSTRACT: The caliber and configuration of retinal blood vessels serve as important
biomarkers for various diseases and medical conditions. A thorough analysis of
the retinal vasculature requires the segmentation of the blood vessels and
their classification into arteries and veins, typically performed on color
fundus images obtained by retinography. However, manually performing these
tasks is labor-intensive and prone to human error. While several automated
methods have been proposed to address this task, the current state of art faces
challenges due to manifest classification errors affecting the topological
consistency of segmentation maps. In this work, we introduce RRWNet, a novel
end-to-end deep learning framework that addresses this limitation. The
framework consists of a fully convolutional neural network that recursively
refines semantic segmentation maps, correcting manifest classification errors
and thus improving topological consistency. In particular, RRWNet is composed
of two specialized subnetworks: a Base subnetwork that generates base
segmentation maps from the input images, and a Recursive Refinement subnetwork
that iteratively and recursively improves these maps. Evaluation on three
different public datasets demonstrates the state-of-the-art performance of the
proposed method, yielding more topologically consistent segmentation maps with
fewer manifest classification errors than existing approaches. In addition, the
Recursive Refinement module within RRWNet proves effective in post-processing
segmentation maps from other methods, further demonstrating its potential. The
model code, weights, and predictions will be publicly available at
https://github.com/j-morano/rrwnet.
| no_new_dataset | 0.949902 |
2402.03848 | David Peer | David Peer, Philemon Sch\"opf, Volckmar Nebendahl, Alexander Rietzler,
Sebastian Stabinger | ANLS* -- A Universal Document Processing Metric for Generative Large
Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Traditionally, discriminative models have been the predominant choice for
tasks like document classification and information extraction. These models
make predictions that fall into a limited number of predefined classes,
facilitating a binary true or false evaluation and enabling the direct
calculation of metrics such as the F1 score. However, recent advancements in
generative large language models (GLLMs) have prompted a shift in the field due
to their enhanced zero-shot capabilities, which eliminate the need for a
downstream dataset and computationally expensive fine-tuning. However,
evaluating GLLMs presents a challenge as the binary true or false evaluation
used for discriminative models is not applicable to the predictions made by
GLLMs.
This paper introduces a new metric for generative models called ANLS* for
evaluating a wide variety of tasks, including information extraction and
classification tasks. The ANLS* metric extends existing ANLS metrics as a
drop-in-replacement and is still compatible with previously reported ANLS
scores. An evaluation of 7 different datasets, and more than 20 different GLLMs
together with 3 different prompting methods using the ANLS* metric is also
provided, demonstrating the importance of the proposed metric.
We also benchmark a novel approach to generate prompts for documents, called
SFT, against other prompting techniques such as LATIN. In almost all cases, SFT
outperforms other techniques and improves the state-of-the-art, sometimes by as
much as $10$ percentage points.
Sources are available at https://github.com/deepopinion/anls_star_metric
| [
{
"version": "v1",
"created": "Tue, 6 Feb 2024 09:50:08 GMT"
},
{
"version": "v2",
"created": "Tue, 27 Feb 2024 13:14:28 GMT"
},
{
"version": "v3",
"created": "Thu, 21 Mar 2024 05:58:10 GMT"
},
{
"version": "v4",
"created": "Tue, 16 Apr 2024 09:14:46 GMT"
},
{
"version": "v5",
"created": "Sat, 25 May 2024 06:31:45 GMT"
},
{
"version": "v6",
"created": "Fri, 28 Jun 2024 06:49:39 GMT"
},
{
"version": "v7",
"created": "Tue, 27 Aug 2024 08:33:29 GMT"
},
{
"version": "v8",
"created": "Mon, 3 Mar 2025 12:50:31 GMT"
},
{
"version": "v9",
"created": "Wed, 12 Mar 2025 08:02:54 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Peer",
"David",
""
],
[
"Schöpf",
"Philemon",
""
],
[
"Nebendahl",
"Volckmar",
""
],
[
"Rietzler",
"Alexander",
""
],
[
"Stabinger",
"Sebastian",
""
]
]
| TITLE: ANLS* -- A Universal Document Processing Metric for Generative Large
Language Models
ABSTRACT: Traditionally, discriminative models have been the predominant choice for
tasks like document classification and information extraction. These models
make predictions that fall into a limited number of predefined classes,
facilitating a binary true or false evaluation and enabling the direct
calculation of metrics such as the F1 score. However, recent advancements in
generative large language models (GLLMs) have prompted a shift in the field due
to their enhanced zero-shot capabilities, which eliminate the need for a
downstream dataset and computationally expensive fine-tuning. However,
evaluating GLLMs presents a challenge as the binary true or false evaluation
used for discriminative models is not applicable to the predictions made by
GLLMs.
This paper introduces a new metric for generative models called ANLS* for
evaluating a wide variety of tasks, including information extraction and
classification tasks. The ANLS* metric extends existing ANLS metrics as a
drop-in-replacement and is still compatible with previously reported ANLS
scores. An evaluation of 7 different datasets, and more than 20 different GLLMs
together with 3 different prompting methods using the ANLS* metric is also
provided, demonstrating the importance of the proposed metric.
We also benchmark a novel approach to generate prompts for documents, called
SFT, against other prompting techniques such as LATIN. In almost all cases, SFT
outperforms other techniques and improves the state-of-the-art, sometimes by as
much as $10$ percentage points.
Sources are available at https://github.com/deepopinion/anls_star_metric
| no_new_dataset | 0.948489 |
2402.15131 | Guanming Xiong | Guanming Xiong, Junwei Bao, Wen Zhao | Interactive-KBQA: Multi-Turn Interactions for Knowledge Base Question
Answering with Large Language Models | This work has been accepted by the ACL 2024 main conference. Code and
data are available at: https://github.com/JimXiongGM/Interactive-KBQA | null | 10.18653/v1/2024.acl-long.569 | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | This study explores the realm of knowledge base question answering (KBQA).
KBQA is considered a challenging task, particularly in parsing intricate
questions into executable logical forms. Traditional semantic parsing
(SP)-based methods require extensive data annotations, which result in
significant costs. Recently, the advent of few-shot in-context learning,
powered by large language models (LLMs), has showcased promising capabilities.
However, fully leveraging LLMs to parse questions into logical forms in
low-resource scenarios poses a substantial challenge. To tackle these hurdles,
we introduce Interactive-KBQA, a framework designed to generate logical forms
through direct interaction with knowledge bases (KBs). Within this framework,
we have developed three generic APIs for KB interaction. For each category of
complex question, we devised exemplars to guide LLMs through the reasoning
processes. Our method achieves competitive results on the WebQuestionsSP,
ComplexWebQuestions, KQA Pro, and MetaQA datasets with a minimal number of
examples (shots). Importantly, our approach supports manual intervention,
allowing for the iterative refinement of LLM outputs. By annotating a dataset
with step-wise reasoning processes, we showcase our model's adaptability and
highlight its potential for contributing significant enhancements to the field.
| [
{
"version": "v1",
"created": "Fri, 23 Feb 2024 06:32:18 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Jul 2024 06:14:20 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 06:15:34 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Xiong",
"Guanming",
""
],
[
"Bao",
"Junwei",
""
],
[
"Zhao",
"Wen",
""
]
]
| TITLE: Interactive-KBQA: Multi-Turn Interactions for Knowledge Base Question
Answering with Large Language Models
ABSTRACT: This study explores the realm of knowledge base question answering (KBQA).
KBQA is considered a challenging task, particularly in parsing intricate
questions into executable logical forms. Traditional semantic parsing
(SP)-based methods require extensive data annotations, which result in
significant costs. Recently, the advent of few-shot in-context learning,
powered by large language models (LLMs), has showcased promising capabilities.
However, fully leveraging LLMs to parse questions into logical forms in
low-resource scenarios poses a substantial challenge. To tackle these hurdles,
we introduce Interactive-KBQA, a framework designed to generate logical forms
through direct interaction with knowledge bases (KBs). Within this framework,
we have developed three generic APIs for KB interaction. For each category of
complex question, we devised exemplars to guide LLMs through the reasoning
processes. Our method achieves competitive results on the WebQuestionsSP,
ComplexWebQuestions, KQA Pro, and MetaQA datasets with a minimal number of
examples (shots). Importantly, our approach supports manual intervention,
allowing for the iterative refinement of LLM outputs. By annotating a dataset
with step-wise reasoning processes, we showcase our model's adaptability and
highlight its potential for contributing significant enhancements to the field.
| no_new_dataset | 0.941007 |
2402.16424 | Qingqing Long | Yuqi Li, Qingqing Long, Yihang Zhou, Meng Xiao, Ran Zhang, Zhiyuan
Ning, Zhihong Zhu, Xuezhi Wang, Yuanchun Zhou | COMAE: COMprehensive Attribute Exploration for Zero-shot Hashing | 18 pages, 7 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Zero-shot hashing (ZSH) has shown excellent success owing to its efficiency
and generalization in large-scale retrieval scenarios. While considerable
success has been achieved, there still exist urgent limitations. Existing works
ignore the locality relationships of representations and attributes, which have
effective transferability between seeable classes and unseeable classes. Also,
the continuous-value attributes are not fully harnessed. In response, we
conduct a COMprehensive Attribute Exploration for ZSH, named COMAE, which
depicts the relationships from seen classes to unseen ones through three
meticulously designed explorations, i.e., point-wise, pair-wise and class-wise
consistency constraints. By regressing attributes from the proposed attribute
prototype network, COMAE learns the local features that are relevant to the
visual attributes. Then COMAE utilizes contrastive learning to comprehensively
depict the context of attributes, rather than instance-independent
optimization. Finally, the class-wise constraint is designed to cohesively
learn the hash code, image representation, and visual attributes more
effectively. Experimental results on the popular ZSH datasets demonstrate that
COMAE outperforms state-of-the-art hashing techniques, especially in scenarios
with a larger number of unseen label classes.
| [
{
"version": "v1",
"created": "Mon, 26 Feb 2024 09:22:57 GMT"
},
{
"version": "v2",
"created": "Wed, 17 Jul 2024 08:23:33 GMT"
},
{
"version": "v3",
"created": "Sun, 21 Jul 2024 12:37:41 GMT"
},
{
"version": "v4",
"created": "Wed, 12 Mar 2025 14:29:30 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Li",
"Yuqi",
""
],
[
"Long",
"Qingqing",
""
],
[
"Zhou",
"Yihang",
""
],
[
"Xiao",
"Meng",
""
],
[
"Zhang",
"Ran",
""
],
[
"Ning",
"Zhiyuan",
""
],
[
"Zhu",
"Zhihong",
""
],
[
"Wang",
"Xuezhi",
""
],
[
"Zhou",
"Yuanchun",
""
]
]
| TITLE: COMAE: COMprehensive Attribute Exploration for Zero-shot Hashing
ABSTRACT: Zero-shot hashing (ZSH) has shown excellent success owing to its efficiency
and generalization in large-scale retrieval scenarios. While considerable
success has been achieved, there still exist urgent limitations. Existing works
ignore the locality relationships of representations and attributes, which have
effective transferability between seeable classes and unseeable classes. Also,
the continuous-value attributes are not fully harnessed. In response, we
conduct a COMprehensive Attribute Exploration for ZSH, named COMAE, which
depicts the relationships from seen classes to unseen ones through three
meticulously designed explorations, i.e., point-wise, pair-wise and class-wise
consistency constraints. By regressing attributes from the proposed attribute
prototype network, COMAE learns the local features that are relevant to the
visual attributes. Then COMAE utilizes contrastive learning to comprehensively
depict the context of attributes, rather than instance-independent
optimization. Finally, the class-wise constraint is designed to cohesively
learn the hash code, image representation, and visual attributes more
effectively. Experimental results on the popular ZSH datasets demonstrate that
COMAE outperforms state-of-the-art hashing techniques, especially in scenarios
with a larger number of unseen label classes.
| no_new_dataset | 0.945197 |
2403.06681 | Jintao Huang | Jintao Huang, Yiu-Ming Cheung, and Chi-Man Vong | PLOOD: Partial Label Learning with Out-of-distribution Objects | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing Partial Label Learning (PLL) methods posit that training and test
data adhere to the same distribution, a premise that frequently does not hold
in practical application where Out-of-Distribution (OOD) objects are present.
We introduce the OODPLL paradigm to tackle this significant yet underexplored
issue. And our newly proposed PLOOD framework enables PLL to tackle OOD objects
through Positive-Negative Sample Augmented (PNSA) feature learning and Partial
Energy (PE)-based label refinement. The PNSA module enhances feature
discrimination and OOD recognition by simulating in- and out-of-distribution
instances, which employ structured positive and negative sample augmentation,
in contrast to conventional PLL methods struggling to distinguish OOD samples.
The PE scoring mechanism combines label confidence with energy-based
uncertainty estimation, thereby reducing the impact of imprecise supervision
and effectively achieving label disambiguation. Experimental results on
CIFAR-10 and CIFAR-100, alongside various OOD datasets, demonstrate that
conventional PLL methods exhibit substantial degradation in OOD scenarios,
underscoring the necessity of incorporating OOD considerations in PLL
approaches. Ablation studies show that PNSA feature learning and PE-based label
refinement are necessary for PLOOD to work, offering a robust solution for
open-set PLL problems.
| [
{
"version": "v1",
"created": "Mon, 11 Mar 2024 12:56:36 GMT"
},
{
"version": "v2",
"created": "Thu, 30 May 2024 11:16:49 GMT"
},
{
"version": "v3",
"created": "Sat, 1 Jun 2024 05:19:24 GMT"
},
{
"version": "v4",
"created": "Wed, 12 Mar 2025 05:54:38 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Huang",
"Jintao",
""
],
[
"Cheung",
"Yiu-Ming",
""
],
[
"Vong",
"Chi-Man",
""
]
]
| TITLE: PLOOD: Partial Label Learning with Out-of-distribution Objects
ABSTRACT: Existing Partial Label Learning (PLL) methods posit that training and test
data adhere to the same distribution, a premise that frequently does not hold
in practical application where Out-of-Distribution (OOD) objects are present.
We introduce the OODPLL paradigm to tackle this significant yet underexplored
issue. And our newly proposed PLOOD framework enables PLL to tackle OOD objects
through Positive-Negative Sample Augmented (PNSA) feature learning and Partial
Energy (PE)-based label refinement. The PNSA module enhances feature
discrimination and OOD recognition by simulating in- and out-of-distribution
instances, which employ structured positive and negative sample augmentation,
in contrast to conventional PLL methods struggling to distinguish OOD samples.
The PE scoring mechanism combines label confidence with energy-based
uncertainty estimation, thereby reducing the impact of imprecise supervision
and effectively achieving label disambiguation. Experimental results on
CIFAR-10 and CIFAR-100, alongside various OOD datasets, demonstrate that
conventional PLL methods exhibit substantial degradation in OOD scenarios,
underscoring the necessity of incorporating OOD considerations in PLL
approaches. Ablation studies show that PNSA feature learning and PE-based label
refinement are necessary for PLOOD to work, offering a robust solution for
open-set PLL problems.
| no_new_dataset | 0.946151 |
2403.20312 | Jaisidh Singh | Jaisidh Singh, Ishaan Shrivastava, Mayank Vatsa, Richa Singh, Aparna
Bharati | Learn "No" to Say "Yes" Better: Improving Vision-Language Models via
Negations | 14 pages + 6 figures in main manuscript (excluding references) | WACV 2025 pages(7991-8001) | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing vision-language models (VLMs) treat text descriptions as a unit,
confusing individual concepts in a prompt and impairing visual semantic
matching and reasoning. An important aspect of reasoning in logic and language
is negations. This paper highlights the limitations of popular VLMs such as
CLIP, at understanding the implications of negations, i.e., the effect of the
word "not" in a given prompt. To enable evaluation of VLMs on fluent prompts
with negations, we present CC-Neg, a dataset containing 228,246 images, true
captions and their corresponding negated captions. Using CC-Neg along with
modifications to the contrastive loss of CLIP, our proposed CoN-CLIP framework,
has an improved understanding of negations. This training paradigm improves
CoN-CLIP's ability to encode semantics reliably, resulting in 3.85% average
gain in top-1 accuracy for zero-shot image classification across 8 datasets.
Further, CoN-CLIP outperforms CLIP on challenging compositionality benchmarks
such as SugarCREPE by 4.4%, showcasing emergent compositional understanding of
objects, relations, and attributes in text. Overall, our work addresses a
crucial limitation of VLMs by introducing a dataset and framework that
strengthens semantic associations between images and text, demonstrating
improved large-scale foundation models with significantly reduced computational
cost, promoting efficiency and accessibility.
| [
{
"version": "v1",
"created": "Fri, 29 Mar 2024 17:33:42 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Singh",
"Jaisidh",
""
],
[
"Shrivastava",
"Ishaan",
""
],
[
"Vatsa",
"Mayank",
""
],
[
"Singh",
"Richa",
""
],
[
"Bharati",
"Aparna",
""
]
]
| TITLE: Learn "No" to Say "Yes" Better: Improving Vision-Language Models via
Negations
ABSTRACT: Existing vision-language models (VLMs) treat text descriptions as a unit,
confusing individual concepts in a prompt and impairing visual semantic
matching and reasoning. An important aspect of reasoning in logic and language
is negations. This paper highlights the limitations of popular VLMs such as
CLIP, at understanding the implications of negations, i.e., the effect of the
word "not" in a given prompt. To enable evaluation of VLMs on fluent prompts
with negations, we present CC-Neg, a dataset containing 228,246 images, true
captions and their corresponding negated captions. Using CC-Neg along with
modifications to the contrastive loss of CLIP, our proposed CoN-CLIP framework,
has an improved understanding of negations. This training paradigm improves
CoN-CLIP's ability to encode semantics reliably, resulting in 3.85% average
gain in top-1 accuracy for zero-shot image classification across 8 datasets.
Further, CoN-CLIP outperforms CLIP on challenging compositionality benchmarks
such as SugarCREPE by 4.4%, showcasing emergent compositional understanding of
objects, relations, and attributes in text. Overall, our work addresses a
crucial limitation of VLMs by introducing a dataset and framework that
strengthens semantic associations between images and text, demonstrating
improved large-scale foundation models with significantly reduced computational
cost, promoting efficiency and accessibility.
| new_dataset | 0.960584 |
2404.11100 | Qiyu Hou | Qiyu Hou, Jun Wang, Meixuan Qiao, Lujun Tian | Synthesizing Realistic Data for Table Recognition | ICDAR 2024 | null | 10.1007/978-3-031-70533-5_22 | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To overcome the limitations and challenges of current automatic table data
annotation methods and random table data synthesis approaches, we propose a
novel method for synthesizing annotation data specifically designed for table
recognition. This method utilizes the structure and content of existing complex
tables, facilitating the efficient creation of tables that closely replicate
the authentic styles found in the target domain. By leveraging the actual
structure and content of tables from Chinese financial announcements, we have
developed the first extensive table annotation dataset in this domain. We used
this dataset to train several recent deep learning-based end-to-end table
recognition models. Additionally, we have established the inaugural benchmark
for real-world complex tables in the Chinese financial announcement domain,
using it to assess the performance of models trained on our synthetic data,
thereby effectively validating our method's practicality and effectiveness.
Furthermore, we applied our synthesis method to augment the FinTabNet dataset,
extracted from English financial announcements, by increasing the proportion of
tables with multiple spanning cells to introduce greater complexity. Our
experiments show that models trained on this augmented dataset achieve
comprehensive improvements in performance, especially in the recognition of
tables with multiple spanning cells.
| [
{
"version": "v1",
"created": "Wed, 17 Apr 2024 06:36:17 GMT"
},
{
"version": "v2",
"created": "Tue, 9 Jul 2024 12:09:32 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Hou",
"Qiyu",
""
],
[
"Wang",
"Jun",
""
],
[
"Qiao",
"Meixuan",
""
],
[
"Tian",
"Lujun",
""
]
]
| TITLE: Synthesizing Realistic Data for Table Recognition
ABSTRACT: To overcome the limitations and challenges of current automatic table data
annotation methods and random table data synthesis approaches, we propose a
novel method for synthesizing annotation data specifically designed for table
recognition. This method utilizes the structure and content of existing complex
tables, facilitating the efficient creation of tables that closely replicate
the authentic styles found in the target domain. By leveraging the actual
structure and content of tables from Chinese financial announcements, we have
developed the first extensive table annotation dataset in this domain. We used
this dataset to train several recent deep learning-based end-to-end table
recognition models. Additionally, we have established the inaugural benchmark
for real-world complex tables in the Chinese financial announcement domain,
using it to assess the performance of models trained on our synthetic data,
thereby effectively validating our method's practicality and effectiveness.
Furthermore, we applied our synthesis method to augment the FinTabNet dataset,
extracted from English financial announcements, by increasing the proportion of
tables with multiple spanning cells to introduce greater complexity. Our
experiments show that models trained on this augmented dataset achieve
comprehensive improvements in performance, especially in the recognition of
tables with multiple spanning cells.
| new_dataset | 0.963575 |
2404.11465 | Arvindh Arun | Arvindh Arun, Saurav Chhatani, Jisun An, Ponnurangam Kumaraguru | X-posing Free Speech: Examining the Impact of Moderation Relaxation on
Online Social Networks | null | null | 10.18653/v1/2024.woah-1.15 | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | We investigate the impact of free speech and the relaxation of moderation on
online social media platforms using Elon Musk's takeover of Twitter as a case
study. By curating a dataset of over 10 million tweets, our study employs a
novel framework combining content and network analysis. Our findings reveal a
significant increase in the distribution of certain forms of hate content,
particularly targeting the LGBTQ+ community and liberals. Network analysis
reveals the formation of cohesive hate communities facilitated by influential
bridge users, with substantial growth in interactions hinting at increased hate
production and diffusion. By tracking the temporal evolution of PageRank, we
identify key influencers, primarily self-identified far-right supporters
disseminating hate against liberals and woke culture. Ironically, embracing
free speech principles appears to have enabled hate speech against the very
concept of freedom of expression and free speech itself. Our findings
underscore the delicate balance platforms must strike between open expression
and robust moderation to curb the proliferation of hate online.
| [
{
"version": "v1",
"created": "Wed, 17 Apr 2024 15:15:47 GMT"
},
{
"version": "v2",
"created": "Thu, 23 May 2024 08:05:39 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Arun",
"Arvindh",
""
],
[
"Chhatani",
"Saurav",
""
],
[
"An",
"Jisun",
""
],
[
"Kumaraguru",
"Ponnurangam",
""
]
]
| TITLE: X-posing Free Speech: Examining the Impact of Moderation Relaxation on
Online Social Networks
ABSTRACT: We investigate the impact of free speech and the relaxation of moderation on
online social media platforms using Elon Musk's takeover of Twitter as a case
study. By curating a dataset of over 10 million tweets, our study employs a
novel framework combining content and network analysis. Our findings reveal a
significant increase in the distribution of certain forms of hate content,
particularly targeting the LGBTQ+ community and liberals. Network analysis
reveals the formation of cohesive hate communities facilitated by influential
bridge users, with substantial growth in interactions hinting at increased hate
production and diffusion. By tracking the temporal evolution of PageRank, we
identify key influencers, primarily self-identified far-right supporters
disseminating hate against liberals and woke culture. Ironically, embracing
free speech principles appears to have enabled hate speech against the very
concept of freedom of expression and free speech itself. Our findings
underscore the delicate balance platforms must strike between open expression
and robust moderation to curb the proliferation of hate online.
| new_dataset | 0.964456 |
2405.08971 | Marvin Pf\"ortner | Marvin Pf\"ortner, Jonathan Wenger, Jon Cockayne, Philipp Hennig | Computation-Aware Kalman Filtering and Smoothing | null | null | null | null | cs.LG cs.NA math.NA stat.ML | http://creativecommons.org/licenses/by/4.0/ | Kalman filtering and smoothing are the foundational mechanisms for efficient
inference in Gauss-Markov models. However, their time and memory complexities
scale prohibitively with the size of the state space. This is particularly
problematic in spatiotemporal regression problems, where the state dimension
scales with the number of spatial observations. Existing approximate frameworks
leverage low-rank approximations of the covariance matrix. But since they do
not model the error introduced by the computational approximation, their
predictive uncertainty estimates can be overly optimistic. In this work, we
propose a probabilistic numerical method for inference in high-dimensional
Gauss-Markov models which mitigates these scaling issues. Our matrix-free
iterative algorithm leverages GPU acceleration and crucially enables a tunable
trade-off between computational cost and predictive uncertainty. Finally, we
demonstrate the scalability of our method on a large-scale climate dataset.
| [
{
"version": "v1",
"created": "Tue, 14 May 2024 21:31:11 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 15:51:20 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Pförtner",
"Marvin",
""
],
[
"Wenger",
"Jonathan",
""
],
[
"Cockayne",
"Jon",
""
],
[
"Hennig",
"Philipp",
""
]
]
| TITLE: Computation-Aware Kalman Filtering and Smoothing
ABSTRACT: Kalman filtering and smoothing are the foundational mechanisms for efficient
inference in Gauss-Markov models. However, their time and memory complexities
scale prohibitively with the size of the state space. This is particularly
problematic in spatiotemporal regression problems, where the state dimension
scales with the number of spatial observations. Existing approximate frameworks
leverage low-rank approximations of the covariance matrix. But since they do
not model the error introduced by the computational approximation, their
predictive uncertainty estimates can be overly optimistic. In this work, we
propose a probabilistic numerical method for inference in high-dimensional
Gauss-Markov models which mitigates these scaling issues. Our matrix-free
iterative algorithm leverages GPU acceleration and crucially enables a tunable
trade-off between computational cost and predictive uncertainty. Finally, we
demonstrate the scalability of our method on a large-scale climate dataset.
| no_new_dataset | 0.944842 |
2405.18458 | Yizhi Wang | Yizhi Wang, Minjia Chen, Chunhui Yao, Jie Ma, Ting Yan, Richard Penty,
Qixiang Cheng | Asymmetrical estimator for training encapsulated deep photonic neural
networks | 23 pages, 6 figures | Nat Commun 16, 2143 (2025) | 10.1038/s41467-025-57459-5 | null | cs.LG physics.optics | http://creativecommons.org/licenses/by/4.0/ | Photonic neural networks (PNNs) are fast in-propagation and high bandwidth
paradigms that aim to popularize reproducible NN acceleration with higher
efficiency and lower cost. However, the training of PNN is known to be
challenging, where the device-to-device and system-to-system variations create
imperfect knowledge of the PNN. Despite backpropagation (BP)-based training
algorithms being the industry standard for their robustness, generality, and
fast gradient convergence for digital training, existing PNN-BP methods rely
heavily on accurate intermediate state extraction or extensive computational
resources for deep PNNs (DPNNs). The truncated photonic signal propagation and
the computation overhead bottleneck DPNN's operation efficiency and increase
system construction cost. Here, we introduce the asymmetrical training (AsyT)
method, tailored for encapsulated DPNNs, where the signal is preserved in the
analogue photonic domain for the entire structure. AsyT offers a lightweight
solution for DPNNs with minimum readouts, fast and energy-efficient operation,
and minimum system footprint. AsyT's ease of operation, error tolerance, and
generality aim to promote PNN acceleration in a widened operational scenario
despite the fabrication variations and imperfect controls. We demonstrated AsyT
for encapsulated DPNN with integrated photonic chips, repeatably enhancing the
performance from in-silico BP for different network structures and datasets.
| [
{
"version": "v1",
"created": "Tue, 28 May 2024 17:27:20 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Aug 2024 10:58:17 GMT"
},
{
"version": "v3",
"created": "Sun, 17 Nov 2024 12:33:25 GMT"
},
{
"version": "v4",
"created": "Thu, 13 Feb 2025 11:59:20 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Wang",
"Yizhi",
""
],
[
"Chen",
"Minjia",
""
],
[
"Yao",
"Chunhui",
""
],
[
"Ma",
"Jie",
""
],
[
"Yan",
"Ting",
""
],
[
"Penty",
"Richard",
""
],
[
"Cheng",
"Qixiang",
""
]
]
| TITLE: Asymmetrical estimator for training encapsulated deep photonic neural
networks
ABSTRACT: Photonic neural networks (PNNs) are fast in-propagation and high bandwidth
paradigms that aim to popularize reproducible NN acceleration with higher
efficiency and lower cost. However, the training of PNN is known to be
challenging, where the device-to-device and system-to-system variations create
imperfect knowledge of the PNN. Despite backpropagation (BP)-based training
algorithms being the industry standard for their robustness, generality, and
fast gradient convergence for digital training, existing PNN-BP methods rely
heavily on accurate intermediate state extraction or extensive computational
resources for deep PNNs (DPNNs). The truncated photonic signal propagation and
the computation overhead bottleneck DPNN's operation efficiency and increase
system construction cost. Here, we introduce the asymmetrical training (AsyT)
method, tailored for encapsulated DPNNs, where the signal is preserved in the
analogue photonic domain for the entire structure. AsyT offers a lightweight
solution for DPNNs with minimum readouts, fast and energy-efficient operation,
and minimum system footprint. AsyT's ease of operation, error tolerance, and
generality aim to promote PNN acceleration in a widened operational scenario
despite the fabrication variations and imperfect controls. We demonstrated AsyT
for encapsulated DPNN with integrated photonic chips, repeatably enhancing the
performance from in-silico BP for different network structures and datasets.
