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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2503.06136 | Ye Tao | Ye Tao, Jiawei Zhang, Yahao Shi, Dongqing Zou, Bin Zhou | GSV3D: Gaussian Splatting-based Geometric Distillation with Stable Video
Diffusion for Single-Image 3D Object Generation | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Image-based 3D generation has vast applications in robotics and gaming, where
high-quality, diverse outputs and consistent 3D representations are crucial.
However, existing methods have limitations: 3D diffusion models are limited by
dataset scarcity and the absence of strong pre-trained priors, while 2D
diffusion-based approaches struggle with geometric consistency. We propose a
method that leverages 2D diffusion models' implicit 3D reasoning ability while
ensuring 3D consistency via Gaussian-splatting-based geometric distillation.
Specifically, the proposed Gaussian Splatting Decoder enforces 3D consistency
by transforming SV3D latent outputs into an explicit 3D representation. Unlike
SV3D, which only relies on implicit 2D representations for video generation,
Gaussian Splatting explicitly encodes spatial and appearance attributes,
enabling multi-view consistency through geometric constraints. These
constraints correct view inconsistencies, ensuring robust geometric
consistency. As a result, our approach simultaneously generates high-quality,
multi-view-consistent images and accurate 3D models, providing a scalable
solution for single-image-based 3D generation and bridging the gap between 2D
Diffusion diversity and 3D structural coherence. Experimental results
demonstrate state-of-the-art multi-view consistency and strong generalization
across diverse datasets. The code will be made publicly available upon
acceptance.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 09:10:31 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Tao",
"Ye",
""
],
[
"Zhang",
"Jiawei",
""
],
[
"Shi",
"Yahao",
""
],
[
"Zou",
"Dongqing",
""
],
[
"Zhou",
"Bin",
""
]
]
| TITLE: GSV3D: Gaussian Splatting-based Geometric Distillation with Stable Video
Diffusion for Single-Image 3D Object Generation
ABSTRACT: Image-based 3D generation has vast applications in robotics and gaming, where
high-quality, diverse outputs and consistent 3D representations are crucial.
However, existing methods have limitations: 3D diffusion models are limited by
dataset scarcity and the absence of strong pre-trained priors, while 2D
diffusion-based approaches struggle with geometric consistency. We propose a
method that leverages 2D diffusion models' implicit 3D reasoning ability while
ensuring 3D consistency via Gaussian-splatting-based geometric distillation.
Specifically, the proposed Gaussian Splatting Decoder enforces 3D consistency
by transforming SV3D latent outputs into an explicit 3D representation. Unlike
SV3D, which only relies on implicit 2D representations for video generation,
Gaussian Splatting explicitly encodes spatial and appearance attributes,
enabling multi-view consistency through geometric constraints. These
constraints correct view inconsistencies, ensuring robust geometric
consistency. As a result, our approach simultaneously generates high-quality,
multi-view-consistent images and accurate 3D models, providing a scalable
solution for single-image-based 3D generation and bridging the gap between 2D
Diffusion diversity and 3D structural coherence. Experimental results
demonstrate state-of-the-art multi-view consistency and strong generalization
across diverse datasets. The code will be made publicly available upon
acceptance.
| no_new_dataset | 0.949059 |
2503.06140 | Xiaosen Wang | Bohan Liu, Xiaosen Wang | Boosting the Local Invariance for Better Adversarial Transferability | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transfer-based attacks pose a significant threat to real-world applications
by directly targeting victim models with adversarial examples generated on
surrogate models. While numerous approaches have been proposed to enhance
adversarial transferability, existing works often overlook the intrinsic
relationship between adversarial perturbations and input images. In this work,
we find that adversarial perturbation often exhibits poor translation
invariance for a given clean image and model, which is attributed to local
invariance. Through empirical analysis, we demonstrate that there is a positive
correlation between the local invariance of adversarial perturbations w.r.t.