| no_new_dataset | 0.951006 |
2405.20903 | Valeria Mascolo | Valeria Mascolo, Alessandro Lovo, Corentin Herbert, Freddy Bouchet | Gaussian Framework and Optimal Projection of Weather Fields for
Prediction of Extreme Events | 40 pages, 11 figures, 6 tables | null | null | null | physics.ao-ph physics.data-an | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Extreme events are the major weather-related hazard for humanity. It is then
of crucial importance to have a good understanding of their statistics and to
be able to forecast them. However, lack of sufficient data makes their study
particularly challenging. In this work, we provide a simple framework for
studying extreme events that tackles the lack of data issue by using the entire
available dataset, rather than focusing on the extremes of the dataset. To do
so, we make the assumption that the set of predictors and the observable used
to define the extreme event follow a jointly Gaussian distribution. This
naturally gives the notion of an optimal projection of the predictors for
forecasting the event. We take as a case study extreme heatwaves over France,
and we test our method on an 8000-year-long intermediate complexity climate
model time series and on the ERA5 reanalysis dataset. For a-posteriori
statistics, we observe and motivate the fact that composite maps of very
extreme events look similar to less extreme ones. For prediction, we show that
our method is competitive with off-the-shelf neural networks on the long
dataset and outperforms them on reanalysis. The optimal projection pattern,
which makes our forecast intrinsically interpretable, highlights the importance
of soil moisture deficit and quasi-stationary Rossby waves as precursors to
extreme heatwaves.
| [
{
"version": "v1",
"created": "Fri, 31 May 2024 15:15:29 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Jun 2024 10:42:48 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 18:18:02 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Mascolo",
"Valeria",
""
],
[
"Lovo",
"Alessandro",
""
],
[
"Herbert",
"Corentin",
""
],
[
"Bouchet",
"Freddy",
""
]
]
| TITLE: Gaussian Framework and Optimal Projection of Weather Fields for
Prediction of Extreme Events
ABSTRACT: Extreme events are the major weather-related hazard for humanity. It is then
of crucial importance to have a good understanding of their statistics and to
be able to forecast them. However, lack of sufficient data makes their study
particularly challenging. In this work, we provide a simple framework for
studying extreme events that tackles the lack of data issue by using the entire
available dataset, rather than focusing on the extremes of the dataset. To do
so, we make the assumption that the set of predictors and the observable used
to define the extreme event follow a jointly Gaussian distribution. This
naturally gives the notion of an optimal projection of the predictors for
forecasting the event. We take as a case study extreme heatwaves over France,
and we test our method on an 8000-year-long intermediate complexity climate
model time series and on the ERA5 reanalysis dataset. For a-posteriori
statistics, we observe and motivate the fact that composite maps of very
extreme events look similar to less extreme ones. For prediction, we show that
our method is competitive with off-the-shelf neural networks on the long
dataset and outperforms them on reanalysis. The optimal projection pattern,
which makes our forecast intrinsically interpretable, highlights the importance
of soil moisture deficit and quasi-stationary Rossby waves as precursors to
extreme heatwaves.
| no_new_dataset | 0.947332 |
2406.08788 | Jay Revolinsky | Jay Revolinsky, Harry Shomer, Jiliang Tang | Towards Understanding Link Predictor Generalizability Under Distribution
Shifts | 23 pages, 8 figures, 17 tables | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | State-of-the-art link prediction (LP) models demonstrate impressive benchmark
results. However, popular benchmark datasets often assume that training,
validation, and testing samples are representative of the overall dataset
distribution. In real-world situations, this assumption is often incorrect;
uncontrolled factors lead new dataset samples to come from a different
distribution than training samples. Additionally, the majority of recent work
with graph dataset shift focuses on node- and graph-level tasks, largely
ignoring link-level tasks. To bridge this gap, we introduce a novel splitting
strategy, known as LPShift, which utilizes structural properties to induce a
controlled distribution shift. We verify LPShift's effect through empirical
evaluation of SOTA LP models on 16 LPShift variants of original dataset splits,
with results indicating drastic changes to model performance. Additional
experiments demonstrate graph structure has a strong influence on the success
of current generalization methods. Source Code Available Here:
https://github.com/revolins/LPShift
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 03:47:12 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 19:49:55 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Revolinsky",
"Jay",
""
],
[
"Shomer",
"Harry",
""
],
[
"Tang",
"Jiliang",
""
]
]
| TITLE: Towards Understanding Link Predictor Generalizability Under Distribution
Shifts
ABSTRACT: State-of-the-art link prediction (LP) models demonstrate impressive benchmark
results. However, popular benchmark datasets often assume that training,
validation, and testing samples are representative of the overall dataset
distribution. In real-world situations, this assumption is often incorrect;
uncontrolled factors lead new dataset samples to come from a different
distribution than training samples. Additionally, the majority of recent work
with graph dataset shift focuses on node- and graph-level tasks, largely
ignoring link-level tasks. To bridge this gap, we introduce a novel splitting
strategy, known as LPShift, which utilizes structural properties to induce a
controlled distribution shift. We verify LPShift's effect through empirical
evaluation of SOTA LP models on 16 LPShift variants of original dataset splits,
with results indicating drastic changes to model performance. Additional
experiments demonstrate graph structure has a strong influence on the success
of current generalization methods. Source Code Available Here:
https://github.com/revolins/LPShift
| no_new_dataset | 0.946695 |
2406.09836 | Zhiwei Zhang | Zhiwei Zhang, Minhua Lin, Junjie Xu, Zongyu Wu, Enyan Dai, Suhang Wang | Robustness Inspired Graph Backdoor Defense | Accepted by ICLR 2025 (Oral) | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph Neural Networks (GNNs) have achieved promising results in tasks such as
node classification and graph classification. However, recent studies reveal
that GNNs are vulnerable to backdoor attacks, posing a significant threat to
their real-world adoption. Despite initial efforts to defend against specific
graph backdoor attacks, there is no work on defending against various types of
backdoor attacks where generated triggers have different properties. Hence, we
first empirically verify that prediction variance under edge dropping is a
crucial indicator for identifying poisoned nodes. With this observation, we
propose using random edge dropping to detect backdoors and theoretically show
that it can efficiently distinguish poisoned nodes from clean ones.
Furthermore, we introduce a novel robust training strategy to efficiently
counteract the impact of the triggers. Extensive experiments on real-world
datasets show that our framework can effectively identify poisoned nodes,
significantly degrade the attack success rate, and maintain clean accuracy when
defending against various types of graph backdoor attacks with different
properties.
| [
{
"version": "v1",
"created": "Fri, 14 Jun 2024 08:46:26 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 02:55:02 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhang",
"Zhiwei",
""
],
[
"Lin",
"Minhua",
""
],
[
"Xu",
"Junjie",
""
],
[
"Wu",
"Zongyu",
""
],
[
"Dai",
"Enyan",
""
],
[
"Wang",
"Suhang",
""
]
]
| TITLE: Robustness Inspired Graph Backdoor Defense
ABSTRACT: Graph Neural Networks (GNNs) have achieved promising results in tasks such as
node classification and graph classification. However, recent studies reveal
that GNNs are vulnerable to backdoor attacks, posing a significant threat to
their real-world adoption. Despite initial efforts to defend against specific
graph backdoor attacks, there is no work on defending against various types of
backdoor attacks where generated triggers have different properties. Hence, we
first empirically verify that prediction variance under edge dropping is a
crucial indicator for identifying poisoned nodes. With this observation, we
propose using random edge dropping to detect backdoors and theoretically show
that it can efficiently distinguish poisoned nodes from clean ones.
Furthermore, we introduce a novel robust training strategy to efficiently
counteract the impact of the triggers. Extensive experiments on real-world
datasets show that our framework can effectively identify poisoned nodes,
significantly degrade the attack success rate, and maintain clean accuracy when
defending against various types of graph backdoor attacks with different
properties.
| no_new_dataset | 0.945298 |
2406.16038 | Delin Qu | Delin Qu, Qizhi Chen, Pingrui Zhang, Xianqiang Gao, Junzhe Li, Bin
Zhao, Dong Wang and Xuelong Li | LiveScene: Language Embedding Interactive Radiance Fields for Physical
Scene Rendering and Control | Accepted at Neurips 2024 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper scales object-level reconstruction to complex scenes, advancing
interactive scene reconstruction. We introduce two datasets, OmniSim and
InterReal, featuring 28 scenes with multiple interactive objects. To tackle the
challenge of inaccurate interactive motion recovery in complex scenes, we
propose LiveScene, a scene-level language-embedded interactive radiance field
that efficiently reconstructs and controls multiple objects. By decomposing the
interactive scene into local deformable fields, LiveScene enables separate
reconstruction of individual object motions, reducing memory consumption.
Additionally, our interaction-aware language embedding localizes individual
interactive objects, allowing for arbitrary control using natural language. Our
approach demonstrates significant superiority in novel view synthesis,
interactive scene control, and language grounding performance through extensive
experiments. Project page: https://livescenes.github.io.
| [
{
"version": "v1",
"created": "Sun, 23 Jun 2024 07:26:13 GMT"
},
{
"version": "v2",
"created": "Sun, 3 Nov 2024 07:37:05 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 03:19:42 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Qu",
"Delin",
""
],
[
"Chen",
"Qizhi",
""
],
[
"Zhang",
"Pingrui",
""
],
[
"Gao",
"Xianqiang",
""
],
[
"Li",
"Junzhe",
""
],
[
"Zhao",
"Bin",
""
],
[
"Wang",
"Dong",
""
],
[
"Li",
"Xuelong",
""
]
]
| TITLE: LiveScene: Language Embedding Interactive Radiance Fields for Physical
Scene Rendering and Control
ABSTRACT: This paper scales object-level reconstruction to complex scenes, advancing
interactive scene reconstruction. We introduce two datasets, OmniSim and
InterReal, featuring 28 scenes with multiple interactive objects. To tackle the
challenge of inaccurate interactive motion recovery in complex scenes, we
propose LiveScene, a scene-level language-embedded interactive radiance field
that efficiently reconstructs and controls multiple objects. By decomposing the
interactive scene into local deformable fields, LiveScene enables separate
reconstruction of individual object motions, reducing memory consumption.
Additionally, our interaction-aware language embedding localizes individual
interactive objects, allowing for arbitrary control using natural language. Our
approach demonstrates significant superiority in novel view synthesis,
interactive scene control, and language grounding performance through extensive
experiments. Project page: https://livescenes.github.io.
| new_dataset | 0.952264 |
2407.02165 | Zihao Huang | Zihao Huang, Shoukang Hu, Guangcong Wang, Tianqi Liu, Yuhang Zang,
Zhiguo Cao, Wei Li, Ziwei Liu | WildAvatar: Learning In-the-wild 3D Avatars from the Web | CVPR2025, Project page: https://wildavatar.github.io/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing research on avatar creation is typically limited to laboratory
datasets, which require high costs against scalability and exhibit insufficient
representation of the real world. On the other hand, the web abounds with
off-the-shelf real-world human videos, but these videos vary in quality and
require accurate annotations for avatar creation. To this end, we propose an
automatic annotating pipeline with filtering protocols to curate these humans
from the web. Our pipeline surpasses state-of-the-art methods on the EMDB
benchmark, and the filtering protocols boost verification metrics on web
videos. We then curate WildAvatar, a web-scale in-the-wild human avatar
creation dataset extracted from YouTube, with $10000+$ different human subjects
and scenes. WildAvatar is at least $10\times$ richer than previous datasets for
3D human avatar creation and closer to the real world. To explore its
potential, we demonstrate the quality and generalizability of avatar creation
methods on WildAvatar. We will publicly release our code, data source links and
annotations to push forward 3D human avatar creation and other related fields
for real-world applications.
| [
{
"version": "v1",
"created": "Tue, 2 Jul 2024 11:17:48 GMT"
},
{
"version": "v2",
"created": "Wed, 10 Jul 2024 09:20:39 GMT"
},
{
"version": "v3",
"created": "Sun, 14 Jul 2024 08:15:12 GMT"
},
{
"version": "v4",
"created": "Wed, 12 Mar 2025 14:19:55 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Huang",
"Zihao",
""
],
[
"Hu",
"Shoukang",
""
],
[
"Wang",
"Guangcong",
""
],
[
"Liu",
"Tianqi",
""
],
[
"Zang",
"Yuhang",
""
],
[
"Cao",
"Zhiguo",
""
],
[
"Li",
"Wei",
""
],
[
"Liu",
"Ziwei",
""
]
]
| TITLE: WildAvatar: Learning In-the-wild 3D Avatars from the Web
ABSTRACT: Existing research on avatar creation is typically limited to laboratory
datasets, which require high costs against scalability and exhibit insufficient
representation of the real world. On the other hand, the web abounds with
off-the-shelf real-world human videos, but these videos vary in quality and
require accurate annotations for avatar creation. To this end, we propose an
automatic annotating pipeline with filtering protocols to curate these humans
from the web. Our pipeline surpasses state-of-the-art methods on the EMDB
benchmark, and the filtering protocols boost verification metrics on web
videos. We then curate WildAvatar, a web-scale in-the-wild human avatar
creation dataset extracted from YouTube, with $10000+$ different human subjects
and scenes. WildAvatar is at least $10\times$ richer than previous datasets for
3D human avatar creation and closer to the real world. To explore its
potential, we demonstrate the quality and generalizability of avatar creation
methods on WildAvatar. We will publicly release our code, data source links and
annotations to push forward 3D human avatar creation and other related fields
for real-world applications.
| new_dataset | 0.96606 |
2407.06060 | Pedro Louro | Pedro Lima Louro, Hugo Redinho, Ricardo Santos, Ricardo Malheiro,
Renato Panda, Rui Pedro Paiva | MERGE -- A Bimodal Dataset for Static Music Emotion Recognition | 16 pages, 4 figures, 13 tables, submitted to IEEE Transactions on
Affective Computing | null | null | null | cs.SD cs.IR cs.LG cs.MM eess.AS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The Music Emotion Recognition (MER) field has seen steady developments in
recent years, with contributions from feature engineering, machine learning,
and deep learning. The landscape has also shifted from audio-centric systems to
bimodal ensembles that combine audio and lyrics. However, a severe lack of
public and sizeable bimodal databases has hampered the development and
improvement of bimodal audio-lyrics systems. This article proposes three new
audio, lyrics, and bimodal MER research datasets, collectively called MERGE,
created using a semi-automatic approach. To comprehensively assess the proposed
datasets and establish a baseline for benchmarking, we conducted several
experiments for each modality, using feature engineering, machine learning, and
deep learning methodologies. In addition, we propose and validate fixed
train-validate-test splits. The obtained results confirm the viability of the
proposed datasets, achieving the best overall result of 79.21% F1-score for
bimodal classification using a deep neural network.
| [
{
"version": "v1",
"created": "Mon, 8 Jul 2024 16:01:04 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 00:52:43 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Louro",
"Pedro Lima",
""
],
[
"Redinho",
"Hugo",
""
],
[
"Santos",
"Ricardo",
""
],
[
"Malheiro",
"Ricardo",
""
],
[
"Panda",
"Renato",
""
],
[
"Paiva",
"Rui Pedro",
""
]
]
| TITLE: MERGE -- A Bimodal Dataset for Static Music Emotion Recognition
ABSTRACT: The Music Emotion Recognition (MER) field has seen steady developments in
recent years, with contributions from feature engineering, machine learning,
and deep learning. The landscape has also shifted from audio-centric systems to
bimodal ensembles that combine audio and lyrics. However, a severe lack of
public and sizeable bimodal databases has hampered the development and
improvement of bimodal audio-lyrics systems. This article proposes three new
audio, lyrics, and bimodal MER research datasets, collectively called MERGE,
created using a semi-automatic approach. To comprehensively assess the proposed
datasets and establish a baseline for benchmarking, we conducted several
experiments for each modality, using feature engineering, machine learning, and
deep learning methodologies. In addition, we propose and validate fixed
train-validate-test splits. The obtained results confirm the viability of the
proposed datasets, achieving the best overall result of 79.21% F1-score for
bimodal classification using a deep neural network.
| new_dataset | 0.961098 |
2407.08952 | Ye Liu | Ye Liu, Jiajun Zhu, Xukai Liu, Haoyu Tang, Yanghai Zhang, Kai Zhang,
Xiaofang Zhou, Enhong Chen | Detect, Investigate, Judge and Determine: A Knowledge-guided Framework
for Few-shot Fake News Detection | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-Shot Fake News Detection (FS-FND) aims to distinguish inaccurate news
from real ones in extremely low-resource scenarios. This task has garnered
increased attention due to the widespread dissemination and harmful impact of
fake news on social media. Large Language Models (LLMs) have demonstrated
competitive performance with the help of their rich prior knowledge and
excellent in-context learning abilities. However, existing methods face
significant limitations, such as the Understanding Ambiguity and Information
Scarcity, which significantly undermine the potential of LLMs. To address these
shortcomings, we propose a Dual-perspective Knowledge-guided Fake News
Detection (DKFND) model, designed to enhance LLMs from both inside and outside
perspectives. Specifically, DKFND first identifies the knowledge concepts of
each news article through a Detection Module. Subsequently, DKFND creatively
designs an Investigation Module to retrieve inside and outside valuable
information concerning to the current news, followed by another Judge Module to
evaluate the relevance and confidence of them. Finally, a Determination Module
further derives two respective predictions and obtain the final result.
Extensive experiments on two public datasets show the efficacy of our proposed
method, particularly in low-resource settings.
| [
{
"version": "v1",
"created": "Fri, 12 Jul 2024 03:15:01 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Feb 2025 04:56:16 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Feb 2025 05:25:32 GMT"
},
{
"version": "v4",
"created": "Tue, 11 Mar 2025 13:06:04 GMT"
},
{
"version": "v5",
"created": "Wed, 12 Mar 2025 04:46:47 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Liu",
"Ye",
""
],
[
"Zhu",
"Jiajun",
""
],
[
"Liu",
"Xukai",
""
],
[
"Tang",
"Haoyu",
""
],
[
"Zhang",
"Yanghai",
""
],
[
"Zhang",
"Kai",
""
],
[
"Zhou",
"Xiaofang",
""
],
[
"Chen",
"Enhong",
""
]
]
| TITLE: Detect, Investigate, Judge and Determine: A Knowledge-guided Framework
for Few-shot Fake News Detection
ABSTRACT: Few-Shot Fake News Detection (FS-FND) aims to distinguish inaccurate news
from real ones in extremely low-resource scenarios. This task has garnered
increased attention due to the widespread dissemination and harmful impact of
fake news on social media. Large Language Models (LLMs) have demonstrated
competitive performance with the help of their rich prior knowledge and
excellent in-context learning abilities. However, existing methods face
significant limitations, such as the Understanding Ambiguity and Information
Scarcity, which significantly undermine the potential of LLMs. To address these
shortcomings, we propose a Dual-perspective Knowledge-guided Fake News
Detection (DKFND) model, designed to enhance LLMs from both inside and outside
perspectives. Specifically, DKFND first identifies the knowledge concepts of
each news article through a Detection Module. Subsequently, DKFND creatively
designs an Investigation Module to retrieve inside and outside valuable
information concerning to the current news, followed by another Judge Module to
evaluate the relevance and confidence of them. Finally, a Determination Module
further derives two respective predictions and obtain the final result.
Extensive experiments on two public datasets show the efficacy of our proposed
method, particularly in low-resource settings.
| no_new_dataset | 0.944074 |
2407.12358 | Chuwei Luo | Yufan Shen, Chuwei Luo, Zhaoqing Zhu, Yang Chen, Qi Zheng, Zhi Yu,
Jiajun Bu, Cong Yao | ProcTag: Process Tagging for Assessing the Efficacy of Document
Instruction Data | AAAI 2025 | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, large language models (LLMs) and multimodal large language models
(MLLMs) have demonstrated promising results on document visual question
answering (VQA) task, particularly after training on document instruction
datasets. An effective evaluation method for document instruction data is
crucial in constructing instruction data with high efficacy, which, in turn,
facilitates the training of LLMs and MLLMs for document VQA. However, most
existing evaluation methods for instruction data are limited to the textual
content of the instructions themselves, thereby hindering the effective
assessment of document instruction datasets and constraining their
construction. In this paper, we propose ProcTag, a data-oriented method that
assesses the efficacy of document instruction data. ProcTag innovatively
performs tagging on the execution process of instructions rather than the
instruction text itself. By leveraging the diversity and complexity of these
tags to assess the efficacy of the given dataset, ProcTag enables selective
sampling or filtering of document instructions. Furthermore, DocLayPrompt, a
novel semi-structured layout-aware document prompting strategy, is proposed for
effectively representing documents. Experiments demonstrate that sampling
existing open-sourced and generated document VQA/instruction datasets with
ProcTag significantly outperforms current methods for evaluating instruction
data. Impressively, with ProcTag-based sampling in the generated document
datasets, only 30.5\% of the document instructions are required to achieve
100\% efficacy compared to the complete dataset. The code is publicly available
at
https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/DocumentUnderstanding/ProcTag.
| [
{
"version": "v1",
"created": "Wed, 17 Jul 2024 07:29:59 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 02:20:28 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Shen",
"Yufan",
""
],
[
"Luo",
"Chuwei",
""
],
[
"Zhu",
"Zhaoqing",
""
],
[
"Chen",
"Yang",
""
],
[
"Zheng",
"Qi",
""
],
[
"Yu",
"Zhi",
""
],
[
"Bu",
"Jiajun",
""
],
[
"Yao",
"Cong",
""
]
]
| TITLE: ProcTag: Process Tagging for Assessing the Efficacy of Document
Instruction Data
ABSTRACT: Recently, large language models (LLMs) and multimodal large language models
(MLLMs) have demonstrated promising results on document visual question
answering (VQA) task, particularly after training on document instruction
datasets. An effective evaluation method for document instruction data is
crucial in constructing instruction data with high efficacy, which, in turn,
facilitates the training of LLMs and MLLMs for document VQA. However, most
existing evaluation methods for instruction data are limited to the textual
content of the instructions themselves, thereby hindering the effective
assessment of document instruction datasets and constraining their
construction. In this paper, we propose ProcTag, a data-oriented method that
assesses the efficacy of document instruction data. ProcTag innovatively
performs tagging on the execution process of instructions rather than the
instruction text itself. By leveraging the diversity and complexity of these
tags to assess the efficacy of the given dataset, ProcTag enables selective
sampling or filtering of document instructions. Furthermore, DocLayPrompt, a
novel semi-structured layout-aware document prompting strategy, is proposed for
effectively representing documents. Experiments demonstrate that sampling
existing open-sourced and generated document VQA/instruction datasets with
ProcTag significantly outperforms current methods for evaluating instruction
data. Impressively, with ProcTag-based sampling in the generated document
datasets, only 30.5\% of the document instructions are required to achieve
100\% efficacy compared to the complete dataset. The code is publicly available
at
https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/DocumentUnderstanding/ProcTag.
| no_new_dataset | 0.924279 |
2407.13329 | Lorenzo Paolini | Lorenzo Paolini, Sahar Vahdati, Angelo Di Iorio, Robert Wardenga, Ivan
Heibi, Silvio Peroni | CiteFusion: An Ensemble Framework for Citation Intent Classification
Harnessing Dual-Model Binary Couples and SHAP Analyses | Submitted to Scientometrics Journal | null | 10.5281/zenodo.15011985 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding the motivations underlying scholarly citations is critical for
evaluating research impact and fostering transparent scholarly communication.
This study introduces CiteFusion, an ensemble framework designed to address the
multiclass Citation Intent Classification (CIC) task on benchmark datasets,
SciCite and ACL-ARC. The framework decomposes the task into binary
classification subtasks, utilizing complementary pairs of SciBERT and XLNet
models fine-tuned independently for each citation intent. These base models are
aggregated through a feedforward neural network meta-classifier, ensuring
robust performance in imbalanced and data-scarce scenarios. To enhance
interpretability, SHAP (SHapley Additive exPlanations) is employed to analyze
token-level contributions and interactions among base models, providing
transparency into classification dynamics. We further investigate the semantic
role of structural context by incorporating section titles into input
sentences, demonstrating their significant impact on classification accuracy
and model reliability. Experimental results show that CiteFusion achieves
state-of-the-art performance, with Macro-F1 scores of 89.60% on SciCite and
76.24% on ACL-ARC. The original intents from both datasets are mapped to
Citation Typing Ontology (CiTO) object properties to ensure interoperability
and reusability. This mapping highlights overlaps between the two datasets
labels, enhancing their understandability and reusability. Finally, we release
a web-based application that classifies citation intents leveraging CiteFusion
models developed on SciCite.
| [
{
"version": "v1",
"created": "Thu, 18 Jul 2024 09:29:33 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 11:59:18 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Paolini",
"Lorenzo",
""
],
[
"Vahdati",
"Sahar",
""
],
[
"Di Iorio",
"Angelo",
""
],
[
"Wardenga",
"Robert",
""
],
[
"Heibi",
"Ivan",
""
],
[
"Peroni",
"Silvio",
""
]
]
| TITLE: CiteFusion: An Ensemble Framework for Citation Intent Classification
Harnessing Dual-Model Binary Couples and SHAP Analyses
ABSTRACT: Understanding the motivations underlying scholarly citations is critical for
evaluating research impact and fostering transparent scholarly communication.
This study introduces CiteFusion, an ensemble framework designed to address the
multiclass Citation Intent Classification (CIC) task on benchmark datasets,
SciCite and ACL-ARC. The framework decomposes the task into binary
classification subtasks, utilizing complementary pairs of SciBERT and XLNet
models fine-tuned independently for each citation intent. These base models are
aggregated through a feedforward neural network meta-classifier, ensuring
robust performance in imbalanced and data-scarce scenarios. To enhance
interpretability, SHAP (SHapley Additive exPlanations) is employed to analyze
token-level contributions and interactions among base models, providing
transparency into classification dynamics. We further investigate the semantic
role of structural context by incorporating section titles into input
sentences, demonstrating their significant impact on classification accuracy
and model reliability. Experimental results show that CiteFusion achieves
state-of-the-art performance, with Macro-F1 scores of 89.60% on SciCite and
76.24% on ACL-ARC. The original intents from both datasets are mapped to
Citation Typing Ontology (CiTO) object properties to ensure interoperability
and reusability. This mapping highlights overlaps between the two datasets
labels, enhancing their understandability and reusability. Finally, we release
a web-based application that classifies citation intents leveraging CiteFusion
models developed on SciCite.
| no_new_dataset | 0.950549 |
2407.14210 | Jos\'e Daniel Pascual-Triana | Jos\'e Daniel Pascual-Triana, Alberto Fern\'andez, Paulo Novais,
Francisco Herrera | Fair Overlap Number of Balls (Fair-ONB): A Data-Morphology-based
Undersampling Method for Bias Reduction | 14 pages, 5 tables, 8 figures | null | 10.1080/02331888.2025.2476029 | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the key issues regarding classification problems in Trustworthy
Artificial Intelligence is ensuring Fairness in the prediction of different
classes when protected (sensitive) features are present. Data quality is
critical in these cases, as biases in training data can be reflected in machine
learning, impacting human lives and failing to comply with current regulations.
One strategy to improve data quality and avoid these problems is preprocessing
the dataset. Instance selection via undersampling can foster balanced learning
of classes and protected feature values. Performing undersampling in class
overlap areas close to the decision boundary should bolster the impact on the
classifier. This work proposes Fair Overlap Number of Balls (Fair-ONB), an
undersampling method that harnesses the data morphology of the different data
groups (obtained from the combination of classes and protected feature values)
to perform guided undersampling in overlap areas. It employs attributes of the
ball coverage of the groups, such as the radius, number of covered instances
and density, to select the most suitable areas for undersampling and reduce
bias. Results show that the Fair-ONB method improves model Fairness with low
impact on the classifier's predictive performance.
| [
{
"version": "v1",
"created": "Fri, 19 Jul 2024 11:16:02 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Sep 2024 16:52:05 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Pascual-Triana",
"José Daniel",
""
],
[
"Fernández",
"Alberto",
""
],
[
"Novais",
"Paulo",
""
],
[
"Herrera",
"Francisco",
""
]
]
| TITLE: Fair Overlap Number of Balls (Fair-ONB): A Data-Morphology-based
Undersampling Method for Bias Reduction
ABSTRACT: One of the key issues regarding classification problems in Trustworthy
Artificial Intelligence is ensuring Fairness in the prediction of different
classes when protected (sensitive) features are present. Data quality is
critical in these cases, as biases in training data can be reflected in machine
learning, impacting human lives and failing to comply with current regulations.
One strategy to improve data quality and avoid these problems is preprocessing
the dataset. Instance selection via undersampling can foster balanced learning
of classes and protected feature values. Performing undersampling in class
overlap areas close to the decision boundary should bolster the impact on the
classifier. This work proposes Fair Overlap Number of Balls (Fair-ONB), an
undersampling method that harnesses the data morphology of the different data
groups (obtained from the combination of classes and protected feature values)
to perform guided undersampling in overlap areas. It employs attributes of the
ball coverage of the groups, such as the radius, number of covered instances
and density, to select the most suitable areas for undersampling and reduce
bias. Results show that the Fair-ONB method improves model Fairness with low
impact on the classifier's predictive performance.
| no_new_dataset | 0.953966 |
2407.15620 | Xihong Yang | Xihong Yang, Yiqi Wang, Jin Chen, Wenqi Fan, Xiangyu Zhao, En Zhu,
Xinwang Liu, Defu Lian | Dual Test-time Training for Out-of-distribution Recommender System | null | null | null | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning has been widely applied in recommender systems, which has
achieved revolutionary progress recently. However, most existing learning-based
methods assume that the user and item distributions remain unchanged between
the training phase and the test phase. However, the distribution of user and
item features can naturally shift in real-world scenarios, potentially
resulting in a substantial decrease in recommendation performance. This
phenomenon can be formulated as an Out-Of-Distribution (OOD) recommendation
problem. To address this challenge, we propose a novel Dual Test-Time-Training
framework for OOD Recommendation, termed DT3OR. In DT3OR, we incorporate a
model adaptation mechanism during the test-time phase to carefully update the
recommendation model, allowing the model to specially adapt to the shifting
user and item features. To be specific, we propose a self-distillation task and
a contrastive task to assist the model learning both the user's invariant
interest preferences and the variant user/item characteristics during the
test-time phase, thus facilitating a smooth adaptation to the shifting
features. Furthermore, we provide theoretical analysis to support the rationale
behind our dual test-time training framework. To the best of our knowledge,
this paper is the first work to address OOD recommendation via a
test-time-training strategy. We conduct experiments on three datasets with
various backbones. Comprehensive experimental results have demonstrated the
effectiveness of DT3OR compared to other state-of-the-art baselines.
| [
{
"version": "v1",
"created": "Mon, 22 Jul 2024 13:27:51 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 14:06:24 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Yang",
"Xihong",
""
],
[
"Wang",
"Yiqi",
""
],
[
"Chen",
"Jin",
""
],
[
"Fan",
"Wenqi",
""
],
[
"Zhao",
"Xiangyu",
""
],
[
"Zhu",
"En",
""
],
[
"Liu",
"Xinwang",
""
],
[
"Lian",
"Defu",
""
]
]
| TITLE: Dual Test-time Training for Out-of-distribution Recommender System
ABSTRACT: Deep learning has been widely applied in recommender systems, which has
achieved revolutionary progress recently. However, most existing learning-based
methods assume that the user and item distributions remain unchanged between
the training phase and the test phase. However, the distribution of user and
item features can naturally shift in real-world scenarios, potentially
resulting in a substantial decrease in recommendation performance. This
phenomenon can be formulated as an Out-Of-Distribution (OOD) recommendation
problem. To address this challenge, we propose a novel Dual Test-Time-Training
framework for OOD Recommendation, termed DT3OR. In DT3OR, we incorporate a
model adaptation mechanism during the test-time phase to carefully update the
recommendation model, allowing the model to specially adapt to the shifting
user and item features. To be specific, we propose a self-distillation task and
a contrastive task to assist the model learning both the user's invariant
interest preferences and the variant user/item characteristics during the
test-time phase, thus facilitating a smooth adaptation to the shifting
features. Furthermore, we provide theoretical analysis to support the rationale
behind our dual test-time training framework. To the best of our knowledge,
this paper is the first work to address OOD recommendation via a
test-time-training strategy. We conduct experiments on three datasets with
various backbones. Comprehensive experimental results have demonstrated the
effectiveness of DT3OR compared to other state-of-the-art baselines.
| no_new_dataset | 0.944331 |
2407.21299 | Kaustav Bhattacharjee | Kaustav Bhattacharjee, Soumya Kundu, Indrasis Chakraborty, and Aritra
Dasgupta | Who should I trust? A Visual Analytics Approach for Comparing Net Load
Forecasting Models | Accepted for publication in the proceedings of 2025 IEEE PES Grid
Edge Technologies Conference & Exposition (Grid Edge) | GridEdge 2025, pp. 1-5, 2025 | 10.1109/GridEdge61154.2025.10887523 | null | cs.HC cs.AI cs.LG cs.SY eess.SP eess.SY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Net load forecasting is crucial for energy planning and facilitating informed
decision-making regarding trade and load distributions. However, evaluating
forecasting models' performance against benchmark models remains challenging,
thereby impeding experts' trust in the model's performance. In this context,
there is a demand for technological interventions that allow scientists to
compare models across various timeframes and solar penetration levels. This
paper introduces a visual analytics-based application designed to compare the
performance of deep-learning-based net load forecasting models with other
models for probabilistic net load forecasting. This application employs
carefully selected visual analytic interventions, enabling users to discern
differences in model performance across different solar penetration levels,
dataset resolutions, and hours of the day over multiple months. We also present
observations made using our application through a case study, demonstrating the
effectiveness of visualizations in aiding scientists in making informed
decisions and enhancing trust in net load forecasting models.
| [
{
"version": "v1",
"created": "Wed, 31 Jul 2024 02:57:21 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Bhattacharjee",
"Kaustav",
""
],
[
"Kundu",
"Soumya",
""
],
[
"Chakraborty",
"Indrasis",
""
],
[
"Dasgupta",
"Aritra",
""
]
]
| TITLE: Who should I trust? A Visual Analytics Approach for Comparing Net Load
Forecasting Models
ABSTRACT: Net load forecasting is crucial for energy planning and facilitating informed
decision-making regarding trade and load distributions. However, evaluating
forecasting models' performance against benchmark models remains challenging,
thereby impeding experts' trust in the model's performance. In this context,
there is a demand for technological interventions that allow scientists to
compare models across various timeframes and solar penetration levels. This
paper introduces a visual analytics-based application designed to compare the
performance of deep-learning-based net load forecasting models with other
models for probabilistic net load forecasting. This application employs
carefully selected visual analytic interventions, enabling users to discern
differences in model performance across different solar penetration levels,
dataset resolutions, and hours of the day over multiple months. We also present
observations made using our application through a case study, demonstrating the
effectiveness of visualizations in aiding scientists in making informed
decisions and enhancing trust in net load forecasting models.
| no_new_dataset | 0.950227 |
2408.00374 | Rahul Bhadani | Xi Chen, Rahul Bhadani, Larry Head | Conformal Trajectory Prediction with Multi-View Data Integration in
Cooperative Driving | null | null | null | null | cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Current research on trajectory prediction primarily relies on data collected
by onboard sensors of an ego vehicle. With the rapid advancement in connected
technologies, such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure
(V2I) communication, valuable information from alternate views becomes
accessible via wireless networks. The integration of information from
alternative views has the potential to overcome the inherent limitations
associated with a single viewpoint, such as occlusions and limited field of
view. In this work, we introduce V2INet, a novel trajectory prediction
framework designed to model multi-view data by extending existing single-view
models. Unlike previous approaches where the multi-view data is manually fused
or formulated as a separate training stage, our model supports end-to-end
training, enhancing both flexibility and performance. Moreover, the predicted
multimodal trajectories are calibrated by a post-hoc conformal prediction
module to get valid and efficient confidence regions. We evaluated the entire
framework using the real-world V2I dataset V2X-Seq. Our results demonstrate
superior performance in terms of Final Displacement Error (FDE) and Miss Rate
(MR) using a single GPU. The code is publicly available at:
https://github.com/xichennn/V2I_trajectory_prediction.
| [
{
"version": "v1",
"created": "Thu, 1 Aug 2024 08:32:03 GMT"
},
{
"version": "v2",
"created": "Fri, 2 Aug 2024 13:00:46 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 18:19:56 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chen",
"Xi",
""
],
[
"Bhadani",
"Rahul",
""
],
[
"Head",
"Larry",
""
]
]
| TITLE: Conformal Trajectory Prediction with Multi-View Data Integration in
Cooperative Driving
ABSTRACT: Current research on trajectory prediction primarily relies on data collected
by onboard sensors of an ego vehicle. With the rapid advancement in connected
technologies, such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure
(V2I) communication, valuable information from alternate views becomes
accessible via wireless networks. The integration of information from
alternative views has the potential to overcome the inherent limitations
associated with a single viewpoint, such as occlusions and limited field of
view. In this work, we introduce V2INet, a novel trajectory prediction
framework designed to model multi-view data by extending existing single-view
models. Unlike previous approaches where the multi-view data is manually fused
or formulated as a separate training stage, our model supports end-to-end
training, enhancing both flexibility and performance. Moreover, the predicted
multimodal trajectories are calibrated by a post-hoc conformal prediction
module to get valid and efficient confidence regions. We evaluated the entire
framework using the real-world V2I dataset V2X-Seq. Our results demonstrate
superior performance in terms of Final Displacement Error (FDE) and Miss Rate
(MR) using a single GPU. The code is publicly available at:
https://github.com/xichennn/V2I_trajectory_prediction.
| no_new_dataset | 0.943086 |
2408.00531 | Max Klabunde | Max Klabunde, Tassilo Wald, Tobias Schumacher, Klaus Maier-Hein,
Markus Strohmaier, Florian Lemmerich | ReSi: A Comprehensive Benchmark for Representational Similarity Measures | ICLR 2025. Code and data at https://github.com/mklabunde/resi | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Measuring the similarity of different representations of neural architectures
is a fundamental task and an open research challenge for the machine learning
community. This paper presents the first comprehensive benchmark for evaluating
representational similarity measures based on well-defined groundings of
similarity. The representational similarity (ReSi) benchmark consists of (i)
six carefully designed tests for similarity measures, (ii) 24 similarity
measures, (iii) 14 neural network architectures, and (iv) seven datasets,
spanning over the graph, language, and vision domains. The benchmark opens up
several important avenues of research on representational similarity that
enable novel explorations and applications of neural architectures. We
demonstrate the utility of the ReSi benchmark by conducting experiments on
various neural network architectures, real world datasets and similarity
measures. All components of the benchmark are publicly available and thereby
facilitate systematic reproduction and production of research results. The
benchmark is extensible, future research can build on and further expand it. We
believe that the ReSi benchmark can serve as a sound platform catalyzing future
research that aims to systematically evaluate existing and explore novel ways
of comparing representations of neural architectures.
| [
{
"version": "v1",
"created": "Thu, 1 Aug 2024 13:08:02 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 20:01:30 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Klabunde",
"Max",
""
],
[
"Wald",
"Tassilo",
""
],
[
"Schumacher",
"Tobias",
""
],
[
"Maier-Hein",
"Klaus",
""
],
[
"Strohmaier",
"Markus",
""
],
[
"Lemmerich",
"Florian",
""
]
]
| TITLE: ReSi: A Comprehensive Benchmark for Representational Similarity Measures
ABSTRACT: Measuring the similarity of different representations of neural architectures
is a fundamental task and an open research challenge for the machine learning
community. This paper presents the first comprehensive benchmark for evaluating
representational similarity measures based on well-defined groundings of
similarity. The representational similarity (ReSi) benchmark consists of (i)
six carefully designed tests for similarity measures, (ii) 24 similarity
measures, (iii) 14 neural network architectures, and (iv) seven datasets,
spanning over the graph, language, and vision domains. The benchmark opens up
several important avenues of research on representational similarity that
enable novel explorations and applications of neural architectures. We
demonstrate the utility of the ReSi benchmark by conducting experiments on
various neural network architectures, real world datasets and similarity
measures. All components of the benchmark are publicly available and thereby
facilitate systematic reproduction and production of research results. The
benchmark is extensible, future research can build on and further expand it. We
believe that the ReSi benchmark can serve as a sound platform catalyzing future
research that aims to systematically evaluate existing and explore novel ways
of comparing representations of neural architectures.
| new_dataset | 0.82566 |
2408.01434 | A'di Dust | Adi Dust, Pat Levitt, Maja Matari\'c | Behind the Smile: Mental Health Implications of Mother-Infant
Interactions Revealed Through Smile Analysis | 9 pages, 2 Figures, Affective Computing & Intelligent Interaction
Conference 2024 | null | 10.1109/ACII63134.2024.00010 | null | cs.CY cs.HC | http://creativecommons.org/licenses/by/4.0/ | Mothers of infants have specific demands in fostering emotional bonds with
their children, characterized by dynamics that are different from adult-adult
interactions, notably requiring heightened maternal emotional regulation. In
this study, we analyzed maternal emotional state by modeling maternal emotion
regulation reflected in smiles. The dataset comprises N=94 videos of
approximately 3 plus or minus 1-minutes, capturing free play interactions
between 6 and 12-month-old infants and their mothers. Corresponding demographic
details of self-reported maternal mental health provide variables for
determining mothers' relations to emotions measured during free play. In this
work, we employ diverse methodological approaches to explore the temporal
evolution of maternal smiles. Our findings reveal a correlation between the
temporal dynamics of mothers' smiles and their emotional state. Furthermore, we
identify specific smile features that correlate with maternal emotional state,
thereby enabling informed inferences with existing literature on general smile
analysis. This study offers insights into emotional labor, defined as the
management of one's own emotions for the benefit of others, and emotion
regulation entailed in mother-infant interactions.
| [
{
"version": "v1",
"created": "Thu, 18 Jul 2024 23:22:57 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 23:31:31 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Dust",
"Adi",
""
],
[
"Levitt",
"Pat",
""
],
[
"Matarić",
"Maja",
""
]
]
| TITLE: Behind the Smile: Mental Health Implications of Mother-Infant
Interactions Revealed Through Smile Analysis
ABSTRACT: Mothers of infants have specific demands in fostering emotional bonds with
their children, characterized by dynamics that are different from adult-adult
interactions, notably requiring heightened maternal emotional regulation. In
this study, we analyzed maternal emotional state by modeling maternal emotion
regulation reflected in smiles. The dataset comprises N=94 videos of
approximately 3 plus or minus 1-minutes, capturing free play interactions
between 6 and 12-month-old infants and their mothers. Corresponding demographic
details of self-reported maternal mental health provide variables for
determining mothers' relations to emotions measured during free play. In this
work, we employ diverse methodological approaches to explore the temporal
evolution of maternal smiles. Our findings reveal a correlation between the
temporal dynamics of mothers' smiles and their emotional state. Furthermore, we
identify specific smile features that correlate with maternal emotional state,
thereby enabling informed inferences with existing literature on general smile
analysis. This study offers insights into emotional labor, defined as the
management of one's own emotions for the benefit of others, and emotion
regulation entailed in mother-infant interactions.
| new_dataset | 0.643861 |
2408.14998 | Alloy Das | Alloy Das, Sanket Biswas, Umapada Pal, Josep Llad\'os, Saumik
Bhattacharya | FastTextSpotter: A High-Efficiency Transformer for Multilingual Scene
Text Spotting | Accepted in ICPR 2024 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The proliferation of scene text in both structured and unstructured
environments presents significant challenges in optical character recognition
(OCR), necessitating more efficient and robust text spotting solutions. This
paper presents FastTextSpotter, a framework that integrates a Swin Transformer
visual backbone with a Transformer Encoder-Decoder architecture, enhanced by a
novel, faster self-attention unit, SAC2, to improve processing speeds while
maintaining accuracy. FastTextSpotter has been validated across multiple
datasets, including ICDAR2015 for regular texts and CTW1500 and TotalText for
arbitrary-shaped texts, benchmarking against current state-of-the-art models.
Our results indicate that FastTextSpotter not only achieves superior accuracy
in detecting and recognizing multilingual scene text (English and Vietnamese)
but also improves model efficiency, thereby setting new benchmarks in the
field. This study underscores the potential of advanced transformer
architectures in improving the adaptability and speed of text spotting
applications in diverse real-world settings. The dataset, code, and pre-trained
models have been released in our Github.
| [
{
"version": "v1",
"created": "Tue, 27 Aug 2024 12:28:41 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 14:56:20 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Das",
"Alloy",
""
],
[
"Biswas",
"Sanket",
""
],
[
"Pal",
"Umapada",
""
],
[
"Lladós",
"Josep",
""
],
[
"Bhattacharya",
"Saumik",
""
]
]
| TITLE: FastTextSpotter: A High-Efficiency Transformer for Multilingual Scene
Text Spotting
ABSTRACT: The proliferation of scene text in both structured and unstructured
environments presents significant challenges in optical character recognition
(OCR), necessitating more efficient and robust text spotting solutions. This
paper presents FastTextSpotter, a framework that integrates a Swin Transformer
visual backbone with a Transformer Encoder-Decoder architecture, enhanced by a
novel, faster self-attention unit, SAC2, to improve processing speeds while
maintaining accuracy. FastTextSpotter has been validated across multiple
datasets, including ICDAR2015 for regular texts and CTW1500 and TotalText for
arbitrary-shaped texts, benchmarking against current state-of-the-art models.
Our results indicate that FastTextSpotter not only achieves superior accuracy
in detecting and recognizing multilingual scene text (English and Vietnamese)
but also improves model efficiency, thereby setting new benchmarks in the
field. This study underscores the potential of advanced transformer
architectures in improving the adaptability and speed of text spotting
applications in diverse real-world settings. The dataset, code, and pre-trained
models have been released in our Github.
| new_dataset | 0.953923 |
2409.04011 | Weijie He | Weijie He, Mushui Liu, Yunlong Yu | Hybrid Mask Generation for Infrared Small Target Detection with
Single-Point Supervision | 11 pages, 9 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Single-frame infrared small target (SIRST) detection poses a significant
challenge due to the requirement to discern minute targets amidst complex
infrared background clutter. In this paper, we focus on a weakly-supervised
paradigm to obtain high-quality pseudo masks from the point-level annotation by
integrating a novel learning-free method with the hybrid of the learning-based
method. The learning-free method adheres to a sequential process, progressing
from a point annotation to the bounding box that encompasses the target, and
subsequently to detailed pseudo masks, while the hybrid is achieved through
filtering out false alarms and retrieving missed detections in the network's
prediction to provide a reliable supplement for learning-free masks. The
experimental results show that our learning-free method generates pseudo masks
with an average Intersection over Union (IoU) that is 4.3% higher than the
second-best learning-free competitor across three datasets, while the hybrid
learning-based method further enhances the quality of pseudo masks, achieving
an additional average IoU increase of 3.4%.
| [
{
"version": "v1",
"created": "Fri, 6 Sep 2024 03:34:44 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:13:29 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"He",
"Weijie",
""
],
[
"Liu",
"Mushui",
""
],
[
"Yu",
"Yunlong",
""
]
]
| TITLE: Hybrid Mask Generation for Infrared Small Target Detection with
Single-Point Supervision
ABSTRACT: Single-frame infrared small target (SIRST) detection poses a significant
challenge due to the requirement to discern minute targets amidst complex
infrared background clutter. In this paper, we focus on a weakly-supervised
paradigm to obtain high-quality pseudo masks from the point-level annotation by
integrating a novel learning-free method with the hybrid of the learning-based
method. The learning-free method adheres to a sequential process, progressing
from a point annotation to the bounding box that encompasses the target, and
subsequently to detailed pseudo masks, while the hybrid is achieved through
filtering out false alarms and retrieving missed detections in the network's
prediction to provide a reliable supplement for learning-free masks. The
experimental results show that our learning-free method generates pseudo masks
with an average Intersection over Union (IoU) that is 4.3% higher than the
second-best learning-free competitor across three datasets, while the hybrid
learning-based method further enhances the quality of pseudo masks, achieving
an additional average IoU increase of 3.4%.
| no_new_dataset | 0.951097 |
2409.04824 | Mahmoud Jahanshahi | Mahmoud Jahanshahi, David Reid, Adam McDaniel, Audris Mockus | OSS License Identification at Scale: A Comprehensive Dataset Using World
of Code | Accepted in 2025 IEEE/ACM 22st International Conference on Mining
Software Repositories (MSR) | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The proliferation of open source software (OSS) and different types of reuse
has made it incredibly difficult to perform an essential legal and compliance
task of accurate license identification within the software supply chain. This
study presents a reusable and comprehensive dataset of OSS licenses, created
using the World of Code (WoC) infrastructure. By scanning all files containing
"license" in their file paths, and applying the approximate matching via
winnowing algorithm to identify the most similar license from the SPDX list, we
found and identified 5.5 million distinct license blobs in OSS projects. The
dataset includes a detailed project-to-license (P2L) map with commit
timestamps, enabling dynamic analysis of license adoption and changes over
time. To verify the accuracy of the dataset we use stratified sampling and
manual review, achieving a final accuracy of 92.08%, with precision of 87.14%,
recall of 95.45%, and an F1 score of 91.11%. This dataset is intended to
support a range of research and practical tasks, including the detection of
license noncompliance, the investigations of license changes, study of
licensing trends, and the development of compliance tools. The dataset is open,
providing a valuable resource for developers, researchers, and legal
professionals in the OSS community.
| [
{
"version": "v1",
"created": "Sat, 7 Sep 2024 13:34:55 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2024 15:04:07 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 20:13:22 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Jahanshahi",
"Mahmoud",
""
],
[
"Reid",
"David",
""
],
[
"McDaniel",
"Adam",
""
],
[
"Mockus",
"Audris",
""
]
]
| TITLE: OSS License Identification at Scale: A Comprehensive Dataset Using World
of Code
ABSTRACT: The proliferation of open source software (OSS) and different types of reuse
has made it incredibly difficult to perform an essential legal and compliance
task of accurate license identification within the software supply chain. This
study presents a reusable and comprehensive dataset of OSS licenses, created
using the World of Code (WoC) infrastructure. By scanning all files containing
"license" in their file paths, and applying the approximate matching via
winnowing algorithm to identify the most similar license from the SPDX list, we
found and identified 5.5 million distinct license blobs in OSS projects. The
dataset includes a detailed project-to-license (P2L) map with commit
timestamps, enabling dynamic analysis of license adoption and changes over
time. To verify the accuracy of the dataset we use stratified sampling and
manual review, achieving a final accuracy of 92.08%, with precision of 87.14%,
recall of 95.45%, and an F1 score of 91.11%. This dataset is intended to
support a range of research and practical tasks, including the detection of
license noncompliance, the investigations of license changes, study of
licensing trends, and the development of compliance tools. The dataset is open,
providing a valuable resource for developers, researchers, and legal
professionals in the OSS community.
| new_dataset | 0.955236 |
2409.07041 | Xinrui Wang | Xinrui Wang, Lanqing Guo, Xiyu Wang, Siyu Huang, Bihan Wen | SoftShadow: Leveraging Soft Masks for Penumbra-Aware Shadow Removal | This paper has been accepted by CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in deep learning have yielded promising results for the
image shadow removal task. However, most existing methods rely on binary
pre-generated shadow masks. The binary nature of such masks could potentially
lead to artifacts near the boundary between shadow and non-shadow areas. In
view of this, inspired by the physical model of shadow formation, we introduce
novel soft shadow masks specifically designed for shadow removal. To achieve
such soft masks, we propose a SoftShadow framework by leveraging the prior
knowledge of pretrained SAM and integrating physical constraints. Specifically,
we jointly tune the SAM and the subsequent shadow removal network using
penumbra formation constraint loss, mask reconstruction loss, and shadow
removal loss. This framework enables accurate predictions of penumbra
(partially shaded) and umbra (fully shaded) areas while simultaneously
facilitating end-to-end shadow removal. Through extensive experiments on
popular datasets, we found that our SoftShadow framework, which generates soft
masks, can better restore boundary artifacts, achieve state-of-the-art
performance, and demonstrate superior generalizability.
| [
{
"version": "v1",
"created": "Wed, 11 Sep 2024 06:12:26 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 07:18:15 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Wang",
"Xinrui",
""
],
[
"Guo",
"Lanqing",
""
],
[
"Wang",
"Xiyu",
""
],
[
"Huang",
"Siyu",
""
],
[
"Wen",
"Bihan",
""
]
]
| TITLE: SoftShadow: Leveraging Soft Masks for Penumbra-Aware Shadow Removal
ABSTRACT: Recent advancements in deep learning have yielded promising results for the
image shadow removal task. However, most existing methods rely on binary
pre-generated shadow masks. The binary nature of such masks could potentially
lead to artifacts near the boundary between shadow and non-shadow areas. In
view of this, inspired by the physical model of shadow formation, we introduce
novel soft shadow masks specifically designed for shadow removal. To achieve
such soft masks, we propose a SoftShadow framework by leveraging the prior
knowledge of pretrained SAM and integrating physical constraints. Specifically,
we jointly tune the SAM and the subsequent shadow removal network using
penumbra formation constraint loss, mask reconstruction loss, and shadow
removal loss. This framework enables accurate predictions of penumbra
(partially shaded) and umbra (fully shaded) areas while simultaneously
facilitating end-to-end shadow removal. Through extensive experiments on
popular datasets, we found that our SoftShadow framework, which generates soft
masks, can better restore boundary artifacts, achieve state-of-the-art
performance, and demonstrate superior generalizability.
| no_new_dataset | 0.948917 |
2409.07989 | Amirreza Fateh | Fatemeh Askari, Amirreza Fateh, Mohammad Reza Mohammadi | Enhancing Few-Shot Image Classification through Learnable Multi-Scale
Embedding and Attention Mechanisms | null | null | 10.1016/j.neunet.2025.107339 | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | In the context of few-shot classification, the goal is to train a classifier
using a limited number of samples while maintaining satisfactory performance.
However, traditional metric-based methods exhibit certain limitations in
achieving this objective. These methods typically rely on a single distance
value between the query feature and support feature, thereby overlooking the
contribution of shallow features. To overcome this challenge, we propose a
novel approach in this paper. Our approach involves utilizing a multi-output
embedding network that maps samples into distinct feature spaces. The proposed
method extracts feature vectors at different stages, enabling the model to
capture both global and abstract features. By utilizing these diverse feature
spaces, our model enhances its performance. Moreover, employing a
self-attention mechanism improves the refinement of features at each stage,
leading to even more robust representations and improved overall performance.
Furthermore, assigning learnable weights to each stage significantly improved
performance and results. We conducted comprehensive evaluations on the
MiniImageNet and FC100 datasets, specifically in the 5-way 1-shot and 5-way
5-shot scenarios. Additionally, we performed cross-domain tasks across eight
benchmark datasets, achieving high accuracy in the testing domains. These
evaluations demonstrate the efficacy of our proposed method in comparison to
state-of-the-art approaches. https://github.com/FatemehAskari/MSENet
| [
{
"version": "v1",
"created": "Thu, 12 Sep 2024 12:34:29 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Jan 2025 14:01:58 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Askari",
"Fatemeh",
""
],
[
"Fateh",
"Amirreza",
""
],
[
"Mohammadi",
"Mohammad Reza",
""
]
]
| TITLE: Enhancing Few-Shot Image Classification through Learnable Multi-Scale
Embedding and Attention Mechanisms
ABSTRACT: In the context of few-shot classification, the goal is to train a classifier
using a limited number of samples while maintaining satisfactory performance.
However, traditional metric-based methods exhibit certain limitations in
achieving this objective. These methods typically rely on a single distance
value between the query feature and support feature, thereby overlooking the
contribution of shallow features. To overcome this challenge, we propose a
novel approach in this paper. Our approach involves utilizing a multi-output
embedding network that maps samples into distinct feature spaces. The proposed
method extracts feature vectors at different stages, enabling the model to
capture both global and abstract features. By utilizing these diverse feature
spaces, our model enhances its performance. Moreover, employing a
self-attention mechanism improves the refinement of features at each stage,
leading to even more robust representations and improved overall performance.