the input image and their transferability across different models. Based on
this finding, we propose a general adversarial transferability boosting
technique called Local Invariance Boosting approach (LI-Boost). Extensive
experiments on the standard ImageNet dataset demonstrate that LI-Boost could
significantly boost various types of transfer-based attacks (e.g.,
gradient-based, input transformation-based, model-related, advanced objective
function, ensemble, etc.) on CNNs, ViTs, and defense mechanisms. Our approach
presents a promising direction for future research in improving adversarial
transferability across different models.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 09:44:45 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Liu",
"Bohan",
""
],
[
"Wang",
"Xiaosen",
""
]
]
| TITLE: Boosting the Local Invariance for Better Adversarial Transferability
ABSTRACT: Transfer-based attacks pose a significant threat to real-world applications
by directly targeting victim models with adversarial examples generated on
surrogate models. While numerous approaches have been proposed to enhance
adversarial transferability, existing works often overlook the intrinsic
relationship between adversarial perturbations and input images. In this work,
we find that adversarial perturbation often exhibits poor translation
invariance for a given clean image and model, which is attributed to local
invariance. Through empirical analysis, we demonstrate that there is a positive
correlation between the local invariance of adversarial perturbations w.r.t.
the input image and their transferability across different models. Based on
this finding, we propose a general adversarial transferability boosting
technique called Local Invariance Boosting approach (LI-Boost). Extensive
experiments on the standard ImageNet dataset demonstrate that LI-Boost could
significantly boost various types of transfer-based attacks (e.g.,
gradient-based, input transformation-based, model-related, advanced objective
function, ensemble, etc.) on CNNs, ViTs, and defense mechanisms. Our approach
presents a promising direction for future research in improving adversarial
transferability across different models.
| no_new_dataset | 0.941115 |
2503.06141 | Mingxing Li | Mingxing Li, Rui Wang, Lei Sun, Yancheng Bai, Xiangxiang Chu | Next Token Is Enough: Realistic Image Quality and Aesthetic Scoring with
Multimodal Large Language Model | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid expansion of mobile internet has resulted in a substantial increase
in user-generated content (UGC) images, thereby making the thorough assessment
of UGC images both urgent and essential. Recently, multimodal large language
models (MLLMs) have shown great potential in image quality assessment (IQA) and
image aesthetic assessment (IAA). Despite this progress, effectively scoring
the quality and aesthetics of UGC images still faces two main challenges: 1) A
single score is inadequate to capture the hierarchical human perception. 2) How
to use MLLMs to output numerical scores, such as mean opinion scores (MOS),
remains an open question. To address these challenges, we introduce a novel
dataset, named Realistic image Quality and Aesthetic (RealQA), including 14,715
UGC images, each of which is annoted with 10 fine-grained attributes. These
attributes span three levels: low level (e.g., image clarity), middle level
(e.g., subject integrity) and high level (e.g., composition). Besides, we
conduct a series of in-depth and comprehensive investigations into how to
effectively predict numerical scores using MLLMs. Surprisingly, by predicting
just two extra significant digits, the next token paradigm can achieve SOTA
performance. Furthermore, with the help of chain of thought (CoT) combined with
the learnt fine-grained attributes, the proposed method can outperform SOTA
methods on five public datasets for IQA and IAA with superior interpretability
and show strong zero-shot generalization for video quality assessment (VQA).
The code and dataset will be released.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 09:49:10 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Mingxing",
""
],
[
"Wang",
"Rui",
""
],
[
"Sun",
"Lei",
""
],
[
"Bai",
"Yancheng",
""
],
[
"Chu",
"Xiangxiang",
""
]
]
| TITLE: Next Token Is Enough: Realistic Image Quality and Aesthetic Scoring with
Multimodal Large Language Model
ABSTRACT: The rapid expansion of mobile internet has resulted in a substantial increase
in user-generated content (UGC) images, thereby making the thorough assessment
of UGC images both urgent and essential. Recently, multimodal large language
models (MLLMs) have shown great potential in image quality assessment (IQA) and
image aesthetic assessment (IAA). Despite this progress, effectively scoring
the quality and aesthetics of UGC images still faces two main challenges: 1) A
single score is inadequate to capture the hierarchical human perception. 2) How
to use MLLMs to output numerical scores, such as mean opinion scores (MOS),
remains an open question. To address these challenges, we introduce a novel
dataset, named Realistic image Quality and Aesthetic (RealQA), including 14,715
UGC images, each of which is annoted with 10 fine-grained attributes. These
attributes span three levels: low level (e.g., image clarity), middle level
(e.g., subject integrity) and high level (e.g., composition). Besides, we
conduct a series of in-depth and comprehensive investigations into how to
effectively predict numerical scores using MLLMs. Surprisingly, by predicting
just two extra significant digits, the next token paradigm can achieve SOTA
performance. Furthermore, with the help of chain of thought (CoT) combined with
the learnt fine-grained attributes, the proposed method can outperform SOTA
methods on five public datasets for IQA and IAA with superior interpretability
and show strong zero-shot generalization for video quality assessment (VQA).