Furthermore, assigning learnable weights to each stage significantly improved
performance and results. We conducted comprehensive evaluations on the
MiniImageNet and FC100 datasets, specifically in the 5-way 1-shot and 5-way
5-shot scenarios. Additionally, we performed cross-domain tasks across eight
benchmark datasets, achieving high accuracy in the testing domains. These
evaluations demonstrate the efficacy of our proposed method in comparison to
state-of-the-art approaches. https://github.com/FatemehAskari/MSENet
| no_new_dataset | 0.947527 |
2409.09479 | Yutian Chen | Yuheng Qiu, Yutian Chen, Zihao Zhang, Wenshan Wang, Sebastian Scherer | MAC-VO: Metrics-aware Covariance for Learning-based Stereo Visual
Odometry | null | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose the MAC-VO, a novel learning-based stereo VO that leverages the
learned metrics-aware matching uncertainty for dual purposes: selecting
keypoint and weighing the residual in pose graph optimization. Compared to
traditional geometric methods prioritizing texture-affluent features like
edges, our keypoint selector employs the learned uncertainty to filter out the
low-quality features based on global inconsistency. In contrast to the
learning-based algorithms that model the scale-agnostic diagonal weight matrix
for covariance, we design a metrics-aware covariance model to capture the
spatial error during keypoint registration and the correlations between
different axes. Integrating this covariance model into pose graph optimization
enhances the robustness and reliability of pose estimation, particularly in
challenging environments with varying illumination, feature density, and motion
patterns. On public benchmark datasets, MAC-VO outperforms existing VO
algorithms and even some SLAM algorithms in challenging environments. The
covariance map also provides valuable information about the reliability of the
estimated poses, which can benefit decision-making for autonomous systems.
| [
{
"version": "v1",
"created": "Sat, 14 Sep 2024 16:49:42 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 04:51:33 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Qiu",
"Yuheng",
""
],
[
"Chen",
"Yutian",
""
],
[
"Zhang",
"Zihao",
""
],
[
"Wang",
"Wenshan",
""
],
[
"Scherer",
"Sebastian",
""
]
]
| TITLE: MAC-VO: Metrics-aware Covariance for Learning-based Stereo Visual
Odometry
ABSTRACT: We propose the MAC-VO, a novel learning-based stereo VO that leverages the
learned metrics-aware matching uncertainty for dual purposes: selecting
keypoint and weighing the residual in pose graph optimization. Compared to
traditional geometric methods prioritizing texture-affluent features like
edges, our keypoint selector employs the learned uncertainty to filter out the
low-quality features based on global inconsistency. In contrast to the
learning-based algorithms that model the scale-agnostic diagonal weight matrix
for covariance, we design a metrics-aware covariance model to capture the
spatial error during keypoint registration and the correlations between
different axes. Integrating this covariance model into pose graph optimization
enhances the robustness and reliability of pose estimation, particularly in
challenging environments with varying illumination, feature density, and motion
patterns. On public benchmark datasets, MAC-VO outperforms existing VO
algorithms and even some SLAM algorithms in challenging environments. The
covariance map also provides valuable information about the reliability of the
estimated poses, which can benefit decision-making for autonomous systems.
| no_new_dataset | 0.954095 |
2409.11889 | Jiaming Zhou | Jiaming Zhou, Shiwan Zhao, Jiabei He, Hui Wang, Wenjia Zeng, Yong
Chen, Haoqin Sun, Aobo Kong, Yong Qin | M2R-Whisper: Multi-stage and Multi-scale Retrieval Augmentation for
Enhancing Whisper | Accepted by ICASSP 2025, oral | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | State-of-the-art models like OpenAI's Whisper exhibit strong performance in
multilingual automatic speech recognition (ASR), but they still face challenges
in accurately recognizing diverse subdialects. In this paper, we propose
M2R-whisper, a novel multi-stage and multi-scale retrieval augmentation
approach designed to enhance ASR performance in low-resource settings. Building
on the principles of in-context learning (ICL) and retrieval-augmented
techniques, our method employs sentence-level ICL in the pre-processing stage
to harness contextual information, while integrating token-level k-Nearest
Neighbors (kNN) retrieval as a post-processing step to further refine the final
output distribution. By synergistically combining sentence-level and
token-level retrieval strategies, M2R-whisper effectively mitigates various
types of recognition errors. Experiments conducted on Mandarin and subdialect
datasets, including AISHELL-1 and KeSpeech, demonstrate substantial
improvements in ASR accuracy, all achieved without any parameter updates.
| [
{
"version": "v1",
"created": "Wed, 18 Sep 2024 11:35:55 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Dec 2024 03:04:54 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 05:22:58 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhou",
"Jiaming",
""
],
[
"Zhao",
"Shiwan",
""
],
[
"He",
"Jiabei",
""
],
[
"Wang",
"Hui",
""
],
[
"Zeng",
"Wenjia",
""
],
[
"Chen",
"Yong",
""
],
[
"Sun",
"Haoqin",
""
],
[
"Kong",
"Aobo",
""
],
[
"Qin",
"Yong",
""
]
]
| TITLE: M2R-Whisper: Multi-stage and Multi-scale Retrieval Augmentation for
Enhancing Whisper
ABSTRACT: State-of-the-art models like OpenAI's Whisper exhibit strong performance in
multilingual automatic speech recognition (ASR), but they still face challenges
in accurately recognizing diverse subdialects. In this paper, we propose
M2R-whisper, a novel multi-stage and multi-scale retrieval augmentation
approach designed to enhance ASR performance in low-resource settings. Building
on the principles of in-context learning (ICL) and retrieval-augmented
techniques, our method employs sentence-level ICL in the pre-processing stage
to harness contextual information, while integrating token-level k-Nearest
Neighbors (kNN) retrieval as a post-processing step to further refine the final
output distribution. By synergistically combining sentence-level and
token-level retrieval strategies, M2R-whisper effectively mitigates various
types of recognition errors. Experiments conducted on Mandarin and subdialect
datasets, including AISHELL-1 and KeSpeech, demonstrate substantial
improvements in ASR accuracy, all achieved without any parameter updates.
| no_new_dataset | 0.942665 |
2409.14572 | Hongchen Wang | Hongchen Wang, Kangming Li, Scott Ramsay, Yao Fehlis, Edward Kim, and
Jason Hattrick-Simpers | Evaluating the Performance and Robustness of LLMs in Materials Science
Q&A and Property Predictions | null | null | null | null | cs.CL cond-mat.mtrl-sci cs.AI cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Large Language Models (LLMs) have the potential to revolutionize scientific
research, yet their robustness and reliability in domain-specific applications
remain insufficiently explored. In this study, we evaluate the performance and
robustness of LLMs for materials science, focusing on domain-specific question
answering and materials property prediction across diverse real-world and
adversarial conditions. Three distinct datasets are used in this study: 1) a
set of multiple-choice questions from undergraduate-level materials science
courses, 2) a dataset including various steel compositions and yield strengths,
and 3) a band gap dataset, containing textual descriptions of material crystal
structures and band gap values. The performance of LLMs is assessed using
various prompting strategies, including zero-shot chain-of-thought, expert
prompting, and few-shot in-context learning. The robustness of these models is
tested against various forms of 'noise', ranging from realistic disturbances to
intentionally adversarial manipulations, to evaluate their resilience and
reliability under real-world conditions. Additionally, the study showcases
unique phenomena of LLMs during predictive tasks, such as mode collapse
behavior when the proximity of prompt examples is altered and performance
recovery from train/test mismatch. The findings aim to provide informed
skepticism for the broad use of LLMs in materials science and to inspire
advancements that enhance their robustness and reliability for practical
applications.
| [
{
"version": "v1",
"created": "Sun, 22 Sep 2024 19:31:16 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 22:03:26 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Wang",
"Hongchen",
""
],
[
"Li",
"Kangming",
""
],
[
"Ramsay",
"Scott",
""
],
[
"Fehlis",
"Yao",
""
],
[
"Kim",
"Edward",
""
],
[
"Hattrick-Simpers",
"Jason",
""
]
]
| TITLE: Evaluating the Performance and Robustness of LLMs in Materials Science
Q&A and Property Predictions
ABSTRACT: Large Language Models (LLMs) have the potential to revolutionize scientific
research, yet their robustness and reliability in domain-specific applications
remain insufficiently explored. In this study, we evaluate the performance and
robustness of LLMs for materials science, focusing on domain-specific question
answering and materials property prediction across diverse real-world and
adversarial conditions. Three distinct datasets are used in this study: 1) a
set of multiple-choice questions from undergraduate-level materials science
courses, 2) a dataset including various steel compositions and yield strengths,
and 3) a band gap dataset, containing textual descriptions of material crystal
structures and band gap values. The performance of LLMs is assessed using
various prompting strategies, including zero-shot chain-of-thought, expert
prompting, and few-shot in-context learning. The robustness of these models is
tested against various forms of 'noise', ranging from realistic disturbances to
intentionally adversarial manipulations, to evaluate their resilience and
reliability under real-world conditions. Additionally, the study showcases
unique phenomena of LLMs during predictive tasks, such as mode collapse
behavior when the proximity of prompt examples is altered and performance
recovery from train/test mismatch. The findings aim to provide informed
skepticism for the broad use of LLMs in materials science and to inspire
advancements that enhance their robustness and reliability for practical
applications.
| new_dataset | 0.963369 |
2409.15949 | Adithi Satish | Danqing Chen, Adithi Satish, Rasul Khanbayov, Carolin M. Schuster and
Georg Groh | Tuning Into Bias: A Computational Study of Gender Bias in Song Lyrics | Accepted to be presented at the 9th Joint SIGHUM Workshop on
Computational Linguistics for Cultural Heritage, Social Sciences, Humanities
and Literature, co-located with NAACL 2025; also accepted and presented as
working paper at the SBP-BRiMS 2024 (see
https://sbp-brims.org/2024/papers/working-papers/Chen_SBP-BRiMS2024_Final_31.pdf
) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The application of text mining methods is becoming increasingly prevalent,
particularly within Humanities and Computational Social Sciences, as well as in
a broader range of disciplines. This paper presents an analysis of gender bias
in English song lyrics using topic modeling and bias measurement techniques.
Leveraging BERTopic, we cluster a dataset of 537,553 English songs into
distinct topics and analyze their temporal evolution. Our results reveal a
significant thematic shift in song lyrics over time, transitioning from
romantic themes to a heightened focus on the sexualization of women.
Additionally, we observe a substantial prevalence of profanity and misogynistic
content across various topics, with a particularly high concentration in the
largest thematic cluster. To further analyse gender bias across topics and
genres in a quantitative way, we employ the Single Category Word Embedding
Association Test (SC-WEAT) to calculate bias scores for word embeddings trained
on the most prominent topics as well as individual genres. The results indicate
a consistent male bias in words associated with intelligence and strength,
while appearance and weakness words show a female bias. Further analysis
highlights variations in these biases across topics, illustrating the interplay
between thematic content and gender stereotypes in song lyrics.
| [
{
"version": "v1",
"created": "Tue, 24 Sep 2024 10:24:53 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 20:54:07 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chen",
"Danqing",
""
],
[
"Satish",
"Adithi",
""
],
[
"Khanbayov",
"Rasul",
""
],
[
"Schuster",
"Carolin M.",
""
],
[
"Groh",
"Georg",
""
]
]
| TITLE: Tuning Into Bias: A Computational Study of Gender Bias in Song Lyrics
ABSTRACT: The application of text mining methods is becoming increasingly prevalent,
particularly within Humanities and Computational Social Sciences, as well as in
a broader range of disciplines. This paper presents an analysis of gender bias
in English song lyrics using topic modeling and bias measurement techniques.
Leveraging BERTopic, we cluster a dataset of 537,553 English songs into
distinct topics and analyze their temporal evolution. Our results reveal a
significant thematic shift in song lyrics over time, transitioning from
romantic themes to a heightened focus on the sexualization of women.
Additionally, we observe a substantial prevalence of profanity and misogynistic
content across various topics, with a particularly high concentration in the
largest thematic cluster. To further analyse gender bias across topics and
genres in a quantitative way, we employ the Single Category Word Embedding
Association Test (SC-WEAT) to calculate bias scores for word embeddings trained
on the most prominent topics as well as individual genres. The results indicate
a consistent male bias in words associated with intelligence and strength,
while appearance and weakness words show a female bias. Further analysis
highlights variations in these biases across topics, illustrating the interplay
between thematic content and gender stereotypes in song lyrics.
| no_new_dataset | 0.945197 |
2410.01425 | Yingdong Hu | Yingdong Hu, Zhening Liu, Jiawei Shao, Zehong Lin, Jun Zhang | EVA-Gaussian: 3D Gaussian-based Real-time Human Novel View Synthesis
under Diverse Multi-view Camera Settings | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Feed-forward based 3D Gaussian Splatting methods have demonstrated
exceptional capability in real-time novel view synthesis for human models.
However, current approaches are confined to either dense viewpoint
configurations or restricted image resolutions. These limitations hinder their
flexibility in free-viewpoint rendering across a wide range of camera view
angle discrepancies, and also restrict their ability to recover fine-grained
human details in real time using commonly available GPUs. To address these
challenges, we propose a novel pipeline named EVA-Gaussian for 3D human novel
view synthesis across diverse multi-view camera settings. Specifically, we
first design an Efficient Cross-View Attention (EVA) module to effectively fuse
cross-view information under high resolution inputs and sparse view settings,
while minimizing temporal and computational overhead. Additionally, we
introduce a feature refinement mechianism to predict the attributes of the 3D
Gaussians and assign a feature value to each Gaussian, enabling the correction
of artifacts caused by geometric inaccuracies in position estimation and
enhancing overall visual fidelity. Experimental results on the THuman2.0 and
THumansit datasets showcase the superiority of EVA-Gaussian in rendering
quality across diverse camera settings. Project page:
https://zhenliuzju.github.io/huyingdong/EVA-Gaussian.
| [
{
"version": "v1",
"created": "Wed, 2 Oct 2024 11:23:08 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 12:14:39 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Hu",
"Yingdong",
""
],
[
"Liu",
"Zhening",
""
],
[
"Shao",
"Jiawei",
""
],
[
"Lin",
"Zehong",
""
],
[
"Zhang",
"Jun",
""
]
]
| TITLE: EVA-Gaussian: 3D Gaussian-based Real-time Human Novel View Synthesis
under Diverse Multi-view Camera Settings
ABSTRACT: Feed-forward based 3D Gaussian Splatting methods have demonstrated
exceptional capability in real-time novel view synthesis for human models.
However, current approaches are confined to either dense viewpoint
configurations or restricted image resolutions. These limitations hinder their
flexibility in free-viewpoint rendering across a wide range of camera view
angle discrepancies, and also restrict their ability to recover fine-grained
human details in real time using commonly available GPUs. To address these
challenges, we propose a novel pipeline named EVA-Gaussian for 3D human novel
view synthesis across diverse multi-view camera settings. Specifically, we
first design an Efficient Cross-View Attention (EVA) module to effectively fuse
cross-view information under high resolution inputs and sparse view settings,
while minimizing temporal and computational overhead. Additionally, we
introduce a feature refinement mechianism to predict the attributes of the 3D
Gaussians and assign a feature value to each Gaussian, enabling the correction
of artifacts caused by geometric inaccuracies in position estimation and
enhancing overall visual fidelity. Experimental results on the THuman2.0 and
THumansit datasets showcase the superiority of EVA-Gaussian in rendering
quality across diverse camera settings. Project page:
https://zhenliuzju.github.io/huyingdong/EVA-Gaussian.
| no_new_dataset | 0.951997 |
2410.02056 | Sreyan Ghosh | Sreyan Ghosh and Sonal Kumar and Zhifeng Kong and Rafael Valle and
Bryan Catanzaro and Dinesh Manocha | Synthio: Augmenting Small-Scale Audio Classification Datasets with
Synthetic Data | Accepted at ICLR 2025. Code and Checkpoints available here:
https://github.com/Sreyan88/Synthio | null | null | null | eess.AS cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | We present Synthio, a novel approach for augmenting small-scale audio
classification datasets with synthetic data. Our goal is to improve audio
classification accuracy with limited labeled data. Traditional data
augmentation techniques, which apply artificial transformations (e.g., adding
random noise or masking segments), struggle to create data that captures the
true diversity present in real-world audios. To address this shortcoming, we
propose to augment the dataset with synthetic audio generated from
text-to-audio (T2A) diffusion models. However, synthesizing effective
augmentations is challenging because not only should the generated data be
acoustically consistent with the underlying small-scale dataset, but they
should also have sufficient compositional diversity. To overcome the first
challenge, we align the generations of the T2A model with the small-scale
dataset using preference optimization. This ensures that the acoustic
characteristics of the generated data remain consistent with the small-scale
dataset. To address the second challenge, we propose a novel caption generation
technique that leverages the reasoning capabilities of Large Language Models to
(1) generate diverse and meaningful audio captions and (2) iteratively refine
their quality. The generated captions are then used to prompt the aligned T2A
model. We extensively evaluate Synthio on ten datasets and four simulated
limited-data settings. Results indicate our method consistently outperforms all
baselines by 0.1%-39% using a T2A model trained only on weakly-captioned
AudioSet.
| [
{
"version": "v1",
"created": "Wed, 2 Oct 2024 22:05:36 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 00:25:08 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Ghosh",
"Sreyan",
""
],
[
"Kumar",
"Sonal",
""
],
[
"Kong",
"Zhifeng",
""
],
[
"Valle",
"Rafael",
""
],
[
"Catanzaro",
"Bryan",
""
],
[
"Manocha",
"Dinesh",
""
]
]
| TITLE: Synthio: Augmenting Small-Scale Audio Classification Datasets with
Synthetic Data
ABSTRACT: We present Synthio, a novel approach for augmenting small-scale audio
classification datasets with synthetic data. Our goal is to improve audio
classification accuracy with limited labeled data. Traditional data
augmentation techniques, which apply artificial transformations (e.g., adding
random noise or masking segments), struggle to create data that captures the
true diversity present in real-world audios. To address this shortcoming, we
propose to augment the dataset with synthetic audio generated from
text-to-audio (T2A) diffusion models. However, synthesizing effective
augmentations is challenging because not only should the generated data be
acoustically consistent with the underlying small-scale dataset, but they
should also have sufficient compositional diversity. To overcome the first
challenge, we align the generations of the T2A model with the small-scale
dataset using preference optimization. This ensures that the acoustic
characteristics of the generated data remain consistent with the small-scale
dataset. To address the second challenge, we propose a novel caption generation
technique that leverages the reasoning capabilities of Large Language Models to
(1) generate diverse and meaningful audio captions and (2) iteratively refine
their quality. The generated captions are then used to prompt the aligned T2A
model. We extensively evaluate Synthio on ten datasets and four simulated
limited-data settings. Results indicate our method consistently outperforms all
baselines by 0.1%-39% using a T2A model trained only on weakly-captioned
AudioSet.
| no_new_dataset | 0.951006 |
2410.05440 | Zihao Zhou | Zihao Zhou, Rose Yu | Can LLMs Understand Time Series Anomalies? | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have gained popularity in time series
forecasting, but their potential for anomaly detection remains largely
unexplored. Our study investigates whether LLMs can understand and detect
anomalies in time series data, focusing on zero-shot and few-shot scenarios.
Inspired by conjectures about LLMs' behavior from time series forecasting
research, we formulate key hypotheses about LLMs' capabilities in time series
anomaly detection. We design and conduct principled experiments to test each of
these hypotheses. Our investigation reveals several surprising findings about
LLMs for time series: (1) LLMs understand time series better as images rather
than as text, (2) LLMs do not demonstrate enhanced performance when prompted to
engage in explicit reasoning about time series analysis. (3) Contrary to common
beliefs, LLMs' understanding of time series does not stem from their repetition
biases or arithmetic abilities. (4) LLMs' behaviors and performance in time
series analysis vary significantly across different models. This study provides
the first comprehensive analysis of contemporary LLM capabilities in time
series anomaly detection. Our results suggest that while LLMs can understand
trivial time series anomalies, we have no evidence that they can understand
more subtle real-world anomalies. Many common conjectures based on their
reasoning capabilities do not hold. All synthetic dataset generators, final
prompts, and evaluation scripts have been made available in
https://github.com/rose-stl-lab/anomllm.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2024 19:16:02 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Oct 2024 23:32:50 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 18:04:52 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhou",
"Zihao",
""
],
[
"Yu",
"Rose",
""
]
]
| TITLE: Can LLMs Understand Time Series Anomalies?
ABSTRACT: Large Language Models (LLMs) have gained popularity in time series
forecasting, but their potential for anomaly detection remains largely
unexplored. Our study investigates whether LLMs can understand and detect
anomalies in time series data, focusing on zero-shot and few-shot scenarios.
Inspired by conjectures about LLMs' behavior from time series forecasting
research, we formulate key hypotheses about LLMs' capabilities in time series
anomaly detection. We design and conduct principled experiments to test each of
these hypotheses. Our investigation reveals several surprising findings about
LLMs for time series: (1) LLMs understand time series better as images rather
than as text, (2) LLMs do not demonstrate enhanced performance when prompted to
engage in explicit reasoning about time series analysis. (3) Contrary to common
beliefs, LLMs' understanding of time series does not stem from their repetition
biases or arithmetic abilities. (4) LLMs' behaviors and performance in time
series analysis vary significantly across different models. This study provides
the first comprehensive analysis of contemporary LLM capabilities in time
series anomaly detection. Our results suggest that while LLMs can understand
trivial time series anomalies, we have no evidence that they can understand
more subtle real-world anomalies. Many common conjectures based on their
reasoning capabilities do not hold. All synthetic dataset generators, final
prompts, and evaluation scripts have been made available in
https://github.com/rose-stl-lab/anomllm.
| no_new_dataset | 0.75183 |
2410.05628 | Jeongeun Park | Jeongeun Park, Sungjoon Choi, Sangdoo Yun | A Unified Framework for Motion Reasoning and Generation in Human
Interaction | https://vim-motion-language.github.io/ | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in large language models (LLMs) have significantly
improved their ability to generate natural and contextually relevant text,
enabling more human-like AI interactions. However, generating and understanding
interactive human-like motion, where multiple individuals engage in coordinated
movements, remains challenging due to the complexity of modeling these
interactions. Additionally, a unified and versatile model is needed to handle
diverse interactive scenarios, such as chat systems that dynamically adapt to
user instructions and assigned roles. To address these challenges, we introduce
VIM, the Versatile Interactive Motion-language model, which integrates both
language and motion modalities to effectively understand, generate, and control
interactive motions in multi-turn conversational contexts. Unlike previous
studies that primarily focus on uni-directional tasks such as text-to-motion or
motion-to-text, VIM employs a unified architecture capable of simultaneously
understanding and generating both motion and text modalities. Given the absence
of an appropriate dataset to support this task, we introduce Inter-MT2, a
large-scale instruction-tuning dataset containing 82.7K multi-turn interactive
motion instructions, covering 153K interactive motion samples. Inter-MT2 spans
diverse instructional scenarios, including motion editing, question answering,
and story generation, leveraging off-the-shelf large language models and motion
diffusion models to construct a broad set of interactive motion instructions.
We extensively evaluate the versatility of VIM across multiple interactive
motion-related tasks, including motion-to-text, text-to-motion, reaction
generation, motion editing, and reasoning about motion sequences.
| [
{
"version": "v1",
"created": "Tue, 8 Oct 2024 02:23:53 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Oct 2024 11:22:39 GMT"
},
{
"version": "v3",
"created": "Thu, 24 Oct 2024 12:47:56 GMT"
},
{
"version": "v4",
"created": "Tue, 11 Mar 2025 15:18:47 GMT"
},
{
"version": "v5",
"created": "Wed, 12 Mar 2025 05:54:44 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Park",
"Jeongeun",
""
],
[
"Choi",
"Sungjoon",
""
],
[
"Yun",
"Sangdoo",
""
]
]
| TITLE: A Unified Framework for Motion Reasoning and Generation in Human
Interaction
ABSTRACT: Recent advancements in large language models (LLMs) have significantly
improved their ability to generate natural and contextually relevant text,
enabling more human-like AI interactions. However, generating and understanding
interactive human-like motion, where multiple individuals engage in coordinated
movements, remains challenging due to the complexity of modeling these
interactions. Additionally, a unified and versatile model is needed to handle
diverse interactive scenarios, such as chat systems that dynamically adapt to
user instructions and assigned roles. To address these challenges, we introduce
VIM, the Versatile Interactive Motion-language model, which integrates both
language and motion modalities to effectively understand, generate, and control
interactive motions in multi-turn conversational contexts. Unlike previous
studies that primarily focus on uni-directional tasks such as text-to-motion or
motion-to-text, VIM employs a unified architecture capable of simultaneously
understanding and generating both motion and text modalities. Given the absence
of an appropriate dataset to support this task, we introduce Inter-MT2, a
large-scale instruction-tuning dataset containing 82.7K multi-turn interactive
motion instructions, covering 153K interactive motion samples. Inter-MT2 spans
diverse instructional scenarios, including motion editing, question answering,
and story generation, leveraging off-the-shelf large language models and motion
diffusion models to construct a broad set of interactive motion instructions.
We extensively evaluate the versatility of VIM across multiple interactive
motion-related tasks, including motion-to-text, text-to-motion, reaction
generation, motion editing, and reasoning about motion sequences.
| new_dataset | 0.961965 |
2410.08917 | Till Raphael Saenger | Till Raphael Saenger, Musashi Hinck, Justin Grimmer and Brandon M.
Stewart | AutoPersuade: A Framework for Evaluating and Explaining Persuasive
Arguments | Published in Proceedings of EMNLP 2024. The official version is
available in the ACL Anthology at
https://aclanthology.org/2024.emnlp-main.913/ | null | 10.18653/v1/2024.emnlp-main.913 | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We introduce AutoPersuade, a three-part framework for constructing persuasive
messages. First, we curate a large dataset of arguments with human evaluations.
Next, we develop a novel topic model to identify argument features that
influence persuasiveness. Finally, we use this model to predict the
effectiveness of new arguments and assess the causal impact of different
components to provide explanations. We validate AutoPersuade through an
experimental study on arguments for veganism, demonstrating its effectiveness
with human studies and out-of-sample predictions.
| [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 15:46:05 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 19:56:48 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Saenger",
"Till Raphael",
""
],
[
"Hinck",
"Musashi",
""
],
[
"Grimmer",
"Justin",
""
],
[
"Stewart",
"Brandon M.",
""
]
]
| TITLE: AutoPersuade: A Framework for Evaluating and Explaining Persuasive
Arguments
ABSTRACT: We introduce AutoPersuade, a three-part framework for constructing persuasive
messages. First, we curate a large dataset of arguments with human evaluations.
Next, we develop a novel topic model to identify argument features that
influence persuasiveness. Finally, we use this model to predict the
effectiveness of new arguments and assess the causal impact of different
components to provide explanations. We validate AutoPersuade through an
experimental study on arguments for veganism, demonstrating its effectiveness
with human studies and out-of-sample predictions.
| new_dataset | 0.73431 |
2410.10182 | Javier Mar\'in | Javier Mar\'in | Hamiltonian Neural Networks for Robust Out-of-Time Credit Scoring | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel credit scoring approach using neural networks to
address class imbalance and out-of-time prediction challenges. We develop a
specific optimizer and loss function inspired by Hamiltonian mechanics that
better captures credit risk dynamics. Testing on the Freddie Mac Single-Family
Loan-Level Dataset shows our model achieves superior discriminative power (AUC)
in out-of-time scenarios compared to conventional methods. The approach has
consistent performance between in-sample and future test sets, maintaining
reliability across time periods. This interdisciplinary method spans physical
systems theory and financial risk management, offering practical advantages for
long-term model stability.
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 06:08:26 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 06:03:20 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Marín",
"Javier",
""
]
]
| TITLE: Hamiltonian Neural Networks for Robust Out-of-Time Credit Scoring
ABSTRACT: This paper presents a novel credit scoring approach using neural networks to
address class imbalance and out-of-time prediction challenges. We develop a
specific optimizer and loss function inspired by Hamiltonian mechanics that
better captures credit risk dynamics. Testing on the Freddie Mac Single-Family
Loan-Level Dataset shows our model achieves superior discriminative power (AUC)
in out-of-time scenarios compared to conventional methods. The approach has
consistent performance between in-sample and future test sets, maintaining
reliability across time periods. This interdisciplinary method spans physical
systems theory and financial risk management, offering practical advantages for
long-term model stability.
| no_new_dataset | 0.950319 |
2410.10782 | Eduardo R. Corral-Soto | Eduardo R. Corral-Soto, Yang Liu, Tongtong Cao, Yuan Ren, Liu Bingbing | 3DArticCyclists: Generating Synthetic Articulated 8D Pose-Controllable
Cyclist Data for Computer Vision Applications | null | null | null | null | cs.CV cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In Autonomous Driving (AD) Perception, cyclists are considered
safety-critical scene objects. Commonly used publicly-available AD datasets
typically contain large amounts of car and vehicle object instances but a low
number of cyclist instances, usually with limited appearance and pose
diversity. This cyclist training data scarcity problem not only limits the
generalization of deep-learning perception models for cyclist semantic
segmentation, pose estimation, and cyclist crossing intention prediction, but
also limits research on new cyclist-related tasks such as fine-grained cyclist
pose estimation and spatio-temporal analysis under complex interactions between
humans and articulated objects. To address this data scarcity problem, in this
paper we propose a framework to generate synthetic dynamic 3D cyclist data
assets that can be used to generate training data for different tasks. In our
framework, we designed a methodology for creating a new part-based multi-view
articulated synthetic 3D bicycle dataset that we call 3DArticBikes that we use
to train a 3D Gaussian Splatting (3DGS)-based reconstruction and image
rendering method. We then propose a parametric bicycle 3DGS composition model
to assemble 8-DoF pose-controllable 3D bicycles. Finally, using dynamic
information from cyclist videos, we build a complete synthetic dynamic 3D
cyclist (rider pedaling a bicycle) by re-posing a selectable synthetic 3D
person, while automatically placing the rider onto one of our new articulated
3D bicycles using a proposed 3D Keypoint optimization-based Inverse Kinematics
pose refinement. We present both, qualitative and quantitative results where we
compare our generated cyclists against those from a recent stable
diffusion-based method.
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 17:50:47 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 01:15:52 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Corral-Soto",
"Eduardo R.",
""
],
[
"Liu",
"Yang",
""
],
[
"Cao",
"Tongtong",
""
],
[
"Ren",
"Yuan",
""
],
[
"Bingbing",
"Liu",
""
]
]
| TITLE: 3DArticCyclists: Generating Synthetic Articulated 8D Pose-Controllable
Cyclist Data for Computer Vision Applications
ABSTRACT: In Autonomous Driving (AD) Perception, cyclists are considered
safety-critical scene objects. Commonly used publicly-available AD datasets
typically contain large amounts of car and vehicle object instances but a low
number of cyclist instances, usually with limited appearance and pose
diversity. This cyclist training data scarcity problem not only limits the
generalization of deep-learning perception models for cyclist semantic
segmentation, pose estimation, and cyclist crossing intention prediction, but
also limits research on new cyclist-related tasks such as fine-grained cyclist
pose estimation and spatio-temporal analysis under complex interactions between
humans and articulated objects. To address this data scarcity problem, in this
paper we propose a framework to generate synthetic dynamic 3D cyclist data
assets that can be used to generate training data for different tasks. In our
framework, we designed a methodology for creating a new part-based multi-view
articulated synthetic 3D bicycle dataset that we call 3DArticBikes that we use
to train a 3D Gaussian Splatting (3DGS)-based reconstruction and image
rendering method. We then propose a parametric bicycle 3DGS composition model
to assemble 8-DoF pose-controllable 3D bicycles. Finally, using dynamic
information from cyclist videos, we build a complete synthetic dynamic 3D
cyclist (rider pedaling a bicycle) by re-posing a selectable synthetic 3D
person, while automatically placing the rider onto one of our new articulated
3D bicycles using a proposed 3D Keypoint optimization-based Inverse Kinematics
pose refinement. We present both, qualitative and quantitative results where we
compare our generated cyclists against those from a recent stable
diffusion-based method.
| new_dataset | 0.963506 |
2410.12459 | Artem Moskalev | Mehdi Yazdani-Jahromi and Mangal Prakash and Tommaso Mansi and Artem
Moskalev and Rui Liao | HELM: Hierarchical Encoding for mRNA Language Modeling | null | null | null | null | cs.LG cs.CE | http://creativecommons.org/licenses/by/4.0/ | Messenger RNA (mRNA) plays a crucial role in protein synthesis, with its
codon structure directly impacting biological properties. While Language Models
(LMs) have shown promise in analyzing biological sequences, existing approaches
fail to account for the hierarchical nature of mRNA's codon structure. We
introduce Hierarchical Encoding for mRNA Language Modeling (HELM), a novel
pre-training strategy that incorporates codon-level hierarchical structure into
language model training. HELM modulates the loss function based on codon
synonymity, aligning the model's learning process with the biological reality
of mRNA sequences. We evaluate HELM on diverse mRNA datasets and tasks,
demonstrating that HELM outperforms standard language model pre-training as
well as existing foundation model baselines on seven diverse downstream
property prediction tasks and an antibody region annotation tasks on average by
around 8%. Additionally, HELM enhances the generative capabilities of language
model, producing diverse mRNA sequences that better align with the underlying
true data distribution compared to non-hierarchical baselines.
| [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 11:16:47 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 10:51:14 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Yazdani-Jahromi",
"Mehdi",
""
],
[
"Prakash",
"Mangal",
""
],
[
"Mansi",
"Tommaso",
""
],
[
"Moskalev",
"Artem",
""
],
[
"Liao",
"Rui",
""
]
]
| TITLE: HELM: Hierarchical Encoding for mRNA Language Modeling
ABSTRACT: Messenger RNA (mRNA) plays a crucial role in protein synthesis, with its
codon structure directly impacting biological properties. While Language Models
(LMs) have shown promise in analyzing biological sequences, existing approaches
fail to account for the hierarchical nature of mRNA's codon structure. We
introduce Hierarchical Encoding for mRNA Language Modeling (HELM), a novel
pre-training strategy that incorporates codon-level hierarchical structure into
language model training. HELM modulates the loss function based on codon
synonymity, aligning the model's learning process with the biological reality
of mRNA sequences. We evaluate HELM on diverse mRNA datasets and tasks,
demonstrating that HELM outperforms standard language model pre-training as
well as existing foundation model baselines on seven diverse downstream
property prediction tasks and an antibody region annotation tasks on average by
around 8%. Additionally, HELM enhances the generative capabilities of language
model, producing diverse mRNA sequences that better align with the underlying
true data distribution compared to non-hierarchical baselines.
| no_new_dataset | 0.952131 |
2410.14634 | Sandeep Nagar Mr. | Sandeep Nagar, Girish Varma | Parallel Backpropagation for Inverse of a Convolution with Application
to Normalizing Flows | 28th International Conference on Artificial Intelligence and
Statistics (AISTATS) 2025 | null | null | null | cs.CV cs.LG cs.MM math.PR | http://creativecommons.org/licenses/by/4.0/ | The inverse of an invertible convolution is an important operation that comes
up in Normalizing Flows, Image Deblurring, etc. The naive algorithm for
backpropagation of this operation using Gaussian elimination has running time
$O(n^3)$ where $n$ is the number of pixels in the image. We give a fast
parallel backpropagation algorithm with running time $O(\sqrt{n})$ for a square
image and provide a GPU implementation of the same. Inverse of Convolutions are
usually used in Normalizing Flows in the sampling pass, making them slow. We
propose to use the Inverse of Convolutions in the forward (image to latent
vector) pass of the Normalizing flow. Since the sampling pass is the inverse of
the forward pass, it will use convolutions only, resulting in efficient
sampling times. We use our parallel backpropagation algorithm to optimize the
inverse of the convolution layer, resulting in fast training times. We
implement this approach in various Normalizing Flow backbones, resulting in our
Inverse-Flow models. We benchmark Inverse-Flow on standard datasets and show
significantly improved sampling times with similar bits per dimension compared
to previous models.
| [
{
"version": "v1",
"created": "Fri, 18 Oct 2024 17:35:33 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Feb 2025 20:58:09 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 06:28:50 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Nagar",
"Sandeep",
""
],
[
"Varma",
"Girish",
""
]
]
| TITLE: Parallel Backpropagation for Inverse of a Convolution with Application
to Normalizing Flows
ABSTRACT: The inverse of an invertible convolution is an important operation that comes
up in Normalizing Flows, Image Deblurring, etc. The naive algorithm for
backpropagation of this operation using Gaussian elimination has running time
$O(n^3)$ where $n$ is the number of pixels in the image. We give a fast
parallel backpropagation algorithm with running time $O(\sqrt{n})$ for a square
image and provide a GPU implementation of the same. Inverse of Convolutions are
usually used in Normalizing Flows in the sampling pass, making them slow. We
propose to use the Inverse of Convolutions in the forward (image to latent
vector) pass of the Normalizing flow. Since the sampling pass is the inverse of
the forward pass, it will use convolutions only, resulting in efficient
sampling times. We use our parallel backpropagation algorithm to optimize the
inverse of the convolution layer, resulting in fast training times. We
implement this approach in various Normalizing Flow backbones, resulting in our
Inverse-Flow models. We benchmark Inverse-Flow on standard datasets and show
significantly improved sampling times with similar bits per dimension compared
to previous models.
| no_new_dataset | 0.949716 |
2410.17448 | Tyler Josephson | Samiha Sharlin, Tyler R. Josephson | In Context Learning and Reasoning for Symbolic Regression with Large
Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) are transformer-based machine learning models
that have shown remarkable performance in tasks for which they were not
explicitly trained. Here, we explore the potential of LLMs to perform symbolic
regression -- a machine-learning method for finding simple and accurate
equations from datasets. We prompt GPT-4 to suggest expressions from data,
which are then optimized and evaluated using external Python tools. These
results are fed back to GPT-4, which proposes improved expressions while
optimizing for complexity and loss. Using chain-of-thought prompting, we
instruct GPT-4 to analyze the data, prior expressions, and the scientific
context (expressed in natural language) for each problem before generating new
expressions. We evaluated the workflow in rediscovery of five well-known
scientific equations from experimental data, and on an additional dataset
without a known equation. GPT-4 successfully rediscovered all five equations,
and in general, performed better when prompted to use a scratchpad and consider
scientific context. We demonstrate how strategic prompting improves the model's
performance and how the natural language interface simplifies integrating
theory with data. We also observe how theory can sometimes offset noisy data
and, in other cases, data can make up for poor context. Although this approach
does not outperform established SR programs where target equations are more
complex, LLMs can nonetheless iterate toward improved solutions while following
instructions and incorporating scientific context in natural language.
| [
{
"version": "v1",
"created": "Tue, 22 Oct 2024 21:50:52 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 13:14:22 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Sharlin",
"Samiha",
""
],
[
"Josephson",
"Tyler R.",
""
]
]
| TITLE: In Context Learning and Reasoning for Symbolic Regression with Large
Language Models
ABSTRACT: Large Language Models (LLMs) are transformer-based machine learning models
that have shown remarkable performance in tasks for which they were not
explicitly trained. Here, we explore the potential of LLMs to perform symbolic
regression -- a machine-learning method for finding simple and accurate
equations from datasets. We prompt GPT-4 to suggest expressions from data,
which are then optimized and evaluated using external Python tools. These
results are fed back to GPT-4, which proposes improved expressions while
optimizing for complexity and loss. Using chain-of-thought prompting, we
instruct GPT-4 to analyze the data, prior expressions, and the scientific
context (expressed in natural language) for each problem before generating new
expressions. We evaluated the workflow in rediscovery of five well-known
scientific equations from experimental data, and on an additional dataset
without a known equation. GPT-4 successfully rediscovered all five equations,
and in general, performed better when prompted to use a scratchpad and consider
scientific context. We demonstrate how strategic prompting improves the model's
performance and how the natural language interface simplifies integrating
theory with data. We also observe how theory can sometimes offset noisy data
and, in other cases, data can make up for poor context. Although this approach
does not outperform established SR programs where target equations are more
complex, LLMs can nonetheless iterate toward improved solutions while following
instructions and incorporating scientific context in natural language.
| no_new_dataset | 0.947721 |
2410.18444 | ChaeHun Park | ChaeHun Park, Hojun Cho, Jaegul Choo | Evaluating Automatic Speech Recognition Systems for Korean
Meteorological Experts | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This paper explores integrating Automatic Speech Recognition (ASR) into
natural language query systems to improve weather forecasting efficiency for
Korean meteorologists. We address challenges in developing ASR systems for the
Korean weather domain, specifically specialized vocabulary and Korean
linguistic intricacies. To tackle these issues, we constructed an evaluation
dataset of spoken queries recorded by native Korean speakers. Using this
dataset, we assessed various configurations of a multilingual ASR model family,
identifying performance limitations related to domain-specific terminology. We
then implemented a simple text-to-speech-based data augmentation method, which
improved the recognition of specialized terms while maintaining general-domain
performance. Our contributions include creating a domain-specific dataset,
comprehensive ASR model evaluations, and an effective augmentation technique.
We believe our work provides a foundation for future advancements in ASR for
the Korean weather forecasting domain.
| [
{
"version": "v1",
"created": "Thu, 24 Oct 2024 05:40:07 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 13:18:04 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Park",
"ChaeHun",
""
],
[
"Cho",
"Hojun",
""
],
[
"Choo",
"Jaegul",
""
]
]
| TITLE: Evaluating Automatic Speech Recognition Systems for Korean
Meteorological Experts
ABSTRACT: This paper explores integrating Automatic Speech Recognition (ASR) into
natural language query systems to improve weather forecasting efficiency for
Korean meteorologists. We address challenges in developing ASR systems for the
Korean weather domain, specifically specialized vocabulary and Korean
linguistic intricacies. To tackle these issues, we constructed an evaluation
dataset of spoken queries recorded by native Korean speakers. Using this
dataset, we assessed various configurations of a multilingual ASR model family,
identifying performance limitations related to domain-specific terminology. We
then implemented a simple text-to-speech-based data augmentation method, which
improved the recognition of specialized terms while maintaining general-domain
performance. Our contributions include creating a domain-specific dataset,
comprehensive ASR model evaluations, and an effective augmentation technique.
We believe our work provides a foundation for future advancements in ASR for
the Korean weather forecasting domain.
| new_dataset | 0.958304 |
2410.18469 | Chung En Sun | Chung-En Sun, Xiaodong Liu, Weiwei Yang, Tsui-Wei Weng, Hao Cheng,
Aidan San, Michel Galley, Jianfeng Gao | Iterative Self-Tuning LLMs for Enhanced Jailbreaking Capabilities | Accepted to NAACL 2025 Main (oral) | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent research has shown that Large Language Models (LLMs) are vulnerable to
automated jailbreak attacks, where adversarial suffixes crafted by algorithms
appended to harmful queries bypass safety alignment and trigger unintended
responses. Current methods for generating these suffixes are computationally
expensive and have low Attack Success Rates (ASR), especially against
well-aligned models like Llama2 and Llama3. To overcome these limitations, we
introduce ADV-LLM, an iterative self-tuning process that crafts adversarial
LLMs with enhanced jailbreak ability. Our framework significantly reduces the
computational cost of generating adversarial suffixes while achieving nearly
100\% ASR on various open-source LLMs. Moreover, it exhibits strong attack
transferability to closed-source models, achieving 99\% ASR on GPT-3.5 and 49\%
ASR on GPT-4, despite being optimized solely on Llama3. Beyond improving
jailbreak ability, ADV-LLM provides valuable insights for future safety
alignment research through its ability to generate large datasets for studying
LLM safety.
| [
{
"version": "v1",
"created": "Thu, 24 Oct 2024 06:36:12 GMT"
},
{
"version": "v2",
"created": "Fri, 25 Oct 2024 23:05:59 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 23:26:25 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Sun",
"Chung-En",
""
],
[
"Liu",
"Xiaodong",
""
],
[
"Yang",
"Weiwei",
""
],
[
"Weng",
"Tsui-Wei",
""
],
[
"Cheng",
"Hao",
""
],
[
"San",
"Aidan",
""
],
[
"Galley",
"Michel",
""
],
[
"Gao",
"Jianfeng",
""
]
]
| TITLE: Iterative Self-Tuning LLMs for Enhanced Jailbreaking Capabilities
ABSTRACT: Recent research has shown that Large Language Models (LLMs) are vulnerable to
automated jailbreak attacks, where adversarial suffixes crafted by algorithms
appended to harmful queries bypass safety alignment and trigger unintended
responses. Current methods for generating these suffixes are computationally
expensive and have low Attack Success Rates (ASR), especially against
well-aligned models like Llama2 and Llama3. To overcome these limitations, we
introduce ADV-LLM, an iterative self-tuning process that crafts adversarial
LLMs with enhanced jailbreak ability. Our framework significantly reduces the
computational cost of generating adversarial suffixes while achieving nearly
100\% ASR on various open-source LLMs. Moreover, it exhibits strong attack
transferability to closed-source models, achieving 99\% ASR on GPT-3.5 and 49\%
ASR on GPT-4, despite being optimized solely on Llama3. Beyond improving
jailbreak ability, ADV-LLM provides valuable insights for future safety
alignment research through its ability to generate large datasets for studying
LLM safety.
| no_new_dataset | 0.944587 |
2410.18857 | Sanghyuk Chun | Sanghyuk Chun and Wonjae Kim and Song Park and Sangdoo Yun | Probabilistic Language-Image Pre-Training | Code: https://github.com/naver-ai/prolip HuggingFace Hub:
https://huggingface.co/collections/SanghyukChun/prolip-6712595dfc87fd8597350291
33 pages, 4.8 MB; LongProLIP paper: arXiv:2503.08048 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-language models (VLMs) embed aligned image-text pairs into a joint
space but often rely on deterministic embeddings, assuming a one-to-one
correspondence between images and texts. This oversimplifies real-world
relationships, which are inherently many-to-many, with multiple captions
describing a single image and vice versa. We introduce Probabilistic
Language-Image Pre-training (ProLIP), the first probabilistic VLM pre-trained
on a billion-scale image-text dataset using only probabilistic objectives,
achieving a strong zero-shot capability (e.g., 74.6% ImageNet zero-shot
accuracy with ViT-B/16). ProLIP efficiently estimates uncertainty by an
"uncertainty token" without extra parameters. We also introduce a novel
inclusion loss that enforces distributional inclusion relationships between
image-text pairs and between original and masked inputs. Experiments
demonstrate that, by leveraging uncertainty estimates, ProLIP benefits
downstream tasks and aligns with intuitive notions of uncertainty, e.g.,
shorter texts being more uncertain and more general inputs including specific
ones. Utilizing text uncertainties, we further improve ImageNet accuracy from
74.6% to 75.8% (under a few-shot setting), supporting the practical advantages
of our probabilistic approach. The code is available at
https://github.com/naver-ai/prolip
| [
{
"version": "v1",
"created": "Thu, 24 Oct 2024 15:42:25 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2024 15:20:28 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 14:03:31 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chun",
"Sanghyuk",
""
],
[
"Kim",
"Wonjae",
""
],
[
"Park",
"Song",
""
],
[
"Yun",
"Sangdoo",
""
]
]
| TITLE: Probabilistic Language-Image Pre-Training
ABSTRACT: Vision-language models (VLMs) embed aligned image-text pairs into a joint
space but often rely on deterministic embeddings, assuming a one-to-one
correspondence between images and texts. This oversimplifies real-world
relationships, which are inherently many-to-many, with multiple captions
describing a single image and vice versa. We introduce Probabilistic
Language-Image Pre-training (ProLIP), the first probabilistic VLM pre-trained
on a billion-scale image-text dataset using only probabilistic objectives,
achieving a strong zero-shot capability (e.g., 74.6% ImageNet zero-shot
accuracy with ViT-B/16). ProLIP efficiently estimates uncertainty by an
"uncertainty token" without extra parameters. We also introduce a novel
inclusion loss that enforces distributional inclusion relationships between
image-text pairs and between original and masked inputs. Experiments
demonstrate that, by leveraging uncertainty estimates, ProLIP benefits
downstream tasks and aligns with intuitive notions of uncertainty, e.g.,
shorter texts being more uncertain and more general inputs including specific
ones. Utilizing text uncertainties, we further improve ImageNet accuracy from
74.6% to 75.8% (under a few-shot setting), supporting the practical advantages
of our probabilistic approach. The code is available at
https://github.com/naver-ai/prolip
| no_new_dataset | 0.949389 |
2410.23746 | Runzhe Zhan | Junchao Wu, Runzhe Zhan, Derek F. Wong, Shu Yang, Xinyi Yang, Yulin
Yuan, Lidia S. Chao | DetectRL: Benchmarking LLM-Generated Text Detection in Real-World
Scenarios | Accepted to NeurIPS 2024 Datasets and Benchmarks Track (Camera-Ready) | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Detecting text generated by large language models (LLMs) is of great recent
interest. With zero-shot methods like DetectGPT, detection capabilities have
reached impressive levels. However, the reliability of existing detectors in
real-world applications remains underexplored. In this study, we present a new
benchmark, DetectRL, highlighting that even state-of-the-art (SOTA) detection
techniques still underperformed in this task. We collected human-written
datasets from domains where LLMs are particularly prone to misuse. Using
popular LLMs, we generated data that better aligns with real-world
applications. Unlike previous studies, we employed heuristic rules to create
adversarial LLM-generated text, simulating various prompts usages, human
revisions like word substitutions, and writing noises like spelling mistakes.
Our development of DetectRL reveals the strengths and limitations of current
SOTA detectors. More importantly, we analyzed the potential impact of writing
styles, model types, attack methods, the text lengths, and real-world human
writing factors on different types of detectors. We believe DetectRL could
serve as an effective benchmark for assessing detectors in real-world
scenarios, evolving with advanced attack methods, thus providing more stressful
evaluation to drive the development of more efficient detectors. Data and code
are publicly available at: https://github.com/NLP2CT/DetectRL.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 09:01:25 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Mar 2025 09:06:03 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 10:08:22 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Wu",
"Junchao",
""
],
[
"Zhan",
"Runzhe",
""
],
[
"Wong",
"Derek F.",
""
],
[
"Yang",
"Shu",
""
],
[
"Yang",
"Xinyi",
""
],
[
"Yuan",
"Yulin",
""
],
[
"Chao",
"Lidia S.",
""
]
]
| TITLE: DetectRL: Benchmarking LLM-Generated Text Detection in Real-World
Scenarios
ABSTRACT: Detecting text generated by large language models (LLMs) is of great recent
interest. With zero-shot methods like DetectGPT, detection capabilities have
reached impressive levels. However, the reliability of existing detectors in
real-world applications remains underexplored. In this study, we present a new
benchmark, DetectRL, highlighting that even state-of-the-art (SOTA) detection
techniques still underperformed in this task. We collected human-written
datasets from domains where LLMs are particularly prone to misuse. Using
popular LLMs, we generated data that better aligns with real-world
applications. Unlike previous studies, we employed heuristic rules to create
adversarial LLM-generated text, simulating various prompts usages, human
revisions like word substitutions, and writing noises like spelling mistakes.
Our development of DetectRL reveals the strengths and limitations of current
SOTA detectors. More importantly, we analyzed the potential impact of writing
styles, model types, attack methods, the text lengths, and real-world human
writing factors on different types of detectors. We believe DetectRL could
serve as an effective benchmark for assessing detectors in real-world
scenarios, evolving with advanced attack methods, thus providing more stressful
evaluation to drive the development of more efficient detectors. Data and code
are publicly available at: https://github.com/NLP2CT/DetectRL.
| new_dataset | 0.544873 |
2411.00144 | Chen Zhao | Chen Zhao, Xuan Wang, Tong Zhang, Saqib Javed, Mathieu Salzmann | Self-Ensembling Gaussian Splatting for Few-Shot Novel View Synthesis | null | null | null | null | cs.CV cs.GR | http://creativecommons.org/licenses/by/4.0/ | 3D Gaussian Splatting (3DGS) has demonstrated remarkable effectiveness in
novel view synthesis (NVS). However, 3DGS tends to overfit when trained with
sparse views, limiting its generalization to novel viewpoints. In this paper,
we address this overfitting issue by introducing Self-Ensembling Gaussian
Splatting (SE-GS). We achieve self-ensembling by incorporating an
uncertainty-aware perturbation strategy during training. A
$\mathbf{\Delta}$-model and a $\mathbf{\Sigma}$-model are jointly trained on
the available images. The $\mathbf{\Delta}$-model is dynamically perturbed
based on rendering uncertainty across training steps, generating diverse
perturbed models with negligible computational overhead. Discrepancies between
the $\mathbf{\Sigma}$-model and these perturbed models are minimized throughout
training, forming a robust ensemble of 3DGS models. This ensemble, represented
by the $\mathbf{\Sigma}$-model, is then used to generate novel-view images
during inference. Experimental results on the LLFF, Mip-NeRF360, DTU, and
MVImgNet datasets demonstrate that our approach enhances NVS quality under
few-shot training conditions, outperforming existing state-of-the-art methods.
The code is released at: https://sailor-z.github.io/projects/SEGS.html.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 18:43:48 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2024 10:39:59 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 03:26:34 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhao",
"Chen",
""
],
[
"Wang",
"Xuan",
""
],
[
"Zhang",
"Tong",
""
],
[
"Javed",
"Saqib",
""
],
[
"Salzmann",
"Mathieu",
""
]
]
| TITLE: Self-Ensembling Gaussian Splatting for Few-Shot Novel View Synthesis
ABSTRACT: 3D Gaussian Splatting (3DGS) has demonstrated remarkable effectiveness in
novel view synthesis (NVS). However, 3DGS tends to overfit when trained with
sparse views, limiting its generalization to novel viewpoints. In this paper,
we address this overfitting issue by introducing Self-Ensembling Gaussian
Splatting (SE-GS). We achieve self-ensembling by incorporating an
uncertainty-aware perturbation strategy during training. A
$\mathbf{\Delta}$-model and a $\mathbf{\Sigma}$-model are jointly trained on
the available images. The $\mathbf{\Delta}$-model is dynamically perturbed
based on rendering uncertainty across training steps, generating diverse
perturbed models with negligible computational overhead. Discrepancies between
the $\mathbf{\Sigma}$-model and these perturbed models are minimized throughout
training, forming a robust ensemble of 3DGS models. This ensemble, represented
by the $\mathbf{\Sigma}$-model, is then used to generate novel-view images
during inference. Experimental results on the LLFF, Mip-NeRF360, DTU, and
MVImgNet datasets demonstrate that our approach enhances NVS quality under
few-shot training conditions, outperforming existing state-of-the-art methods.
The code is released at: https://sailor-z.github.io/projects/SEGS.html.
| no_new_dataset | 0.947137 |
2411.01126 | Davin Hill | Davin Hill, Josh Bone, Aria Masoomi, Max Torop, Jennifer Dy | Axiomatic Explainer Globalness via Optimal Transport | Proceedings of the 28th International Conference on Artificial
Intelligence and Statistics (AISTATS) 2025 | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Explainability methods are often challenging to evaluate and compare. With a
multitude of explainers available, practitioners must often compare and select
explainers based on quantitative evaluation metrics. One particular
differentiator between explainers is the diversity of explanations for a given
dataset; i.e. whether all explanations are identical, unique and uniformly
distributed, or somewhere between these two extremes. In this work, we define a
complexity measure for explainers, globalness, which enables deeper
understanding of the distribution of explanations produced by feature
attribution and feature selection methods for a given dataset. We establish the
axiomatic properties that any such measure should possess and prove that our
proposed measure, Wasserstein Globalness, meets these criteria. We validate the
utility of Wasserstein Globalness using image, tabular, and synthetic datasets,
empirically showing that it both facilitates meaningful comparison between
explainers and improves the selection process for explainability methods.
| [
{
"version": "v1",
"created": "Sat, 2 Nov 2024 04:01:38 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 03:46:50 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Hill",
"Davin",
""
],
[
"Bone",
"Josh",
""
],
[
"Masoomi",
"Aria",
""
],
[
"Torop",
"Max",
""
],
[
"Dy",
"Jennifer",
""
]
]
| TITLE: Axiomatic Explainer Globalness via Optimal Transport
ABSTRACT: Explainability methods are often challenging to evaluate and compare. With a
multitude of explainers available, practitioners must often compare and select
explainers based on quantitative evaluation metrics. One particular
differentiator between explainers is the diversity of explanations for a given
dataset; i.e. whether all explanations are identical, unique and uniformly
distributed, or somewhere between these two extremes. In this work, we define a
complexity measure for explainers, globalness, which enables deeper
understanding of the distribution of explanations produced by feature
attribution and feature selection methods for a given dataset. We establish the
axiomatic properties that any such measure should possess and prove that our
proposed measure, Wasserstein Globalness, meets these criteria. We validate the
utility of Wasserstein Globalness using image, tabular, and synthetic datasets,
empirically showing that it both facilitates meaningful comparison between
explainers and improves the selection process for explainability methods.
| no_new_dataset | 0.951006 |
2411.01293 | Rafa{\l} Karczewski | Rafa{\l} Karczewski, Markus Heinonen, Vikas Garg | Diffusion Models as Cartoonists: The Curious Case of High Density
Regions | ICLR 2025 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | We investigate what kind of images lie in the high-density regions of
diffusion models. We introduce a theoretical mode-tracking process capable of
pinpointing the exact mode of the denoising distribution, and we propose a
practical high-density sampler that consistently generates images of higher
likelihood than usual samplers. Our empirical findings reveal the existence of
significantly higher likelihood samples that typical samplers do not produce,
often manifesting as cartoon-like drawings or blurry images depending on the
noise level. Curiously, these patterns emerge in datasets devoid of such
examples. We also present a novel approach to track sample likelihoods in
diffusion SDEs, which remarkably incurs no additional computational cost.
| [
{
"version": "v1",
"created": "Sat, 2 Nov 2024 16:02:47 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2025 10:41:45 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 12:08:55 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Karczewski",
"Rafał",
""
],
[
"Heinonen",
"Markus",
""
],
[
"Garg",
"Vikas",
""
]
]
| TITLE: Diffusion Models as Cartoonists: The Curious Case of High Density
Regions
ABSTRACT: We investigate what kind of images lie in the high-density regions of
diffusion models. We introduce a theoretical mode-tracking process capable of
pinpointing the exact mode of the denoising distribution, and we propose a
practical high-density sampler that consistently generates images of higher
likelihood than usual samplers. Our empirical findings reveal the existence of
significantly higher likelihood samples that typical samplers do not produce,
often manifesting as cartoon-like drawings or blurry images depending on the
noise level. Curiously, these patterns emerge in datasets devoid of such
examples. We also present a novel approach to track sample likelihoods in
diffusion SDEs, which remarkably incurs no additional computational cost.
| no_new_dataset | 0.950319 |
2411.08127 | SangHyun Park | Shih-Ying Yeh, Sang-Hyun Park, Yi Li, Giyeong Oh, Xuehai Wang, Min
Song, Youngjae Yu | TIPO: Text to Image with Text Presampling for Prompt Optimization | 41 pages, 32 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | TIPO (Text-to-Image Prompt Optimization) introduces an efficient approach for
automatic prompt refinement in text-to-image (T2I) generation. Starting from
simple user prompts, TIPO leverages a lightweight pre-trained model to expand
these prompts into richer, detailed versions. Conceptually, TIPO samples
refined prompts from a targeted sub-distribution within the broader semantic
space, preserving the original intent while significantly improving visual
quality, coherence, and detail. Unlike resource-intensive methods based on
large language models (LLMs) or reinforcement learning (RL), TIPO provides
computational efficiency and scalability, opening new possibilities for
effective, automated prompt engineering in T2I tasks.