The code and dataset will be released.
| new_dataset | 0.961171 |
2503.06142 | Xinan He | Xinan He, Yue Zhou, Bing Fan, Bin Li, Guopu Zhu, Feng Ding | VLForgery Face Triad: Detection, Localization and Attribution via
Multimodal Large Language Models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Faces synthesized by diffusion models (DMs) with high-quality and
controllable attributes pose a significant challenge for Deepfake detection.
Most state-of-the-art detectors only yield a binary decision, incapable of
forgery localization, attribution of forgery methods, and providing analysis on
the cause of forgeries. In this work, we integrate Multimodal Large Language
Models (MLLMs) within DM-based face forensics, and propose a fine-grained
analysis triad framework called VLForgery, that can 1) predict falsified facial
images; 2) locate the falsified face regions subjected to partial synthesis;
and 3) attribute the synthesis with specific generators. To achieve the above
goals, we introduce VLF (Visual Language Forensics), a novel and diverse
synthesis face dataset designed to facilitate rich interactions between Visual
and Language modalities in MLLMs. Additionally, we propose an extrinsic
knowledge-guided description method, termed EkCot, which leverages knowledge
from the image generation pipeline to enable MLLMs to quickly capture image
content. Furthermore, we introduce a low-level vision comparison pipeline
designed to identify differential features between real and fake that MLLMs can
inherently understand. These features are then incorporated into EkCot,
enhancing its ability to analyze forgeries in a structured manner, following
the sequence of detection, localization, and attribution. Extensive experiments
demonstrate that VLForgery outperforms other state-of-the-art forensic
approaches in detection accuracy, with additional potential for falsified
region localization and attribution analysis.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 09:55:19 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"He",
"Xinan",
""
],
[
"Zhou",
"Yue",
""
],
[
"Fan",
"Bing",
""
],
[
"Li",
"Bin",
""
],
[
"Zhu",
"Guopu",
""
],
[
"Ding",
"Feng",
""
]
]
| TITLE: VLForgery Face Triad: Detection, Localization and Attribution via
Multimodal Large Language Models
ABSTRACT: Faces synthesized by diffusion models (DMs) with high-quality and
controllable attributes pose a significant challenge for Deepfake detection.
Most state-of-the-art detectors only yield a binary decision, incapable of
forgery localization, attribution of forgery methods, and providing analysis on
the cause of forgeries. In this work, we integrate Multimodal Large Language
Models (MLLMs) within DM-based face forensics, and propose a fine-grained
analysis triad framework called VLForgery, that can 1) predict falsified facial
images; 2) locate the falsified face regions subjected to partial synthesis;
and 3) attribute the synthesis with specific generators. To achieve the above
goals, we introduce VLF (Visual Language Forensics), a novel and diverse
synthesis face dataset designed to facilitate rich interactions between Visual
and Language modalities in MLLMs. Additionally, we propose an extrinsic
knowledge-guided description method, termed EkCot, which leverages knowledge
from the image generation pipeline to enable MLLMs to quickly capture image
content. Furthermore, we introduce a low-level vision comparison pipeline
designed to identify differential features between real and fake that MLLMs can
inherently understand. These features are then incorporated into EkCot,
enhancing its ability to analyze forgeries in a structured manner, following
the sequence of detection, localization, and attribution. Extensive experiments
demonstrate that VLForgery outperforms other state-of-the-art forensic
approaches in detection accuracy, with additional potential for falsified
region localization and attribution analysis.
| new_dataset | 0.966188 |
2503.06149 | Xudong Wang | Xudong Wang, Jiacheng Wang, Lei Feng, Dusit Niyato, Ruichen Zhang,
Jiawen Kang, Zehui Xiong, Hongyang Du, Shiwen Mao | Wireless Hallucination in Generative AI-enabled Communications:
Concepts, Issues, and Solutions | 7 pages, 4 figures | null | null | null | cs.IT eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generative AI (GenAI) is driving the intelligence of wireless communications.