We provide visual results, human preference report to investigate TIPO's
effectiveness. Experimental evaluations on benchmark datasets demonstrate
substantial improvements in aesthetic quality, significant reduction of visual
artifacts, and enhanced alignment with target distributions along with
significant human preference proficiency. These results highlight the
importance of targeted prompt engineering in text-to-image tasks and indicate
broader opportunities for automated prompt refinement.
| [
{
"version": "v1",
"created": "Tue, 12 Nov 2024 19:09:45 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2024 14:58:31 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 18:21:57 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Yeh",
"Shih-Ying",
""
],
[
"Park",
"Sang-Hyun",
""
],
[
"Li",
"Yi",
""
],
[
"Oh",
"Giyeong",
""
],
[
"Wang",
"Xuehai",
""
],
[
"Song",
"Min",
""
],
[
"Yu",
"Youngjae",
""
]
]
| TITLE: TIPO: Text to Image with Text Presampling for Prompt Optimization
ABSTRACT: TIPO (Text-to-Image Prompt Optimization) introduces an efficient approach for
automatic prompt refinement in text-to-image (T2I) generation. Starting from
simple user prompts, TIPO leverages a lightweight pre-trained model to expand
these prompts into richer, detailed versions. Conceptually, TIPO samples
refined prompts from a targeted sub-distribution within the broader semantic
space, preserving the original intent while significantly improving visual
quality, coherence, and detail. Unlike resource-intensive methods based on
large language models (LLMs) or reinforcement learning (RL), TIPO provides
computational efficiency and scalability, opening new possibilities for
effective, automated prompt engineering in T2I tasks.
We provide visual results, human preference report to investigate TIPO's
effectiveness. Experimental evaluations on benchmark datasets demonstrate
substantial improvements in aesthetic quality, significant reduction of visual
artifacts, and enhanced alignment with target distributions along with
significant human preference proficiency. These results highlight the
importance of targeted prompt engineering in text-to-image tasks and indicate
broader opportunities for automated prompt refinement.
| no_new_dataset | 0.957038 |
2411.10153 | Gerardo Duran-Martin | Gerardo Duran-Martin, Leandro S\'anchez-Betancourt, Alexander Y.
Shestopaloff, Kevin Murphy | A unifying framework for generalised Bayesian online learning in
non-stationary environments | Published in Transactions on Machine Learning Research (03/2025) | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | We propose a unifying framework for methods that perform probabilistic online
learning in non-stationary environments. We call the framework BONE, which
stands for generalised (B)ayesian (O)nline learning in (N)on-stationary
(E)nvironments. BONE provides a common structure to tackle a variety of
problems, including online continual learning, prequential forecasting, and
contextual bandits. The framework requires specifying three modelling choices:
(i) a model for measurements (e.g., a neural network), (ii) an auxiliary
process to model non-stationarity (e.g., the time since the last changepoint),
and (iii) a conditional prior over model parameters (e.g., a multivariate
Gaussian). The framework also requires two algorithmic choices, which we use to
carry out approximate inference under this framework: (i) an algorithm to
estimate beliefs (posterior distribution) about the model parameters given the
auxiliary variable, and (ii) an algorithm to estimate beliefs about the
auxiliary variable. We show how the modularity of our framework allows for many
existing methods to be reinterpreted as instances of BONE, and it allows us to
propose new methods. We compare experimentally existing methods with our
proposed new method on several datasets, providing insights into the situations
that make each method more suitable for a specific task. We provide a Jax open
source library to facilitate the adoption of this framework.
| [
{
"version": "v1",
"created": "Fri, 15 Nov 2024 12:52:02 GMT"
},
{
"version": "v2",
"created": "Mon, 18 Nov 2024 10:16:14 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 10:05:37 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Duran-Martin",
"Gerardo",
""
],
[
"Sánchez-Betancourt",
"Leandro",
""
],
[
"Shestopaloff",
"Alexander Y.",
""
],
[
"Murphy",
"Kevin",
""
]
]
| TITLE: A unifying framework for generalised Bayesian online learning in
non-stationary environments
ABSTRACT: We propose a unifying framework for methods that perform probabilistic online
learning in non-stationary environments. We call the framework BONE, which
stands for generalised (B)ayesian (O)nline learning in (N)on-stationary
(E)nvironments. BONE provides a common structure to tackle a variety of
problems, including online continual learning, prequential forecasting, and
contextual bandits. The framework requires specifying three modelling choices:
(i) a model for measurements (e.g., a neural network), (ii) an auxiliary
process to model non-stationarity (e.g., the time since the last changepoint),
and (iii) a conditional prior over model parameters (e.g., a multivariate
Gaussian). The framework also requires two algorithmic choices, which we use to
carry out approximate inference under this framework: (i) an algorithm to
estimate beliefs (posterior distribution) about the model parameters given the
auxiliary variable, and (ii) an algorithm to estimate beliefs about the
auxiliary variable. We show how the modularity of our framework allows for many
existing methods to be reinterpreted as instances of BONE, and it allows us to
propose new methods. We compare experimentally existing methods with our
proposed new method on several datasets, providing insights into the situations
that make each method more suitable for a specific task. We provide a Jax open
source library to facilitate the adoption of this framework.
| no_new_dataset | 0.945349 |
2411.10224 | Kang Liu | Qiguang Miao and Kang Liu and Zhuoqi Ma and Yunan Li and Xiaolu Kang
and Ruixuan Liu and Tianyi Liu and Kun Xie and Zhicheng Jiao | EVOKE: Elevating Chest X-ray Report Generation via Multi-View
Contrastive Learning and Patient-Specific Knowledge | The code is available at https://github.com/mk-runner/EVOKE | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Radiology reports are crucial for planning treatment strategies and
facilitating effective doctor-patient communication. However, the manual
creation of these reports places a significant burden on radiologists. While
automatic radiology report generation presents a promising solution, existing
methods often rely on single-view radiographs, which constrain diagnostic
accuracy. To address this challenge, we propose \textbf{EVOKE}, a novel chest
X-ray report generation framework that incorporates multi-view contrastive
learning and patient-specific knowledge. Specifically, we introduce a
multi-view contrastive learning method that enhances visual representation by
aligning multi-view radiographs with their corresponding report. After that, we
present a knowledge-guided report generation module that integrates available
patient-specific indications (e.g., symptom descriptions) to trigger the
production of accurate and coherent radiology reports. To support research in
multi-view report generation, we construct Multi-view CXR and Two-view CXR
datasets using publicly available sources. Our proposed EVOKE surpasses recent
state-of-the-art methods across multiple datasets, achieving a 2.9\%
F\textsubscript{1} RadGraph improvement on MIMIC-CXR, a 7.3\% BLEU-1
improvement on MIMIC-ABN, a 3.1\% BLEU-4 improvement on Multi-view CXR, and an
8.2\% F\textsubscript{1,mic-14} CheXbert improvement on Two-view CXR.
| [
{
"version": "v1",
"created": "Fri, 15 Nov 2024 14:38:13 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 09:38:02 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Miao",
"Qiguang",
""
],
[
"Liu",
"Kang",
""
],
[
"Ma",
"Zhuoqi",
""
],
[
"Li",
"Yunan",
""
],
[
"Kang",
"Xiaolu",
""
],
[
"Liu",
"Ruixuan",
""
],
[
"Liu",
"Tianyi",
""
],
[
"Xie",
"Kun",
""
],
[
"Jiao",
"Zhicheng",
""
]
]
| TITLE: EVOKE: Elevating Chest X-ray Report Generation via Multi-View
Contrastive Learning and Patient-Specific Knowledge
ABSTRACT: Radiology reports are crucial for planning treatment strategies and
facilitating effective doctor-patient communication. However, the manual
creation of these reports places a significant burden on radiologists. While
automatic radiology report generation presents a promising solution, existing
methods often rely on single-view radiographs, which constrain diagnostic
accuracy. To address this challenge, we propose \textbf{EVOKE}, a novel chest
X-ray report generation framework that incorporates multi-view contrastive
learning and patient-specific knowledge. Specifically, we introduce a
multi-view contrastive learning method that enhances visual representation by
aligning multi-view radiographs with their corresponding report. After that, we
present a knowledge-guided report generation module that integrates available
patient-specific indications (e.g., symptom descriptions) to trigger the
production of accurate and coherent radiology reports. To support research in
multi-view report generation, we construct Multi-view CXR and Two-view CXR
datasets using publicly available sources. Our proposed EVOKE surpasses recent
state-of-the-art methods across multiple datasets, achieving a 2.9\%
F\textsubscript{1} RadGraph improvement on MIMIC-CXR, a 7.3\% BLEU-1
improvement on MIMIC-ABN, a 3.1\% BLEU-4 improvement on Multi-view CXR, and an
8.2\% F\textsubscript{1,mic-14} CheXbert improvement on Two-view CXR.
| no_new_dataset | 0.942981 |
2411.15232 | Taha Koleilat | Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz, Yiming Xiao | BiomedCoOp: Learning to Prompt for Biomedical Vision-Language Models | Accepted to CVPR 2025 | null | null | null | cs.CV cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent advancements in vision-language models (VLMs), such as CLIP, have
demonstrated substantial success in self-supervised representation learning for
vision tasks. However, effectively adapting VLMs to downstream applications
remains challenging, as their accuracy often depends on time-intensive and
expertise-demanding prompt engineering, while full model fine-tuning is costly.
This is particularly true for biomedical images, which, unlike natural images,
typically suffer from limited annotated datasets, unintuitive image contrasts,
and nuanced visual features. Recent prompt learning techniques, such as Context
Optimization (CoOp) intend to tackle these issues, but still fall short in
generalizability. Meanwhile, explorations in prompt learning for biomedical
image analysis are still highly limited. In this work, we propose BiomedCoOp, a
novel prompt learning framework that enables efficient adaptation of BiomedCLIP
for accurate and highly generalizable few-shot biomedical image classification.
Our approach achieves effective prompt context learning by leveraging semantic
consistency with average prompt ensembles from Large Language Models (LLMs) and
knowledge distillation with a statistics-based prompt selection strategy. We
conducted comprehensive validation of our proposed framework on 11 medical
datasets across 9 modalities and 10 organs against existing state-of-the-art
methods, demonstrating significant improvements in both accuracy and
generalizability. The code is publicly available at
https://github.com/HealthX-Lab/BiomedCoOp.
| [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 19:13:04 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 03:28:09 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Koleilat",
"Taha",
""
],
[
"Asgariandehkordi",
"Hojat",
""
],
[
"Rivaz",
"Hassan",
""
],
[
"Xiao",
"Yiming",
""
]
]
| TITLE: BiomedCoOp: Learning to Prompt for Biomedical Vision-Language Models
ABSTRACT: Recent advancements in vision-language models (VLMs), such as CLIP, have
demonstrated substantial success in self-supervised representation learning for
vision tasks. However, effectively adapting VLMs to downstream applications
remains challenging, as their accuracy often depends on time-intensive and
expertise-demanding prompt engineering, while full model fine-tuning is costly.
This is particularly true for biomedical images, which, unlike natural images,
typically suffer from limited annotated datasets, unintuitive image contrasts,
and nuanced visual features. Recent prompt learning techniques, such as Context
Optimization (CoOp) intend to tackle these issues, but still fall short in
generalizability. Meanwhile, explorations in prompt learning for biomedical
image analysis are still highly limited. In this work, we propose BiomedCoOp, a
novel prompt learning framework that enables efficient adaptation of BiomedCLIP
for accurate and highly generalizable few-shot biomedical image classification.
Our approach achieves effective prompt context learning by leveraging semantic
consistency with average prompt ensembles from Large Language Models (LLMs) and
knowledge distillation with a statistics-based prompt selection strategy. We
conducted comprehensive validation of our proposed framework on 11 medical
datasets across 9 modalities and 10 organs against existing state-of-the-art
methods, demonstrating significant improvements in both accuracy and
generalizability. The code is publicly available at
https://github.com/HealthX-Lab/BiomedCoOp.
| no_new_dataset | 0.944536 |
2411.16370 | Amaan Valiuddin | M.M.A. Valiuddin, R.J.G. van Sloun, C.G.A. Viviers, P.H.N. de With, F.
van der Sommen | A Review of Bayesian Uncertainty Quantification in Deep Probabilistic
Image Segmentation | 20 pages, revised | null | null | null | cs.CV cs.AI cs.LG eess.IV stat.ML | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Advancements in image segmentation play an integral role within the broad
scope of Deep Learning-based Computer Vision. Furthermore, their widespread
applicability in critical real-world tasks has resulted in challenges related
to the reliability of such algorithms. Hence, uncertainty quantification has
been extensively studied within this context, enabling the expression of model
ignorance (epistemic uncertainty) or data ambiguity (aleatoric uncertainty) to
prevent uninformed decision-making. Due to the rapid adoption of Convolutional
Neural Network (CNN)-based segmentation models in high-stake applications, a
substantial body of research has been published on this very topic, causing its
swift expansion into a distinct field. This work provides a comprehensive
overview of probabilistic segmentation, by discussing fundamental concepts of
uncertainty quantification, governing advancements in the field as well as the
application to various tasks. Moreover, literature on both types of
uncertainties trace back to four key applications: (1) to quantify statistical
inconsistencies in the annotation process due ambiguous images, (2) correlating
prediction error with uncertainty, (3) expanding the model hypothesis space for
better generalization, and (4) Active Learning. An extensive discussion follows
that includes an overview of utilized datasets for each of the applications and
evaluation of the available methods. We also highlight challenges related to
architectures, uncertainty quantification methods, standardization and
benchmarking, and finally end with recommendations for future work such as
methods based on single forward passes and models that appropriately leverage
volumetric data.
| [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 13:26:09 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Jan 2025 09:34:51 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 09:51:17 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Valiuddin",
"M. M. A.",
""
],
[
"van Sloun",
"R. J. G.",
""
],
[
"Viviers",
"C. G. A.",
""
],
[
"de With",
"P. H. N.",
""
],
[
"van der Sommen",
"F.",
""
]
]
| TITLE: A Review of Bayesian Uncertainty Quantification in Deep Probabilistic
Image Segmentation
ABSTRACT: Advancements in image segmentation play an integral role within the broad
scope of Deep Learning-based Computer Vision. Furthermore, their widespread
applicability in critical real-world tasks has resulted in challenges related
to the reliability of such algorithms. Hence, uncertainty quantification has
been extensively studied within this context, enabling the expression of model
ignorance (epistemic uncertainty) or data ambiguity (aleatoric uncertainty) to
prevent uninformed decision-making. Due to the rapid adoption of Convolutional
Neural Network (CNN)-based segmentation models in high-stake applications, a
substantial body of research has been published on this very topic, causing its
swift expansion into a distinct field. This work provides a comprehensive
overview of probabilistic segmentation, by discussing fundamental concepts of
uncertainty quantification, governing advancements in the field as well as the
application to various tasks. Moreover, literature on both types of
uncertainties trace back to four key applications: (1) to quantify statistical
inconsistencies in the annotation process due ambiguous images, (2) correlating
prediction error with uncertainty, (3) expanding the model hypothesis space for
better generalization, and (4) Active Learning. An extensive discussion follows
that includes an overview of utilized datasets for each of the applications and
evaluation of the available methods. We also highlight challenges related to
architectures, uncertainty quantification methods, standardization and
benchmarking, and finally end with recommendations for future work such as
methods based on single forward passes and models that appropriately leverage
volumetric data.
| no_new_dataset | 0.945248 |
2411.16901 | Abdesselam Ferdi | Abdesselam Ferdi | Deep Convolutional Neural Networks Structured Pruning via Gravity
Regularization | null | null | 10.1109/ICMLANT63295.2024.00009 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Structured pruning is a widely employed strategy for accelerating deep
convolutional neural networks (DCNNs). However, existing methods often
necessitate modifications to the original architectures, involve complex
implementations, and require lengthy fine-tuning stages. To address these
challenges, we propose a novel physics-inspired approach that integrates the
concept of gravity into the training stage of DCNNs. In this approach, the
gravity is directly proportional to the product of the masses of the
convolution filter and the attracting filter, and inversely proportional to the
square of the distance between them. We applied this force to the convolution
filters, either drawing filters closer to the attracting filter (experiencing
weaker gravity) toward non-zero weights or pulling filters farther away
(subject to stronger gravity) toward zero weights. As a result, filters
experiencing stronger gravity have their weights reduced to zero, enabling
their removal, while filters under weaker gravity retain significant weights
and preserve important information. Our method simultaneously optimizes the
filter weights and ranks their importance, eliminating the need for complex
implementations or extensive fine-tuning. We validated the proposed approach on
popular DCNN architectures using the CIFAR dataset, achieving competitive
results compared to existing methods.
| [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 20:10:10 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Ferdi",
"Abdesselam",
""
]
]
| TITLE: Deep Convolutional Neural Networks Structured Pruning via Gravity
Regularization
ABSTRACT: Structured pruning is a widely employed strategy for accelerating deep
convolutional neural networks (DCNNs). However, existing methods often
necessitate modifications to the original architectures, involve complex
implementations, and require lengthy fine-tuning stages. To address these
challenges, we propose a novel physics-inspired approach that integrates the
concept of gravity into the training stage of DCNNs. In this approach, the
gravity is directly proportional to the product of the masses of the
convolution filter and the attracting filter, and inversely proportional to the
square of the distance between them. We applied this force to the convolution
filters, either drawing filters closer to the attracting filter (experiencing
weaker gravity) toward non-zero weights or pulling filters farther away
(subject to stronger gravity) toward zero weights. As a result, filters
experiencing stronger gravity have their weights reduced to zero, enabling
their removal, while filters under weaker gravity retain significant weights
and preserve important information. Our method simultaneously optimizes the
filter weights and ranks their importance, eliminating the need for complex
implementations or extensive fine-tuning. We validated the proposed approach on
popular DCNN architectures using the CIFAR dataset, achieving competitive
results compared to existing methods.
| no_new_dataset | 0.952309 |
2411.17489 | Nicolai Hermann | Nicolai Hermann, Jorge Condor, Piotr Didyk | Puzzle Similarity: A Perceptually-guided Cross-Reference Metric for
Artifact Detection in 3D Scene Reconstructions | null | null | null | null | cs.CV cs.AI cs.GR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern reconstruction techniques can effectively model complex 3D scenes from
sparse 2D views. However, automatically assessing the quality of novel views
and identifying artifacts is challenging due to the lack of ground truth images
and the limitations of No-Reference image metrics in predicting reliable
artifact maps. The absence of such metrics hinders the assessment of the
quality of novel views and limits the adoption of post-processing techniques,
such as inpainting, to enhance reconstruction quality. To tackle this, recent
work has established a new category of metrics (Cross-Reference), predicting
image quality solely by leveraging context from alternate viewpoint captures
(arXiv:2404.14409). In this work, we propose a new Cross-Reference metric,
Puzzle Similarity, which is designed to localize artifacts in novel views. Our
approach utilizes image patch statistics from the input views to establish a
scene-specific distribution, later used to identify poorly reconstructed
regions in the novel views. Given the lack of good measures to evaluate
Cross-Reference methods in the context of 3D reconstruction, we collected a
novel human-labeled dataset of artifact and distortion maps in unseen
reconstructed views. Through this dataset, we demonstrate that our method
achieves state-of-the-art localization of artifacts in novel views, correlating
with human assessment, even without aligned references. We can leverage our new
metric to enhance applications like automatic image restoration, guided
acquisition, or 3D reconstruction from sparse inputs. Find the project page at
https://nihermann.github.io/puzzlesim/ .
| [
{
"version": "v1",
"created": "Tue, 26 Nov 2024 14:57:30 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 09:04:43 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Hermann",
"Nicolai",
""
],
[
"Condor",
"Jorge",
""
],
[
"Didyk",
"Piotr",
""
]
]
| TITLE: Puzzle Similarity: A Perceptually-guided Cross-Reference Metric for
Artifact Detection in 3D Scene Reconstructions
ABSTRACT: Modern reconstruction techniques can effectively model complex 3D scenes from
sparse 2D views. However, automatically assessing the quality of novel views
and identifying artifacts is challenging due to the lack of ground truth images
and the limitations of No-Reference image metrics in predicting reliable
artifact maps. The absence of such metrics hinders the assessment of the
quality of novel views and limits the adoption of post-processing techniques,
such as inpainting, to enhance reconstruction quality. To tackle this, recent
work has established a new category of metrics (Cross-Reference), predicting
image quality solely by leveraging context from alternate viewpoint captures
(arXiv:2404.14409). In this work, we propose a new Cross-Reference metric,
Puzzle Similarity, which is designed to localize artifacts in novel views. Our
approach utilizes image patch statistics from the input views to establish a
scene-specific distribution, later used to identify poorly reconstructed
regions in the novel views. Given the lack of good measures to evaluate
Cross-Reference methods in the context of 3D reconstruction, we collected a
novel human-labeled dataset of artifact and distortion maps in unseen
reconstructed views. Through this dataset, we demonstrate that our method
achieves state-of-the-art localization of artifacts in novel views, correlating
with human assessment, even without aligned references. We can leverage our new
metric to enhance applications like automatic image restoration, guided
acquisition, or 3D reconstruction from sparse inputs. Find the project page at
https://nihermann.github.io/puzzlesim/ .
| new_dataset | 0.963403 |
2412.00139 | Muhammad Huzaifa | Muhammad Huzaifa, Yova Kementchedjhieva | EFSA: Episodic Few-Shot Adaptation for Text-to-Image Retrieval | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Text-to-image retrieval is a critical task for managing diverse visual
content, but common benchmarks for the task rely on small, single-domain
datasets that fail to capture real-world complexity. Pre-trained
vision-language models tend to perform well with easy negatives but struggle
with hard negatives--visually similar yet incorrect images--especially in
open-domain scenarios. To address this, we introduce Episodic Few-Shot
Adaptation (EFSA), a novel test-time framework that adapts pre-trained models
dynamically to a query's domain by fine-tuning on top-k retrieved candidates
and synthetic captions generated for them. EFSA improves performance across
diverse domains while preserving generalization, as shown in evaluations on
queries from eight highly distinct visual domains and an open-domain retrieval
pool of over one million images. Our work highlights the potential of episodic
few-shot adaptation to enhance robustness in the critical and understudied task
of open-domain text-to-image retrieval.
| [
{
"version": "v1",
"created": "Thu, 28 Nov 2024 17:09:20 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 09:54:42 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Huzaifa",
"Muhammad",
""
],
[
"Kementchedjhieva",
"Yova",
""
]
]
| TITLE: EFSA: Episodic Few-Shot Adaptation for Text-to-Image Retrieval
ABSTRACT: Text-to-image retrieval is a critical task for managing diverse visual
content, but common benchmarks for the task rely on small, single-domain
datasets that fail to capture real-world complexity. Pre-trained
vision-language models tend to perform well with easy negatives but struggle
with hard negatives--visually similar yet incorrect images--especially in
open-domain scenarios. To address this, we introduce Episodic Few-Shot
Adaptation (EFSA), a novel test-time framework that adapts pre-trained models
dynamically to a query's domain by fine-tuning on top-k retrieved candidates
and synthetic captions generated for them. EFSA improves performance across
diverse domains while preserving generalization, as shown in evaluations on
queries from eight highly distinct visual domains and an open-domain retrieval
pool of over one million images. Our work highlights the potential of episodic
few-shot adaptation to enhance robustness in the critical and understudied task
of open-domain text-to-image retrieval.
| no_new_dataset | 0.9463 |
2412.00418 | Yu Shi | Yu Shi, Yiqi Wang, WeiXuan Lang, Jiaxin Zhang, Pan Dong, Aiping Li | Mixture of Experts for Node Classification | null | null | null | null | cs.SI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nodes in the real-world graphs exhibit diverse patterns in numerous aspects,
such as degree and homophily. However, most existent node predictors fail to
capture a wide range of node patterns or to make predictions based on distinct
node patterns, resulting in unsatisfactory classification performance. In this
paper, we reveal that different node predictors are good at handling nodes with
specific patterns and only apply one node predictor uniformly could lead to
suboptimal result. To mitigate this gap, we propose a mixture of experts
framework, MoE-NP, for node classification. Specifically, MoE-NP combines a
mixture of node predictors and strategically selects models based on node
patterns. Experimental results from a range of real-world datasets demonstrate
significant performance improvements from MoE-NP.
| [
{
"version": "v1",
"created": "Sat, 30 Nov 2024 10:05:03 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 12:33:46 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Shi",
"Yu",
""
],
[
"Wang",
"Yiqi",
""
],
[
"Lang",
"WeiXuan",
""
],
[
"Zhang",
"Jiaxin",
""
],
[
"Dong",
"Pan",
""
],
[
"Li",
"Aiping",
""
]
]
| TITLE: Mixture of Experts for Node Classification
ABSTRACT: Nodes in the real-world graphs exhibit diverse patterns in numerous aspects,
such as degree and homophily. However, most existent node predictors fail to
capture a wide range of node patterns or to make predictions based on distinct
node patterns, resulting in unsatisfactory classification performance. In this
paper, we reveal that different node predictors are good at handling nodes with
specific patterns and only apply one node predictor uniformly could lead to
suboptimal result. To mitigate this gap, we propose a mixture of experts
framework, MoE-NP, for node classification. Specifically, MoE-NP combines a
mixture of node predictors and strategically selects models based on node
patterns. Experimental results from a range of real-world datasets demonstrate
significant performance improvements from MoE-NP.
| no_new_dataset | 0.95803 |
2412.01562 | Miroslav Purkrabek | Miroslav Purkrabek and Jiri Matas | Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing
the Virtuous Circle | Code: https://mirapurkrabek.github.io/BBox-Mask-Pose | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Human pose estimation methods work well on isolated people but struggle with
multiple-bodies-in-proximity scenarios. Previous work has addressed this
problem by conditioning pose estimation by detected bounding boxes or
keypoints, but overlooked instance masks. We propose to iteratively enforce
mutual consistency of bounding boxes, instance masks, and poses. The introduced
BBox-Mask-Pose (BMP) method uses three specialized models that improve each
other's output in a closed loop. All models are adapted for mutual
conditioning, which improves robustness in multi-body scenes. MaskPose, a new
mask-conditioned pose estimation model, is the best among top-down approaches
on OCHuman. BBox-Mask-Pose pushes SOTA on OCHuman dataset in all three tasks -
detection, instance segmentation, and pose estimation. It also achieves SOTA
performance on COCO pose estimation. The method is especially good in scenes
with large instances overlap, where it improves detection by 39% over the
baseline detector. With small specialized models and faster runtime, BMP is an
effective alternative to large human-centered foundational models. Code and
models are available on https://MiraPurkrabek.github.io/BBox-Mask-Pose.
| [
{
"version": "v1",
"created": "Mon, 2 Dec 2024 14:50:15 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 14:28:25 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Purkrabek",
"Miroslav",
""
],
[
"Matas",
"Jiri",
""
]
]
| TITLE: Detection, Pose Estimation and Segmentation for Multiple Bodies: Closing
the Virtuous Circle
ABSTRACT: Human pose estimation methods work well on isolated people but struggle with
multiple-bodies-in-proximity scenarios. Previous work has addressed this
problem by conditioning pose estimation by detected bounding boxes or
keypoints, but overlooked instance masks. We propose to iteratively enforce
mutual consistency of bounding boxes, instance masks, and poses. The introduced
BBox-Mask-Pose (BMP) method uses three specialized models that improve each
other's output in a closed loop. All models are adapted for mutual
conditioning, which improves robustness in multi-body scenes. MaskPose, a new
mask-conditioned pose estimation model, is the best among top-down approaches
on OCHuman. BBox-Mask-Pose pushes SOTA on OCHuman dataset in all three tasks -
detection, instance segmentation, and pose estimation. It also achieves SOTA
performance on COCO pose estimation. The method is especially good in scenes
with large instances overlap, where it improves detection by 39% over the
baseline detector. With small specialized models and faster runtime, BMP is an
effective alternative to large human-centered foundational models. Code and
models are available on https://MiraPurkrabek.github.io/BBox-Mask-Pose.
| no_new_dataset | 0.952706 |
2412.02386 | Blanca Lasheras-Hernandez | Blanca Lasheras-Hernandez, Klaus H. Strobl, Sergio Izquierdo, Tim
Bodenm\"uller, Rudolph Triebel, and Javier Civera | Single-Shot Metric Depth from Focused Plenoptic Cameras | 8 pages (6 for text + 2 for references), 6 figures, 2 tables.
Accepted at IEEE ICRA 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Metric depth estimation from visual sensors is crucial for robots to
perceive, navigate, and interact with their environment. Traditional range
imaging setups, such as stereo or structured light cameras, face hassles
including calibration, occlusions, and hardware demands, with accuracy limited
by the baseline between cameras. Single- and multi-view monocular depth offers
a more compact alternative, but is constrained by the unobservability of the
metric scale. Light field imaging provides a promising solution for estimating
metric depth by using a unique lens configuration through a single device.
However, its application to single-view dense metric depth is under-addressed
mainly due to the technology's high cost, the lack of public benchmarks, and
proprietary geometrical models and software. Our work explores the potential of
focused plenoptic cameras for dense metric depth. We propose a novel pipeline
that predicts metric depth from a single plenoptic camera shot by first
generating a sparse metric point cloud using machine learning, which is then
used to scale and align a dense relative depth map regressed by a foundation
depth model, resulting in dense metric depth. To validate it, we curated the
Light Field & Stereo Image Dataset (LFS) of real-world light field images with
stereo depth labels, filling a current gap in existing resources. Experimental
results show that our pipeline produces accurate metric depth predictions,
laying a solid groundwork for future research in this field.
| [
{
"version": "v1",
"created": "Tue, 3 Dec 2024 11:21:17 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 13:31:15 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Lasheras-Hernandez",
"Blanca",
""
],
[
"Strobl",
"Klaus H.",
""
],
[
"Izquierdo",
"Sergio",
""
],
[
"Bodenmüller",
"Tim",
""
],
[
"Triebel",
"Rudolph",
""
],
[
"Civera",
"Javier",
""
]
]
| TITLE: Single-Shot Metric Depth from Focused Plenoptic Cameras
ABSTRACT: Metric depth estimation from visual sensors is crucial for robots to
perceive, navigate, and interact with their environment. Traditional range
imaging setups, such as stereo or structured light cameras, face hassles
including calibration, occlusions, and hardware demands, with accuracy limited
by the baseline between cameras. Single- and multi-view monocular depth offers
a more compact alternative, but is constrained by the unobservability of the
metric scale. Light field imaging provides a promising solution for estimating
metric depth by using a unique lens configuration through a single device.
However, its application to single-view dense metric depth is under-addressed
mainly due to the technology's high cost, the lack of public benchmarks, and
proprietary geometrical models and software. Our work explores the potential of
focused plenoptic cameras for dense metric depth. We propose a novel pipeline
that predicts metric depth from a single plenoptic camera shot by first
generating a sparse metric point cloud using machine learning, which is then
used to scale and align a dense relative depth map regressed by a foundation
depth model, resulting in dense metric depth. To validate it, we curated the
Light Field & Stereo Image Dataset (LFS) of real-world light field images with
stereo depth labels, filling a current gap in existing resources. Experimental
results show that our pipeline produces accurate metric depth predictions,
laying a solid groundwork for future research in this field.
| no_new_dataset | 0.888662 |
2412.04106 | Haoning Wu | Haoning Wu, Ziheng Zhao, Ya Zhang, Yanfeng Wang, Weidi Xie | MRGen: Segmentation Data Engine For Underrepresented MRI Modalities | Technical Report; Project Page:
https://haoningwu3639.github.io/MRGen/ | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training medical image segmentation models for rare yet clinically
significant imaging modalities is challenging due to the scarcity of annotated
data, and manual mask annotations can be costly and labor-intensive to acquire.
This paper investigates leveraging generative models to synthesize training
data, to train segmentation models for underrepresented modalities,
particularly on annotation-scarce MRI. Concretely, our contributions are
threefold: (i) we introduce MRGen-DB, a large-scale radiology image-text
dataset comprising extensive samples with rich metadata, including modality
labels, attributes, regions, and organs information, with a subset having
pixelwise mask annotations; (ii) we present MRGen, a diffusion-based data
engine for controllable medical image synthesis, conditioned on text prompts
and segmentation masks. MRGen can generate realistic images for diverse MRI
modalities lacking mask annotations, facilitating segmentation training in
low-source domains; (iii) extensive experiments across multiple modalities
demonstrate that MRGen significantly improves segmentation performance on
unannotated modalities by providing high-quality synthetic data. We believe
that our method bridges a critical gap in medical image analysis, extending
segmentation capabilities to scenarios that are challenging to acquire manual
annotations.
| [
{
"version": "v1",
"created": "Wed, 4 Dec 2024 16:34:22 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 11:59:46 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Wu",
"Haoning",
""
],
[
"Zhao",
"Ziheng",
""
],
[
"Zhang",
"Ya",
""
],
[
"Wang",
"Yanfeng",
""
],
[
"Xie",
"Weidi",
""
]
]
| TITLE: MRGen: Segmentation Data Engine For Underrepresented MRI Modalities
ABSTRACT: Training medical image segmentation models for rare yet clinically
significant imaging modalities is challenging due to the scarcity of annotated
data, and manual mask annotations can be costly and labor-intensive to acquire.
This paper investigates leveraging generative models to synthesize training
data, to train segmentation models for underrepresented modalities,
particularly on annotation-scarce MRI. Concretely, our contributions are
threefold: (i) we introduce MRGen-DB, a large-scale radiology image-text
dataset comprising extensive samples with rich metadata, including modality
labels, attributes, regions, and organs information, with a subset having
pixelwise mask annotations; (ii) we present MRGen, a diffusion-based data
engine for controllable medical image synthesis, conditioned on text prompts
and segmentation masks. MRGen can generate realistic images for diverse MRI
modalities lacking mask annotations, facilitating segmentation training in
low-source domains; (iii) extensive experiments across multiple modalities
demonstrate that MRGen significantly improves segmentation performance on
unannotated modalities by providing high-quality synthetic data. We believe
that our method bridges a critical gap in medical image analysis, extending
segmentation capabilities to scenarios that are challenging to acquire manual
annotations.
| new_dataset | 0.964656 |
2412.06485 | Aylar Partovizadeh | Aylar Partovizadeh, Sebastian Sch\"ops, Dimitrios Loukrezis | Fourier-enhanced reduced-order surrogate modeling for uncertainty
quantification in electric machine design | null | null | 10.1007/s00366-025-02123-1 | null | cs.CE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This work proposes a data-driven surrogate modeling framework for
cost-effectively inferring the torque of a permanent magnet synchronous machine
under geometric design variations. The framework is separated into a
reduced-order modeling and an inference part. Given a dataset of torque
signals, each corresponding to a different set of design parameters, torque
dimension is first reduced by post-processing a discrete Fourier transform and
keeping a reduced number of frequency components. This allows to take advantage
of torque periodicity and preserve physical information contained in the
frequency components. Next, a response surface model is computed by means of
machine learning regression, which maps the design parameters to the reduced
frequency components. The response surface models of choice are polynomial
chaos expansions, feedforward neural networks, and Gaussian processes. Torque
inference is performed by evaluating the response surface model for new design
parameters and then inverting the dimension reduction. Numerical results show
that the resulting surrogate models lead to sufficiently accurate torque
predictions for previously unseen design configurations. The framework is found
to be significantly advantageous compared to approximating the original (not
reduced) torque signal directly, as well as slightly advantageous compared to
using principal component analysis for dimension reduction. The combination of
discrete Fourier transform-based dimension reduction with Gaussian
process-based response surfaces yields the best-in-class surrogate model for
this use case. The surrogate models replace the original, high-fidelity model
in Monte Carlo-based uncertainty quantification studies, where they provide
accurate torque statistics estimates at significantly reduced computational
cost.
| [
{
"version": "v1",
"created": "Mon, 9 Dec 2024 13:35:28 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:57:50 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Partovizadeh",
"Aylar",
""
],
[
"Schöps",
"Sebastian",
""
],
[
"Loukrezis",
"Dimitrios",
""
]
]
| TITLE: Fourier-enhanced reduced-order surrogate modeling for uncertainty
quantification in electric machine design
ABSTRACT: This work proposes a data-driven surrogate modeling framework for
cost-effectively inferring the torque of a permanent magnet synchronous machine
under geometric design variations. The framework is separated into a
reduced-order modeling and an inference part. Given a dataset of torque
signals, each corresponding to a different set of design parameters, torque
dimension is first reduced by post-processing a discrete Fourier transform and
keeping a reduced number of frequency components. This allows to take advantage
of torque periodicity and preserve physical information contained in the
frequency components. Next, a response surface model is computed by means of
machine learning regression, which maps the design parameters to the reduced
frequency components. The response surface models of choice are polynomial
chaos expansions, feedforward neural networks, and Gaussian processes. Torque
inference is performed by evaluating the response surface model for new design
parameters and then inverting the dimension reduction. Numerical results show
that the resulting surrogate models lead to sufficiently accurate torque
predictions for previously unseen design configurations. The framework is found
to be significantly advantageous compared to approximating the original (not
reduced) torque signal directly, as well as slightly advantageous compared to
using principal component analysis for dimension reduction. The combination of
discrete Fourier transform-based dimension reduction with Gaussian
process-based response surfaces yields the best-in-class surrogate model for
this use case. The surrogate models replace the original, high-fidelity model
in Monte Carlo-based uncertainty quantification studies, where they provide
accurate torque statistics estimates at significantly reduced computational
cost.
| no_new_dataset | 0.949342 |
2412.07923 | Sagi Shaier | Sagi Shaier, Mario Sanz-Guerrero, Katharina von der Wense | Asking Again and Again: Exploring LLM Robustness to Repeated Questions | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This study investigates whether repeating questions within prompts influences
the performance of large language models (LLMs). We hypothesize that
reiterating a question within a single prompt might enhance the model's focus
on key elements of the query. We evaluate five recent LLMs -- including
GPT-4o-mini, DeepSeek-V3, and smaller open-source models -- on three reading
comprehension datasets under different prompt settings, varying question
repetition levels (1, 3, or 5 times per prompt). Our results demonstrate that
question repetition can increase models' accuracy by up to $6\%$. However,
across all models, settings, and datasets, we do not find the result
statistically significant. These findings provide insights into prompt design
and LLM behavior, suggesting that repetition alone does not significantly
impact output quality.
| [
{
"version": "v1",
"created": "Tue, 10 Dec 2024 21:09:12 GMT"
},
{
"version": "v2",
"created": "Sat, 8 Mar 2025 16:42:51 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 13:48:12 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Shaier",
"Sagi",
""
],
[
"Sanz-Guerrero",
"Mario",
""
],
[
"von der Wense",
"Katharina",
""
]
]
| TITLE: Asking Again and Again: Exploring LLM Robustness to Repeated Questions
ABSTRACT: This study investigates whether repeating questions within prompts influences
the performance of large language models (LLMs). We hypothesize that
reiterating a question within a single prompt might enhance the model's focus
on key elements of the query. We evaluate five recent LLMs -- including
GPT-4o-mini, DeepSeek-V3, and smaller open-source models -- on three reading
comprehension datasets under different prompt settings, varying question
repetition levels (1, 3, or 5 times per prompt). Our results demonstrate that
question repetition can increase models' accuracy by up to $6\%$. However,
across all models, settings, and datasets, we do not find the result
statistically significant. These findings provide insights into prompt design
and LLM behavior, suggesting that repetition alone does not significantly
impact output quality.
| no_new_dataset | 0.943556 |
2412.10211 | Paula Daud\'en-Oliver | Paula Daud\'en-Oliver and David Agost-Beltran and Emilio
Sansano-Sansano and Valero Laparra and Jes\'us Malo and Marina
Mart\'inez-Garcia | RAID-Database: human Responses to Affine Image Distortions | null | null | null | null | cs.CV q-bio.NC q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Image quality databases are used to train models for predicting subjective
human perception. However, most existing databases focus on distortions
commonly found in digital media and not in natural conditions. Affine
transformations are particularly relevant to study, as they are among the most
commonly encountered by human observers in everyday life. This Data Descriptor
presents a set of human responses to suprathreshold affine image transforms
(rotation, translation, scaling) and Gaussian noise as convenient reference to
compare with previously existing image quality databases. The responses were
measured using well established psychophysics: the Maximum Likelihood
Difference Scaling method. The set contains responses to 864 distorted images.
The experiments involved 105 observers and more than 20000 comparisons of
quadruples of images. The quality of the dataset is ensured because (a) it
reproduces the classical Pi\'eron's law, (b) it reproduces classical absolute
detection thresholds, and (c) it is consistent with conventional image quality
databases but improves them according to Group-MAD experiments.
| [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 15:34:34 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 15:12:43 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Daudén-Oliver",
"Paula",
""
],
[
"Agost-Beltran",
"David",
""
],
[
"Sansano-Sansano",
"Emilio",
""
],
[
"Laparra",
"Valero",
""
],
[
"Malo",
"Jesús",
""
],
[
"Martínez-Garcia",
"Marina",
""
]
]
| TITLE: RAID-Database: human Responses to Affine Image Distortions
ABSTRACT: Image quality databases are used to train models for predicting subjective
human perception. However, most existing databases focus on distortions
commonly found in digital media and not in natural conditions. Affine
transformations are particularly relevant to study, as they are among the most
commonly encountered by human observers in everyday life. This Data Descriptor
presents a set of human responses to suprathreshold affine image transforms
(rotation, translation, scaling) and Gaussian noise as convenient reference to
compare with previously existing image quality databases. The responses were
measured using well established psychophysics: the Maximum Likelihood
Difference Scaling method. The set contains responses to 864 distorted images.
The experiments involved 105 observers and more than 20000 comparisons of
quadruples of images. The quality of the dataset is ensured because (a) it
reproduces the classical Pi\'eron's law, (b) it reproduces classical absolute
detection thresholds, and (c) it is consistent with conventional image quality
databases but improves them according to Group-MAD experiments.
| no_new_dataset | 0.876264 |
2412.10488 | Zehao Chen | Zehao Chen, Rong Pan | SVGBuilder: Component-Based Colored SVG Generation with Text-Guided
Autoregressive Transformers | Project: https://svgbuilder.github.io | null | null | null | cs.CV cs.AI cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Scalable Vector Graphics (SVG) are essential XML-based formats for versatile
graphics, offering resolution independence and scalability. Unlike raster
images, SVGs use geometric shapes and support interactivity, animation, and
manipulation via CSS and JavaScript. Current SVG generation methods face
challenges related to high computational costs and complexity. In contrast,
human designers use component-based tools for efficient SVG creation. Inspired
by this, SVGBuilder introduces a component-based, autoregressive model for
generating high-quality colored SVGs from textual input. It significantly
reduces computational overhead and improves efficiency compared to traditional
methods. Our model generates SVGs up to 604 times faster than
optimization-based approaches. To address the limitations of existing SVG
datasets and support our research, we introduce ColorSVG-100K, the first
large-scale dataset of colored SVGs, comprising 100,000 graphics. This dataset
fills the gap in color information for SVG generation models and enhances
diversity in model training. Evaluation against state-of-the-art models
demonstrates SVGBuilder's superior performance in practical applications,
highlighting its efficiency and quality in generating complex SVG graphics.
| [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 15:24:11 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Dec 2024 16:13:15 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 14:34:11 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chen",
"Zehao",
""
],
[
"Pan",
"Rong",
""
]
]
| TITLE: SVGBuilder: Component-Based Colored SVG Generation with Text-Guided
Autoregressive Transformers
ABSTRACT: Scalable Vector Graphics (SVG) are essential XML-based formats for versatile
graphics, offering resolution independence and scalability. Unlike raster
images, SVGs use geometric shapes and support interactivity, animation, and
manipulation via CSS and JavaScript. Current SVG generation methods face
challenges related to high computational costs and complexity. In contrast,
human designers use component-based tools for efficient SVG creation. Inspired
by this, SVGBuilder introduces a component-based, autoregressive model for
generating high-quality colored SVGs from textual input. It significantly
reduces computational overhead and improves efficiency compared to traditional
methods. Our model generates SVGs up to 604 times faster than
optimization-based approaches. To address the limitations of existing SVG
datasets and support our research, we introduce ColorSVG-100K, the first
large-scale dataset of colored SVGs, comprising 100,000 graphics. This dataset
fills the gap in color information for SVG generation models and enhances
diversity in model training. Evaluation against state-of-the-art models
demonstrates SVGBuilder's superior performance in practical applications,
highlighting its efficiency and quality in generating complex SVG graphics.
| new_dataset | 0.956877 |
2412.11464 | Quan-Sheng Zeng | Quan-Sheng Zeng, Yunheng Li, Daquan Zhou, Guanbin Li, Qibin Hou,
Ming-Ming Cheng | High-Quality Mask Tuning Matters for Open-Vocabulary Segmentation | Revised version according to comments from reviewers of ICLR2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Open-vocabulary image segmentation has been advanced through the synergy
between mask generators and vision-language models like Contrastive
Language-Image Pre-training (CLIP). Previous approaches focus on generating
masks while aligning mask features with text embeddings during training. In
this paper, we observe that relying on generated low-quality masks can weaken
the alignment of vision and language in regional representations. This
motivates us to present a new fine-tuning framework, named MaskCLIP++, which
uses ground-truth masks instead of generated masks to enhance the mask
classification capability of CLIP. Due to the limited diversity of image
segmentation datasets with mask annotations, we propose incorporating a
consistency alignment principle during fine-tuning, which alleviates
categorical bias toward the fine-tuning dataset. After low-cost fine-tuning,
MaskCLIP++ significantly improves the mask classification performance on
multi-domain datasets. Combining with the mask generator in previous
state-of-the-art mask-based open vocabulary segmentation methods, we achieve
performance improvements of +1.7, +2.3, +2.1, +3.1, and +0.3 mIoU on the A-847,
PC-459, A-150, PC-59, and PAS-20 datasets, respectively. Code is avaliable at
https://github.com/HVision-NKU/MaskCLIPpp .
| [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 05:44:45 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Dec 2024 04:13:08 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 08:04:32 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zeng",
"Quan-Sheng",
""
],
[
"Li",
"Yunheng",
""
],
[
"Zhou",
"Daquan",
""
],
[
"Li",
"Guanbin",
""
],
[
"Hou",
"Qibin",
""
],
[
"Cheng",
"Ming-Ming",
""
]
]
| TITLE: High-Quality Mask Tuning Matters for Open-Vocabulary Segmentation
ABSTRACT: Open-vocabulary image segmentation has been advanced through the synergy
between mask generators and vision-language models like Contrastive
Language-Image Pre-training (CLIP). Previous approaches focus on generating
masks while aligning mask features with text embeddings during training. In
this paper, we observe that relying on generated low-quality masks can weaken
the alignment of vision and language in regional representations. This
motivates us to present a new fine-tuning framework, named MaskCLIP++, which
uses ground-truth masks instead of generated masks to enhance the mask
classification capability of CLIP. Due to the limited diversity of image
segmentation datasets with mask annotations, we propose incorporating a
consistency alignment principle during fine-tuning, which alleviates
categorical bias toward the fine-tuning dataset. After low-cost fine-tuning,
MaskCLIP++ significantly improves the mask classification performance on
multi-domain datasets. Combining with the mask generator in previous
state-of-the-art mask-based open vocabulary segmentation methods, we achieve
performance improvements of +1.7, +2.3, +2.1, +3.1, and +0.3 mIoU on the A-847,
PC-459, A-150, PC-59, and PAS-20 datasets, respectively. Code is avaliable at
https://github.com/HVision-NKU/MaskCLIPpp .
| no_new_dataset | 0.948585 |
2412.14569 | Chaoqun Liu | Chaoqun Liu, Xuanpeng Li, Chen Gong, Guangyu Li | Global Spatio-Temporal Fusion-based Traffic Prediction Algorithm with
Anomaly Aware | null | GLOBECOM 2024 - 2024 IEEE Global Communications Conference | 10.1109/GLOBECOM52923.2024.10901114 | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Traffic prediction is an indispensable component of urban planning and
traffic management. Achieving accurate traffic prediction hinges on the ability
to capture the potential spatio-temporal relationships among road sensors.
However, the majority of existing works focus on local short-term
spatio-temporal correlations, failing to fully consider the interactions of
different sensors in the long-term state. In addition, these works do not
analyze the influences of anomalous factors, or have insufficient ability to
extract personalized features of anomalous factors, which make them
ineffectively capture their spatio-temporal influences on traffic prediction.
To address the aforementioned issues, We propose a global spatio-temporal
fusion-based traffic prediction algorithm that incorporates anomaly awareness.
Initially, based on the designed anomaly detection network, we construct an
efficient anomalous factors impacting module (AFIM), to evaluate the
spatio-temporal impact of unexpected external events on traffic prediction.
Furthermore, we propose a multi-scale spatio-temporal feature fusion module
(MTSFFL) based on the transformer architecture, to obtain all possible both
long and short term correlations among different sensors in a wide-area traffic
environment for accurate prediction of traffic flow. Finally, experiments are
implemented based on real-scenario public transportation datasets (PEMS04 and
PEMS08) to demonstrate that our approach can achieve state-of-the-art
performance.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 06:40:21 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Liu",
"Chaoqun",
""
],
[
"Li",
"Xuanpeng",
""
],
[
"Gong",
"Chen",
""
],
[
"Li",
"Guangyu",
""
]
]
| TITLE: Global Spatio-Temporal Fusion-based Traffic Prediction Algorithm with
Anomaly Aware
ABSTRACT: Traffic prediction is an indispensable component of urban planning and
traffic management. Achieving accurate traffic prediction hinges on the ability
to capture the potential spatio-temporal relationships among road sensors.
However, the majority of existing works focus on local short-term
spatio-temporal correlations, failing to fully consider the interactions of
different sensors in the long-term state. In addition, these works do not
analyze the influences of anomalous factors, or have insufficient ability to
extract personalized features of anomalous factors, which make them
ineffectively capture their spatio-temporal influences on traffic prediction.
To address the aforementioned issues, We propose a global spatio-temporal
fusion-based traffic prediction algorithm that incorporates anomaly awareness.
Initially, based on the designed anomaly detection network, we construct an
efficient anomalous factors impacting module (AFIM), to evaluate the
spatio-temporal impact of unexpected external events on traffic prediction.
Furthermore, we propose a multi-scale spatio-temporal feature fusion module
(MTSFFL) based on the transformer architecture, to obtain all possible both
long and short term correlations among different sensors in a wide-area traffic
environment for accurate prediction of traffic flow. Finally, experiments are
implemented based on real-scenario public transportation datasets (PEMS04 and
PEMS08) to demonstrate that our approach can achieve state-of-the-art
performance.
| no_new_dataset | 0.945901 |
2412.15341 | Reza Shirkavand | Reza Shirkavand, Peiran Yu, Shangqian Gao, Gowthami Somepalli, Tom
Goldstein, Heng Huang | Efficient Fine-Tuning and Concept Suppression for Pruned Diffusion
Models | CVPR 2025 | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in diffusion generative models have yielded remarkable
progress. While the quality of generated content continues to improve, these
models have grown considerably in size and complexity. This increasing
computational burden poses significant challenges, particularly in
resource-constrained deployment scenarios such as mobile devices. The
combination of model pruning and knowledge distillation has emerged as a
promising solution to reduce computational demands while preserving generation
quality. However, this technique inadvertently propagates undesirable
behaviors, including the generation of copyrighted content and unsafe concepts,
even when such instances are absent from the fine-tuning dataset. In this
paper, we propose a novel bilevel optimization framework for pruned diffusion
models that consolidates the fine-tuning and unlearning processes into a
unified phase. Our approach maintains the principal advantages of
distillation-namely, efficient convergence and style transfer
capabilities-while selectively suppressing the generation of unwanted content.
This plug-in framework is compatible with various pruning and concept
unlearning methods, facilitating efficient, safe deployment of diffusion models
in controlled environments.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 19:13:18 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 20:52:10 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Shirkavand",
"Reza",
""
],
[
"Yu",
"Peiran",
""
],
[
"Gao",
"Shangqian",
""
],
[
"Somepalli",
"Gowthami",
""
],
[
"Goldstein",
"Tom",
""
],
[
"Huang",
"Heng",
""
]
]
| TITLE: Efficient Fine-Tuning and Concept Suppression for Pruned Diffusion
Models
ABSTRACT: Recent advances in diffusion generative models have yielded remarkable
progress. While the quality of generated content continues to improve, these
models have grown considerably in size and complexity. This increasing
computational burden poses significant challenges, particularly in
resource-constrained deployment scenarios such as mobile devices. The
combination of model pruning and knowledge distillation has emerged as a
promising solution to reduce computational demands while preserving generation
quality. However, this technique inadvertently propagates undesirable
behaviors, including the generation of copyrighted content and unsafe concepts,
even when such instances are absent from the fine-tuning dataset. In this
paper, we propose a novel bilevel optimization framework for pruned diffusion
models that consolidates the fine-tuning and unlearning processes into a
unified phase. Our approach maintains the principal advantages of
distillation-namely, efficient convergence and style transfer
capabilities-while selectively suppressing the generation of unwanted content.
This plug-in framework is compatible with various pruning and concept
unlearning methods, facilitating efficient, safe deployment of diffusion models
in controlled environments.
| no_new_dataset | 0.945651 |
2412.19950 | Christian Friedrich | Eric Hirsch and Christian Friedrich | Data-driven tool wear prediction in milling, based on a
process-integrated single-sensor approach | This preprint has been submitted to Robotics and Computer-Integrated
Manufacturing for possible publication ,14 pages, 12 figures | null | null | null | cs.LG cs.RO eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate tool wear prediction is essential for maintaining productivity and
minimizing costs in machining. However, the complex nature of the tool wear
process poses significant challenges to achieving reliable predictions. This
study explores data-driven methods, in particular deep learning, for tool wear
prediction. Traditional data-driven approaches often focus on a single process,
relying on multi-sensor setups and extensive data generation, which limits
generalization to new settings. Moreover, multi-sensor integration is often
impractical in industrial environments. To address these limitations, this
research investigates the transferability of predictive models using minimal
training data, validated across two processes. Furthermore, it uses a simple
setup with a single acceleration sensor to establish a low-cost data generation
approach that facilitates the generalization of models to other processes via
transfer learning. The study evaluates several machine learning models,
including transformer-inspired convolutional neural networks (CNN), long
short-term memory networks (LSTM), support vector machines (SVM), and decision
trees, trained on different input formats such as feature vectors and
short-time Fourier transform (STFT). The performance of the models is evaluated
on two machines and on different amounts of training data, including scenarios
with significantly reduced datasets, providing insight into their effectiveness
under constrained data conditions. The results demonstrate the potential of
specific models and configurations for effective tool wear prediction,
contributing to the development of more adaptable and efficient predictive
maintenance strategies in machining. Notably, the ConvNeXt model has an
exceptional performance, achieving 99.1\% accuracy in identifying tool wear
using data from only four milling tools operated until they are worn.
| [
{
"version": "v1",
"created": "Fri, 27 Dec 2024 23:10:32 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Jan 2025 14:35:01 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Mar 2025 18:20:38 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Hirsch",
"Eric",
""
],
[
"Friedrich",
"Christian",
""
]
]
| TITLE: Data-driven tool wear prediction in milling, based on a
process-integrated single-sensor approach
ABSTRACT: Accurate tool wear prediction is essential for maintaining productivity and
minimizing costs in machining. However, the complex nature of the tool wear
process poses significant challenges to achieving reliable predictions. This
study explores data-driven methods, in particular deep learning, for tool wear
prediction. Traditional data-driven approaches often focus on a single process,
relying on multi-sensor setups and extensive data generation, which limits
generalization to new settings. Moreover, multi-sensor integration is often
impractical in industrial environments. To address these limitations, this
research investigates the transferability of predictive models using minimal
training data, validated across two processes. Furthermore, it uses a simple
setup with a single acceleration sensor to establish a low-cost data generation
approach that facilitates the generalization of models to other processes via
transfer learning. The study evaluates several machine learning models,
including transformer-inspired convolutional neural networks (CNN), long
short-term memory networks (LSTM), support vector machines (SVM), and decision
trees, trained on different input formats such as feature vectors and
short-time Fourier transform (STFT). The performance of the models is evaluated
on two machines and on different amounts of training data, including scenarios
with significantly reduced datasets, providing insight into their effectiveness
under constrained data conditions. The results demonstrate the potential of
specific models and configurations for effective tool wear prediction,
contributing to the development of more adaptable and efficient predictive
maintenance strategies in machining. Notably, the ConvNeXt model has an
exceptional performance, achieving 99.1\% accuracy in identifying tool wear
using data from only four milling tools operated until they are worn.
| no_new_dataset | 0.944791 |
2412.21197 | Yang Chen | Yang Chen, Sheng Guo, Bo Zheng and Limin Wang | A Large-Scale Study on Video Action Dataset Condensation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recently, dataset condensation has made significant progress in the image
domain. Unlike images, videos possess an additional temporal dimension, which
harbors considerable redundant information, making condensation even more
crucial. However, video dataset condensation still remains an underexplored
area. We aim to bridge this gap by providing a large-scale study with
systematic design and fair comparison. Specifically, our work delves into three
key aspects to provide valuable empirical insights: (1) temporal processing of
video data, (2) the evaluation protocol for video dataset condensation, and (3)
adaptation of condensation algorithms to the space-time domain. From this
study, we derive several intriguing observations: (i) labeling methods greatly
influence condensation performance, (ii) simple sliding-window sampling is
effective for temporal processing, and (iii) dataset distillation methods
perform better in challenging scenarios, while sample selection methods excel
in easier ones. Furthermore, we propose a unified evaluation protocol for the
fair comparison of different condensation algorithms and achieve
state-of-the-art results on four widely-used action recognition datasets:
HMDB51, UCF101, SSv2 and K400. Our code is available at
https://github.com/MCG-NJU/Video-DC.
| [
{
"version": "v1",
"created": "Mon, 30 Dec 2024 18:58:29 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 03:28:28 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chen",
"Yang",
""
],
[
"Guo",
"Sheng",
""
],
[
"Zheng",
"Bo",
""
],
[
"Wang",
"Limin",
""
]
]
| TITLE: A Large-Scale Study on Video Action Dataset Condensation
ABSTRACT: Recently, dataset condensation has made significant progress in the image
domain. Unlike images, videos possess an additional temporal dimension, which
harbors considerable redundant information, making condensation even more
crucial. However, video dataset condensation still remains an underexplored
area. We aim to bridge this gap by providing a large-scale study with
systematic design and fair comparison. Specifically, our work delves into three
key aspects to provide valuable empirical insights: (1) temporal processing of
video data, (2) the evaluation protocol for video dataset condensation, and (3)
adaptation of condensation algorithms to the space-time domain. From this
study, we derive several intriguing observations: (i) labeling methods greatly
influence condensation performance, (ii) simple sliding-window sampling is
effective for temporal processing, and (iii) dataset distillation methods
perform better in challenging scenarios, while sample selection methods excel
in easier ones. Furthermore, we propose a unified evaluation protocol for the
fair comparison of different condensation algorithms and achieve
state-of-the-art results on four widely-used action recognition datasets:
HMDB51, UCF101, SSv2 and K400. Our code is available at
https://github.com/MCG-NJU/Video-DC.
| no_new_dataset | 0.9463 |
2501.01046 | Youngjun Son | Youngjun Son, Chaewon Kim, Jaejin Lee | FED: Fast and Efficient Dataset Deduplication Framework with GPU
Acceleration | 13 pages, 4 figures | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Dataset deduplication plays a crucial role in enhancing data quality,
ultimately improving the training performance and efficiency of large language
models. A commonly used method for data deduplication is the MinHash LSH
algorithm. Recently, NVIDIA introduced a GPU-based MinHash LSH deduplication
method, but it remains suboptimal, leaving room for further improvement in
processing efficiency. This paper proposes a GPU-accelerated deduplication
framework, FED, that optimizes MinHash LSH for GPU clusters and leverages
computationally efficient, partially reusable non-cryptographic hash functions.