Due to data limitations, random generation, and dynamic environments, GenAI may
generate channel information or optimization strategies that violate physical
laws or deviate from actual real-world requirements. We refer to this
phenomenon as wireless hallucination, which results in invalid channel
information, spectrum wastage, and low communication reliability but remains
underexplored. To address this gap, this article provides a comprehensive
concept of wireless hallucinations in GenAI-driven communications, focusing on
hallucination mitigation. Specifically, we first introduce the fundamental,
analyze its causes based on the GenAI workflow, and propose mitigation
solutions at the data, model, and post-generation levels. Then, we
systematically examines representative hallucination scenarios in GenAI-enabled
communications and their corresponding solutions. Finally, we propose a novel
integrated mitigation solution for GenAI-based channel estimation. At the data
level, we establish a channel estimation hallucination dataset and employ
generative adversarial networks (GANs)-based data augmentation. Additionally,
we incorporate attention mechanisms and large language models (LLMs) to enhance
both training and inference performance. Experimental results demonstrate that
the proposed hybrid solutions reduce the normalized mean square error (NMSE) by
0.19, effectively reducing wireless hallucinations.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 10:16:24 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Xudong",
""
],
[
"Wang",
"Jiacheng",
""
],
[
"Feng",
"Lei",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Zhang",
"Ruichen",
""
],
[
"Kang",
"Jiawen",
""
],
[
"Xiong",
"Zehui",
""
],
[
"Du",
"Hongyang",
""
],
[
"Mao",
"Shiwen",
""
]
]
| TITLE: Wireless Hallucination in Generative AI-enabled Communications:
Concepts, Issues, and Solutions
ABSTRACT: Generative AI (GenAI) is driving the intelligence of wireless communications.
Due to data limitations, random generation, and dynamic environments, GenAI may
generate channel information or optimization strategies that violate physical
laws or deviate from actual real-world requirements. We refer to this
phenomenon as wireless hallucination, which results in invalid channel
information, spectrum wastage, and low communication reliability but remains
underexplored. To address this gap, this article provides a comprehensive
concept of wireless hallucinations in GenAI-driven communications, focusing on
hallucination mitigation. Specifically, we first introduce the fundamental,
analyze its causes based on the GenAI workflow, and propose mitigation
solutions at the data, model, and post-generation levels. Then, we
systematically examines representative hallucination scenarios in GenAI-enabled
communications and their corresponding solutions. Finally, we propose a novel
integrated mitigation solution for GenAI-based channel estimation. At the data
level, we establish a channel estimation hallucination dataset and employ
generative adversarial networks (GANs)-based data augmentation. Additionally,
we incorporate attention mechanisms and large language models (LLMs) to enhance
both training and inference performance. Experimental results demonstrate that
the proposed hybrid solutions reduce the normalized mean square error (NMSE) by
0.19, effectively reducing wireless hallucinations.
| new_dataset | 0.957794 |
2503.06154 | Zidu Wang | Zidu Wang, Jiankuo Zhao, Miao Xu, Xiangyu Zhu, Zhen Lei | SRM-Hair: Single Image Head Mesh Reconstruction via 3D Morphable Hair | Under review | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | 3D Morphable Models (3DMMs) have played a pivotal role as a fundamental
representation or initialization for 3D avatar animation and reconstruction.
However, extending 3DMMs to hair remains challenging due to the difficulty of
enforcing vertex-level consistent semantic meaning across hair shapes. This
paper introduces a novel method, Semantic-consistent Ray Modeling of Hair
(SRM-Hair), for making 3D hair morphable and controlled by coefficients. The
key contribution lies in semantic-consistent ray modeling, which extracts
ordered hair surface vertices and exhibits notable properties such as
additivity for hairstyle fusion, adaptability, flipping, and thickness
modification. We collect a dataset of over 250 high-fidelity real hair scans
paired with 3D face data to serve as a prior for the 3D morphable hair. Based
on this, SRM-Hair can reconstruct a hair mesh combined with a 3D head from a
single image. Note that SRM-Hair produces an independent hair mesh,
facilitating applications in virtual avatar creation, realistic animation, and
high-fidelity hair rendering. Both quantitative and qualitative experiments
demonstrate that SRM-Hair achieves state-of-the-art performance in 3D mesh
reconstruction. Our project is available at
https://github.com/wang-zidu/SRM-Hair
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 10:37:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Zidu",
""
],
[
"Zhao",
"Jiankuo",
""
],
[
"Xu",
"Miao",
""
],
[
"Zhu",
"Xiangyu",
""
],
[
"Lei",
"Zhen",
""
]
]
| TITLE: SRM-Hair: Single Image Head Mesh Reconstruction via 3D Morphable Hair
ABSTRACT: 3D Morphable Models (3DMMs) have played a pivotal role as a fundamental
representation or initialization for 3D avatar animation and reconstruction.