FED significantly outperforms the CPU-based deduplication tool in SlimPajama
(using 64 logical CPU cores) by up to 107.2 times and the GPU-based tool in
NVIDIA NeMo Curator by up to 6.3 times when processing 30 million documents on
a node with four GPUs. Notably, our method dramatically accelerates the
previously time-consuming MinHash signature generation phase, achieving
speed-ups of up to 260 compared to the CPU baseline. Despite these gains in
efficiency, FED maintains high deduplication quality, with the duplicate
document sets reaching a Jaccard similarity of over 0.96 compared to those
identified by the standard MinHash algorithm. In large-scale experiments, the
deduplication of 1.2 trillion tokens is completed in just 6 hours in a
four-node, 16-GPU environment. The related code is publicly available on GitHub
(\href{https://github.com/mcrl/FED}{https://github.com/mcrl/FED}).
| [
{
"version": "v1",
"created": "Thu, 2 Jan 2025 04:11:23 GMT"
},
{
"version": "v2",
"created": "Sun, 16 Feb 2025 07:56:11 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 13:36:32 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Son",
"Youngjun",
""
],
[
"Kim",
"Chaewon",
""
],
[
"Lee",
"Jaejin",
""
]
]
| TITLE: FED: Fast and Efficient Dataset Deduplication Framework with GPU
Acceleration
ABSTRACT: Dataset deduplication plays a crucial role in enhancing data quality,
ultimately improving the training performance and efficiency of large language
models. A commonly used method for data deduplication is the MinHash LSH
algorithm. Recently, NVIDIA introduced a GPU-based MinHash LSH deduplication
method, but it remains suboptimal, leaving room for further improvement in
processing efficiency. This paper proposes a GPU-accelerated deduplication
framework, FED, that optimizes MinHash LSH for GPU clusters and leverages
computationally efficient, partially reusable non-cryptographic hash functions.
FED significantly outperforms the CPU-based deduplication tool in SlimPajama
(using 64 logical CPU cores) by up to 107.2 times and the GPU-based tool in
NVIDIA NeMo Curator by up to 6.3 times when processing 30 million documents on
a node with four GPUs. Notably, our method dramatically accelerates the
previously time-consuming MinHash signature generation phase, achieving
speed-ups of up to 260 compared to the CPU baseline. Despite these gains in
efficiency, FED maintains high deduplication quality, with the duplicate
document sets reaching a Jaccard similarity of over 0.96 compared to those
identified by the standard MinHash algorithm. In large-scale experiments, the
deduplication of 1.2 trillion tokens is completed in just 6 hours in a
four-node, 16-GPU environment. The related code is publicly available on GitHub
(\href{https://github.com/mcrl/FED}{https://github.com/mcrl/FED}).
| no_new_dataset | 0.952264 |
2501.02749 | Zhongjin Xu | Hao Luo, Jianjun Wei, Shuchen Zhao, Ankai Liang, Zhongjin Xu, Ruxue
Jiang | Intelligent logistics management robot path planning algorithm
integrating transformer and GCN network | 21 pages | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This research delves into advanced route optimization for robots in smart
logistics, leveraging a fusion of Transformer architectures, Graph Neural
Networks (GNNs), and Generative Adversarial Networks (GANs). The approach
utilizes a graph-based representation encompassing geographical data, cargo
allocation, and robot dynamics, addressing both spatial and resource
limitations to refine route efficiency. Through extensive testing with
authentic logistics datasets, the proposed method achieves notable
improvements, including a 15% reduction in travel distance, a 20% boost in time
efficiency, and a 10% decrease in energy consumption. These findings highlight
the algorithm's effectiveness, promoting enhanced performance in intelligent
logistics operations.
| [
{
"version": "v1",
"created": "Mon, 6 Jan 2025 03:53:02 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 03:29:21 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Luo",
"Hao",
""
],
[
"Wei",
"Jianjun",
""
],
[
"Zhao",
"Shuchen",
""
],
[
"Liang",
"Ankai",
""
],
[
"Xu",
"Zhongjin",
""
],
[
"Jiang",
"Ruxue",
""
]
]
| TITLE: Intelligent logistics management robot path planning algorithm
integrating transformer and GCN network
ABSTRACT: This research delves into advanced route optimization for robots in smart
logistics, leveraging a fusion of Transformer architectures, Graph Neural
Networks (GNNs), and Generative Adversarial Networks (GANs). The approach
utilizes a graph-based representation encompassing geographical data, cargo
allocation, and robot dynamics, addressing both spatial and resource
limitations to refine route efficiency. Through extensive testing with
authentic logistics datasets, the proposed method achieves notable
improvements, including a 15% reduction in travel distance, a 20% boost in time
efficiency, and a 10% decrease in energy consumption. These findings highlight
the algorithm's effectiveness, promoting enhanced performance in intelligent
logistics operations.
| no_new_dataset | 0.94868 |
2501.05712 | Hyunwoo Ko | Guijin Son, Hyunwoo Ko, Dasol Choi | Multi-Step Reasoning in Korean and the Emergent Mirage | C3NLP @ NAACL 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | We introduce HRMCR (HAE-RAE Multi-Step Commonsense Reasoning), a benchmark
designed to evaluate large language models' ability to perform multi-step
reasoning in culturally specific contexts, focusing on Korean. The questions
are automatically generated via templates and algorithms, requiring LLMs to
integrate Korean cultural knowledge into sequential reasoning steps. Consistent
with prior observations on emergent abilities, our experiments reveal that
models trained on fewer than \(2 \cdot 10^{25}\) training FLOPs struggle to
solve any questions, showing near-zero performance. Beyond this threshold,
performance improves sharply. State-of-the-art models (e.g., O1) still score
under 50\%, underscoring the difficulty of our tasks. Notably, stepwise
analysis suggests the observed emergent behavior may stem from compounding
errors across multiple steps rather than reflecting a genuinely new capability.
We publicly release the benchmark and commit to regularly updating the dataset
to prevent contamination.
| [
{
"version": "v1",
"created": "Fri, 10 Jan 2025 05:07:27 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:45:28 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Son",
"Guijin",
""
],
[
"Ko",
"Hyunwoo",
""
],
[
"Choi",
"Dasol",
""
]
]
| TITLE: Multi-Step Reasoning in Korean and the Emergent Mirage
ABSTRACT: We introduce HRMCR (HAE-RAE Multi-Step Commonsense Reasoning), a benchmark
designed to evaluate large language models' ability to perform multi-step
reasoning in culturally specific contexts, focusing on Korean. The questions
are automatically generated via templates and algorithms, requiring LLMs to
integrate Korean cultural knowledge into sequential reasoning steps. Consistent
with prior observations on emergent abilities, our experiments reveal that
models trained on fewer than \(2 \cdot 10^{25}\) training FLOPs struggle to
solve any questions, showing near-zero performance. Beyond this threshold,
performance improves sharply. State-of-the-art models (e.g., O1) still score
under 50\%, underscoring the difficulty of our tasks. Notably, stepwise
analysis suggests the observed emergent behavior may stem from compounding
errors across multiple steps rather than reflecting a genuinely new capability.
We publicly release the benchmark and commit to regularly updating the dataset
to prevent contamination.
| new_dataset | 0.946892 |
2501.05757 | Seungjoo Shin | Seungjoo Shin, Jaesik Park, Sunghyun Cho | Locality-aware Gaussian Compression for Fast and High-quality Rendering | Accepted to ICLR 2025. Project page:
https://seungjooshin.github.io/LocoGS | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present LocoGS, a locality-aware 3D Gaussian Splatting (3DGS) framework
that exploits the spatial coherence of 3D Gaussians for compact modeling of
volumetric scenes. To this end, we first analyze the local coherence of 3D
Gaussian attributes, and propose a novel locality-aware 3D Gaussian
representation that effectively encodes locally-coherent Gaussian attributes
using a neural field representation with a minimal storage requirement. On top
of the novel representation, LocoGS is carefully designed with additional
components such as dense initialization, an adaptive spherical harmonics
bandwidth scheme and different encoding schemes for different Gaussian
attributes to maximize compression performance. Experimental results
demonstrate that our approach outperforms the rendering quality of existing
compact Gaussian representations for representative real-world 3D datasets
while achieving from 54.6$\times$ to 96.6$\times$ compressed storage size and
from 2.1$\times$ to 2.4$\times$ rendering speed than 3DGS. Even our approach
also demonstrates an averaged 2.4$\times$ higher rendering speed than the
state-of-the-art compression method with comparable compression performance.
| [
{
"version": "v1",
"created": "Fri, 10 Jan 2025 07:19:41 GMT"
},
{
"version": "v2",
"created": "Mon, 3 Mar 2025 07:07:28 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Mar 2025 11:12:31 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Shin",
"Seungjoo",
""
],
[
"Park",
"Jaesik",
""
],
[
"Cho",
"Sunghyun",
""
]
]
| TITLE: Locality-aware Gaussian Compression for Fast and High-quality Rendering
ABSTRACT: We present LocoGS, a locality-aware 3D Gaussian Splatting (3DGS) framework
that exploits the spatial coherence of 3D Gaussians for compact modeling of
volumetric scenes. To this end, we first analyze the local coherence of 3D
Gaussian attributes, and propose a novel locality-aware 3D Gaussian
representation that effectively encodes locally-coherent Gaussian attributes
using a neural field representation with a minimal storage requirement. On top
of the novel representation, LocoGS is carefully designed with additional
components such as dense initialization, an adaptive spherical harmonics
bandwidth scheme and different encoding schemes for different Gaussian
attributes to maximize compression performance. Experimental results
demonstrate that our approach outperforms the rendering quality of existing
compact Gaussian representations for representative real-world 3D datasets
while achieving from 54.6$\times$ to 96.6$\times$ compressed storage size and
from 2.1$\times$ to 2.4$\times$ rendering speed than 3DGS. Even our approach
also demonstrates an averaged 2.4$\times$ higher rendering speed than the
state-of-the-art compression method with comparable compression performance.
| no_new_dataset | 0.94743 |
2501.06557 | Marco Giordano | Marco Giordano, Claudia Rinaldi | A Survey on Spoken Italian Datasets and Corpora | Published on IEEE Access Journal on Feb 2025 | in IEEE Access, vol. 13, pp. 29190-29205, 2025 | 10.1109/ACCESS.2025.3538952 | null | cs.CL cs.AI cs.DL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Spoken language datasets are vital for advancing linguistic research, Natural
Language Processing, and speech technology. However, resources dedicated to
Italian, a linguistically rich and diverse Romance language, remain
underexplored compared to major languages like English or Mandarin. This survey
provides a comprehensive analysis of 66 spoken Italian datasets, highlighting
their characteristics, methodologies, and applications. The datasets are
categorized by speech type, source and context, and demographic and linguistic
features, with a focus on their utility in fields such as Automatic Speech
Recognition, emotion detection, and education. Challenges related to dataset
scarcity, representativeness, and accessibility are discussed alongside
recommendations for enhancing dataset creation and utilization. The full
dataset inventory is publicly accessible via GitHub and archived on Zenodo,
serving as a valuable resource for researchers and developers. By addressing
current gaps and proposing future directions, this work aims to support the
advancement of Italian speech technologies and linguistic research.
| [
{
"version": "v1",
"created": "Sat, 11 Jan 2025 14:33:57 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 13:59:29 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Giordano",
"Marco",
""
],
[
"Rinaldi",
"Claudia",
""
]
]
| TITLE: A Survey on Spoken Italian Datasets and Corpora
ABSTRACT: Spoken language datasets are vital for advancing linguistic research, Natural
Language Processing, and speech technology. However, resources dedicated to
Italian, a linguistically rich and diverse Romance language, remain
underexplored compared to major languages like English or Mandarin. This survey
provides a comprehensive analysis of 66 spoken Italian datasets, highlighting
their characteristics, methodologies, and applications. The datasets are
categorized by speech type, source and context, and demographic and linguistic
features, with a focus on their utility in fields such as Automatic Speech
Recognition, emotion detection, and education. Challenges related to dataset
scarcity, representativeness, and accessibility are discussed alongside
recommendations for enhancing dataset creation and utilization. The full
dataset inventory is publicly accessible via GitHub and archived on Zenodo,
serving as a valuable resource for researchers and developers. By addressing
current gaps and proposing future directions, this work aims to support the
advancement of Italian speech technologies and linguistic research.
| no_new_dataset | 0.938745 |
2501.08333 | Hyeonwoo Kim | Hyeonwoo Kim, Sangwon Beak, Hanbyul Joo | DAViD: Modeling Dynamic Affordance of 3D Objects using Pre-trained Video
Diffusion Models | Project Page: https://snuvclab.github.io/david/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Modeling how humans interact with objects is crucial for AI to effectively
assist or mimic human behaviors. Existing studies for learning such ability
primarily focus on static human-object interaction (HOI) patterns, such as
contact and spatial relationships, while dynamic HOI patterns, capturing the
movement of humans and objects over time, remain relatively underexplored. In
this paper, we present a novel framework for learning Dynamic Affordance across
various target object categories. To address the scarcity of 4D HOI datasets,
our method learns the 3D dynamic affordance from synthetically generated 4D HOI
samples. Specifically, we propose a pipeline that first generates 2D HOI videos
from a given 3D target object using a pre-trained video diffusion model, then
lifts them into 3D to generate 4D HOI samples. Leveraging these synthesized 4D
HOI samples, we train DAViD, our generative 4D human-object interaction model,
which is composed of two key components: (1) a human motion diffusion model
(MDM) with Low-Rank Adaptation (LoRA) module to fine-tune a pre-trained MDM to
learn the HOI motion concepts from limited HOI motion samples, (2) a motion
diffusion model for 4D object poses conditioned by produced human interaction
motions. Interestingly, DAViD can integrate newly learned HOI motion concepts
with pre-trained human motions to create novel HOI motions, even for multiple
HOI motion concepts, demonstrating the advantage of our pipeline with LoRA in
integrating dynamic HOI concepts. Through extensive experiments, we demonstrate
that DAViD outperforms baselines in synthesizing HOI motion.
| [
{
"version": "v1",
"created": "Tue, 14 Jan 2025 18:59:59 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 21:35:21 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Kim",
"Hyeonwoo",
""
],
[
"Beak",
"Sangwon",
""
],
[
"Joo",
"Hanbyul",
""
]
]
| TITLE: DAViD: Modeling Dynamic Affordance of 3D Objects using Pre-trained Video
Diffusion Models
ABSTRACT: Modeling how humans interact with objects is crucial for AI to effectively
assist or mimic human behaviors. Existing studies for learning such ability
primarily focus on static human-object interaction (HOI) patterns, such as
contact and spatial relationships, while dynamic HOI patterns, capturing the
movement of humans and objects over time, remain relatively underexplored. In
this paper, we present a novel framework for learning Dynamic Affordance across
various target object categories. To address the scarcity of 4D HOI datasets,
our method learns the 3D dynamic affordance from synthetically generated 4D HOI
samples. Specifically, we propose a pipeline that first generates 2D HOI videos
from a given 3D target object using a pre-trained video diffusion model, then
lifts them into 3D to generate 4D HOI samples. Leveraging these synthesized 4D
HOI samples, we train DAViD, our generative 4D human-object interaction model,
which is composed of two key components: (1) a human motion diffusion model
(MDM) with Low-Rank Adaptation (LoRA) module to fine-tune a pre-trained MDM to
learn the HOI motion concepts from limited HOI motion samples, (2) a motion
diffusion model for 4D object poses conditioned by produced human interaction
motions. Interestingly, DAViD can integrate newly learned HOI motion concepts
with pre-trained human motions to create novel HOI motions, even for multiple
HOI motion concepts, demonstrating the advantage of our pipeline with LoRA in
integrating dynamic HOI concepts. Through extensive experiments, we demonstrate
that DAViD outperforms baselines in synthesizing HOI motion.
| no_new_dataset | 0.93835 |
2501.12106 | Stefan Lenz | Stefan Lenz, Arsenij Ustjanzew, Marco Jeray, Torsten Panholzer | Can open source large language models be used for tumor documentation in
Germany? -- An evaluation on urological doctors' notes | 48 pages, 5 figures | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tumor documentation in Germany is largely done manually, requiring reading
patient records and entering data into structured databases. Large language
models (LLMs) could potentially enhance this process by improving efficiency
and reliability. This evaluation tests eleven different open source LLMs with
sizes ranging from 1-70 billion model parameters on three basic tasks of the
tumor documentation process: identifying tumor diagnoses, assigning ICD-10
codes, and extracting the date of first diagnosis. For evaluating the LLMs on
these tasks, a dataset of annotated text snippets based on anonymized doctors'
notes from urology was prepared. Different prompting strategies were used to
investigate the effect of the number of examples in few-shot prompting and to
explore the capabilities of the LLMs in general. The models Llama 3.1 8B,
Mistral 7B, and Mistral NeMo 12 B performed comparably well in the tasks.
Models with less extensive training data or having fewer than 7 billion
parameters showed notably lower performance, while larger models did not
display performance gains. Examples from a different medical domain than
urology could also improve the outcome in few-shot prompting, which
demonstrates the ability of LLMs to handle tasks needed for tumor
documentation. Open source LLMs show a strong potential for automating tumor
documentation. Models from 7-12 billion parameters could offer an optimal
balance between performance and resource efficiency. With tailored fine-tuning
and well-designed prompting, these models might become important tools for
clinical documentation in the future. The code for the evaluation is available
from https://github.com/stefan-m-lenz/UroLlmEval. We also release the dataset
as a new valuable resource that addresses the shortage of authentic and easily
accessible benchmarks in German-language medical NLP.
| [
{
"version": "v1",
"created": "Tue, 21 Jan 2025 12:56:47 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 08:48:46 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Lenz",
"Stefan",
""
],
[
"Ustjanzew",
"Arsenij",
""
],
[
"Jeray",
"Marco",
""
],
[
"Panholzer",
"Torsten",
""
]
]
| TITLE: Can open source large language models be used for tumor documentation in
Germany? -- An evaluation on urological doctors' notes
ABSTRACT: Tumor documentation in Germany is largely done manually, requiring reading
patient records and entering data into structured databases. Large language
models (LLMs) could potentially enhance this process by improving efficiency
and reliability. This evaluation tests eleven different open source LLMs with
sizes ranging from 1-70 billion model parameters on three basic tasks of the
tumor documentation process: identifying tumor diagnoses, assigning ICD-10
codes, and extracting the date of first diagnosis. For evaluating the LLMs on
these tasks, a dataset of annotated text snippets based on anonymized doctors'
notes from urology was prepared. Different prompting strategies were used to
investigate the effect of the number of examples in few-shot prompting and to
explore the capabilities of the LLMs in general. The models Llama 3.1 8B,
Mistral 7B, and Mistral NeMo 12 B performed comparably well in the tasks.
Models with less extensive training data or having fewer than 7 billion
parameters showed notably lower performance, while larger models did not
display performance gains. Examples from a different medical domain than
urology could also improve the outcome in few-shot prompting, which
demonstrates the ability of LLMs to handle tasks needed for tumor
documentation. Open source LLMs show a strong potential for automating tumor
documentation. Models from 7-12 billion parameters could offer an optimal
balance between performance and resource efficiency. With tailored fine-tuning
and well-designed prompting, these models might become important tools for
clinical documentation in the future. The code for the evaluation is available
from https://github.com/stefan-m-lenz/UroLlmEval. We also release the dataset
as a new valuable resource that addresses the shortage of authentic and easily
accessible benchmarks in German-language medical NLP.
| no_new_dataset | 0.641113 |
2501.14198 | Zeyun Deng | Zeyun Deng, Joseph Campbell | Sparse Mixture-of-Experts for Non-Uniform Noise Reduction in MRI Images | Accepted to the WACV Workshop on Image Quality | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Magnetic Resonance Imaging (MRI) is an essential diagnostic tool in clinical
settings but its utility is often hindered by noise artifacts introduced during
the imaging process. Effective denoising is critical for enhancing image
quality while preserving anatomical structures. However traditional denoising
methods which typically assume uniform noise distributions struggle to handle
the non-uniform noise commonly present in MRI images. In this paper we
introduce a novel approach leveraging a sparse mixture-of-experts framework for
MRI image denoising. Each expert is a specialized denoising convolutional
neural network fine-tuned to target specific noise characteristics associated
with different image regions. Our method demonstrates superior performance over
state-of-the-art denoising techniques on both synthetic and real-world MRI
datasets. Furthermore we show that it generalizes effectively to unseen
datasets highlighting its robustness and adaptability.
| [
{
"version": "v1",
"created": "Fri, 24 Jan 2025 03:04:44 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 02:32:20 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Deng",
"Zeyun",
""
],
[
"Campbell",
"Joseph",
""
]
]
| TITLE: Sparse Mixture-of-Experts for Non-Uniform Noise Reduction in MRI Images
ABSTRACT: Magnetic Resonance Imaging (MRI) is an essential diagnostic tool in clinical
settings but its utility is often hindered by noise artifacts introduced during
the imaging process. Effective denoising is critical for enhancing image
quality while preserving anatomical structures. However traditional denoising
methods which typically assume uniform noise distributions struggle to handle
the non-uniform noise commonly present in MRI images. In this paper we
introduce a novel approach leveraging a sparse mixture-of-experts framework for
MRI image denoising. Each expert is a specialized denoising convolutional
neural network fine-tuned to target specific noise characteristics associated
with different image regions. Our method demonstrates superior performance over
state-of-the-art denoising techniques on both synthetic and real-world MRI
datasets. Furthermore we show that it generalizes effectively to unseen
datasets highlighting its robustness and adaptability.
| no_new_dataset | 0.946646 |
2501.14729 | Xin Zhou | Xin Zhou, Dingkang Liang, Sifan Tu, Xiwu Chen, Yikang Ding, Dingyuan
Zhang, Feiyang Tan, Hengshuang Zhao, Xiang Bai | HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene
Understanding and Generation | The code will be available at https://github.com/LMD0311/HERMES | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Driving World Models (DWMs) have become essential for autonomous driving by
enabling future scene prediction. However, existing DWMs are limited to scene
generation and fail to incorporate scene understanding, which involves
interpreting and reasoning about the driving environment. In this paper, we
present a unified Driving World Model named HERMES. We seamlessly integrate 3D
scene understanding and future scene evolution (generation) through a unified
framework in driving scenarios. Specifically, HERMES leverages a Bird's-Eye
View (BEV) representation to consolidate multi-view spatial information while
preserving geometric relationships and interactions. We also introduce world
queries, which incorporate world knowledge into BEV features via causal
attention in the Large Language Model, enabling contextual enrichment for
understanding and generation tasks. We conduct comprehensive studies on
nuScenes and OmniDrive-nuScenes datasets to validate the effectiveness of our
method. HERMES achieves state-of-the-art performance, reducing generation error
by 32.4% and improving understanding metrics such as CIDEr by 8.0%. The model
and code will be publicly released at https://github.com/LMD0311/HERMES.
| [
{
"version": "v1",
"created": "Fri, 24 Jan 2025 18:59:51 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 17:58:02 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Zhou",
"Xin",
""
],
[
"Liang",
"Dingkang",
""
],
[
"Tu",
"Sifan",
""
],
[
"Chen",
"Xiwu",
""
],
[
"Ding",
"Yikang",
""
],
[
"Zhang",
"Dingyuan",
""
],
[
"Tan",
"Feiyang",
""
],
[
"Zhao",
"Hengshuang",
""
],
[
"Bai",
"Xiang",
""
]
]
| TITLE: HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene
Understanding and Generation
ABSTRACT: Driving World Models (DWMs) have become essential for autonomous driving by
enabling future scene prediction. However, existing DWMs are limited to scene
generation and fail to incorporate scene understanding, which involves
interpreting and reasoning about the driving environment. In this paper, we
present a unified Driving World Model named HERMES. We seamlessly integrate 3D
scene understanding and future scene evolution (generation) through a unified
framework in driving scenarios. Specifically, HERMES leverages a Bird's-Eye
View (BEV) representation to consolidate multi-view spatial information while
preserving geometric relationships and interactions. We also introduce world
queries, which incorporate world knowledge into BEV features via causal
attention in the Large Language Model, enabling contextual enrichment for
understanding and generation tasks. We conduct comprehensive studies on
nuScenes and OmniDrive-nuScenes datasets to validate the effectiveness of our
method. HERMES achieves state-of-the-art performance, reducing generation error
by 32.4% and improving understanding metrics such as CIDEr by 8.0%. The model
and code will be publicly released at https://github.com/LMD0311/HERMES.
| no_new_dataset | 0.9463 |
2501.17202 | Chen Chen | Chen Chen, Yuchen Hu, Siyin Wang, Helin Wang, Zhehuai Chen, Chao
Zhang, Chao-Han Huck Yang, and Eng Siong Chng | Audio Large Language Models Can Be Descriptive Speech Quality Evaluators | ICLR 2025 | null | null | null | cs.SD cs.CL eess.AS | http://creativecommons.org/licenses/by/4.0/ | An ideal multimodal agent should be aware of the quality of its input
modalities. Recent advances have enabled large language models (LLMs) to
incorporate auditory systems for handling various speech-related tasks.
However, most audio LLMs remain unaware of the quality of the speech they
process. This limitation arises because speech quality evaluation is typically
excluded from multi-task training due to the lack of suitable datasets. To
address this, we introduce the first natural language-based speech evaluation
corpus, generated from authentic human ratings. In addition to the overall Mean
Opinion Score (MOS), this corpus offers detailed analysis across multiple
dimensions and identifies causes of quality degradation. It also enables
descriptive comparisons between two speech samples (A/B tests) with human-like
judgment. Leveraging this corpus, we propose an alignment approach with LLM
distillation (ALLD) to guide the audio LLM in extracting relevant information
from raw speech and generating meaningful responses. Experimental results
demonstrate that ALLD outperforms the previous state-of-the-art regression
model in MOS prediction, with a mean square error of 0.17 and an A/B test
accuracy of 98.6%. Additionally, the generated responses achieve BLEU scores of
25.8 and 30.2 on two tasks, surpassing the capabilities of task-specific
models. This work advances the comprehensive perception of speech signals by
audio LLMs, contributing to the development of real-world auditory and sensory
intelligent agents.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 22:47:51 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 02:01:46 GMT"
}
]
| 2025-03-13T00:00:00 | [
[
"Chen",
"Chen",
""
],
[
"Hu",
"Yuchen",
""
],
[
"Wang",
"Siyin",
""
],
[
"Wang",
"Helin",
""
],
[
"Chen",
"Zhehuai",
""
],
[
"Zhang",
"Chao",
""
],
[
"Yang",
"Chao-Han Huck",
""
],
[
"Chng",
"Eng Siong",
""
]
]
| TITLE: Audio Large Language Models Can Be Descriptive Speech Quality Evaluators
ABSTRACT: An ideal multimodal agent should be aware of the quality of its input
modalities. Recent advances have enabled large language models (LLMs) to
incorporate auditory systems for handling various speech-related tasks.
However, most audio LLMs remain unaware of the quality of the speech they
process. This limitation arises because speech quality evaluation is typically
excluded from multi-task training due to the lack of suitable datasets. To
address this, we introduce the first natural language-based speech evaluation
corpus, generated from authentic human ratings. In addition to the overall Mean
Opinion Score (MOS), this corpus offers detailed analysis across multiple
dimensions and identifies causes of quality degradation. It also enables
descriptive comparisons between two speech samples (A/B tests) with human-like
judgment. Leveraging this corpus, we propose an alignment approach with LLM
distillation (ALLD) to guide the audio LLM in extracting relevant information
from raw speech and generating meaningful responses. Experimental results
demonstrate that ALLD outperforms the previous state-of-the-art regression
model in MOS prediction, with a mean square error of 0.17 and an A/B test
accuracy of 98.6%. Additionally, the generated responses achieve BLEU scores of
25.8 and 30.2 on two tasks, surpassing the capabilities of task-specific
models. This work advances the comprehensive perception of speech signals by
audio LLMs, contributing to the development of real-world auditory and sensory
intelligent agents.
| no_new_dataset | 0.652075 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.