However, extending 3DMMs to hair remains challenging due to the difficulty of
enforcing vertex-level consistent semantic meaning across hair shapes. This
paper introduces a novel method, Semantic-consistent Ray Modeling of Hair
(SRM-Hair), for making 3D hair morphable and controlled by coefficients. The
key contribution lies in semantic-consistent ray modeling, which extracts
ordered hair surface vertices and exhibits notable properties such as
additivity for hairstyle fusion, adaptability, flipping, and thickness
modification. We collect a dataset of over 250 high-fidelity real hair scans
paired with 3D face data to serve as a prior for the 3D morphable hair. Based
on this, SRM-Hair can reconstruct a hair mesh combined with a 3D head from a
single image. Note that SRM-Hair produces an independent hair mesh,
facilitating applications in virtual avatar creation, realistic animation, and
high-fidelity hair rendering. Both quantitative and qualitative experiments
demonstrate that SRM-Hair achieves state-of-the-art performance in 3D mesh
reconstruction. Our project is available at
https://github.com/wang-zidu/SRM-Hair
| new_dataset | 0.966156 |
2503.06161 | William Han | Kai Li, Junhao Wang, William Han, Ding Zhao | Feature-EndoGaussian: Feature Distilled Gaussian Splatting in Surgical
Deformable Scene Reconstruction | 14 pages, 5 figures | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Minimally invasive surgery (MIS) has transformed clinical practice by
reducing recovery times, minimizing complications, and enhancing precision.
Nonetheless, MIS inherently relies on indirect visualization and precise
instrument control, posing unique challenges. Recent advances in artificial
intelligence have enabled real-time surgical scene understanding through
techniques such as image classification, object detection, and segmentation,
with scene reconstruction emerging as a key element for enhanced intraoperative
guidance. Although neural radiance fields (NeRFs) have been explored for this
purpose, their substantial data requirements and slow rendering inhibit
real-time performance. In contrast, 3D Gaussian Splatting (3DGS) offers a more
efficient alternative, achieving state-of-the-art performance in dynamic
surgical scene reconstruction. In this work, we introduce Feature-EndoGaussian
(FEG), an extension of 3DGS that integrates 2D segmentation cues into 3D
rendering to enable real-time semantic and scene reconstruction. By leveraging
pretrained segmentation foundation models, FEG incorporates semantic feature
distillation within the Gaussian deformation framework, thereby enhancing both
reconstruction fidelity and segmentation accuracy. On the EndoNeRF dataset, FEG
achieves superior performance (SSIM of 0.97, PSNR of 39.08, and LPIPS of 0.03)
compared to leading methods. Additionally, on the EndoVis18 dataset, FEG
demonstrates competitive class-wise segmentation metrics while balancing model
size and real-time performance.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 10:50:19 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Kai",
""
],
[
"Wang",
"Junhao",
""
],
[
"Han",
"William",
""
],
[
"Zhao",
"Ding",
""
]
]
| TITLE: Feature-EndoGaussian: Feature Distilled Gaussian Splatting in Surgical
Deformable Scene Reconstruction
ABSTRACT: Minimally invasive surgery (MIS) has transformed clinical practice by
reducing recovery times, minimizing complications, and enhancing precision.
Nonetheless, MIS inherently relies on indirect visualization and precise
instrument control, posing unique challenges. Recent advances in artificial
intelligence have enabled real-time surgical scene understanding through
techniques such as image classification, object detection, and segmentation,
with scene reconstruction emerging as a key element for enhanced intraoperative
guidance. Although neural radiance fields (NeRFs) have been explored for this
purpose, their substantial data requirements and slow rendering inhibit
real-time performance. In contrast, 3D Gaussian Splatting (3DGS) offers a more
efficient alternative, achieving state-of-the-art performance in dynamic
surgical scene reconstruction. In this work, we introduce Feature-EndoGaussian
(FEG), an extension of 3DGS that integrates 2D segmentation cues into 3D
rendering to enable real-time semantic and scene reconstruction. By leveraging
pretrained segmentation foundation models, FEG incorporates semantic feature
distillation within the Gaussian deformation framework, thereby enhancing both
reconstruction fidelity and segmentation accuracy. On the EndoNeRF dataset, FEG
achieves superior performance (SSIM of 0.97, PSNR of 39.08, and LPIPS of 0.03)
compared to leading methods. Additionally, on the EndoVis18 dataset, FEG
demonstrates competitive class-wise segmentation metrics while balancing model
size and real-time performance.
| no_new_dataset | 0.944995 |
2503.06179 | Wongi Park | Wongi Park, Myeongseok Nam, Siwon Kim, Sangwoo Jo, Soomok Lee | ForestSplats: Deformable transient field for Gaussian Splatting in the
Wild | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recently, 3D Gaussian Splatting (3D-GS) has emerged, showing real-time
rendering speeds and high-quality results in static scenes. Although 3D-GS
shows effectiveness in static scenes, their performance significantly degrades
in real-world environments due to transient objects, lighting variations, and
diverse levels of occlusion. To tackle this, existing methods estimate
occluders or transient elements by leveraging pre-trained models or integrating
additional transient field pipelines. However, these methods still suffer from
two defects: 1) Using semantic features from the Vision Foundation model (VFM)
causes additional computational costs. 2) The transient field requires
significant memory to handle transient elements with per-view Gaussians and
struggles to define clear boundaries for occluders, solely relying on
photometric errors. To address these problems, we propose ForestSplats, a novel
approach that leverages the deformable transient field and a superpixel-aware
mask to efficiently represent transient elements in the 2D scene across
unconstrained image collections and effectively decompose static scenes from
transient distractors without VFM. We designed the transient field to be
deformable, capturing per-view transient elements. Furthermore, we introduce a
superpixel-aware mask that clearly defines the boundaries of occluders by
considering photometric errors and superpixels. Additionally, we propose
uncertainty-aware densification to avoid generating Gaussians within the
boundaries of occluders during densification. Through extensive experiments
across several benchmark datasets, we demonstrate that ForestSplats outperforms
existing methods without VFM and shows significant memory efficiency in
representing transient elements.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 11:44:57 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Park",
"Wongi",
""
],
[
"Nam",
"Myeongseok",
""
],
[
"Kim",
"Siwon",
""
],
[
"Jo",
"Sangwoo",
""
],
[
"Lee",
"Soomok",
""
]
]
| TITLE: ForestSplats: Deformable transient field for Gaussian Splatting in the
Wild
ABSTRACT: Recently, 3D Gaussian Splatting (3D-GS) has emerged, showing real-time
rendering speeds and high-quality results in static scenes. Although 3D-GS
shows effectiveness in static scenes, their performance significantly degrades
in real-world environments due to transient objects, lighting variations, and
diverse levels of occlusion. To tackle this, existing methods estimate
occluders or transient elements by leveraging pre-trained models or integrating
additional transient field pipelines. However, these methods still suffer from
two defects: 1) Using semantic features from the Vision Foundation model (VFM)
causes additional computational costs. 2) The transient field requires
significant memory to handle transient elements with per-view Gaussians and
struggles to define clear boundaries for occluders, solely relying on
photometric errors. To address these problems, we propose ForestSplats, a novel
approach that leverages the deformable transient field and a superpixel-aware
mask to efficiently represent transient elements in the 2D scene across
unconstrained image collections and effectively decompose static scenes from
transient distractors without VFM. We designed the transient field to be
deformable, capturing per-view transient elements. Furthermore, we introduce a
superpixel-aware mask that clearly defines the boundaries of occluders by
considering photometric errors and superpixels. Additionally, we propose
uncertainty-aware densification to avoid generating Gaussians within the
boundaries of occluders during densification. Through extensive experiments
across several benchmark datasets, we demonstrate that ForestSplats outperforms
existing methods without VFM and shows significant memory efficiency in
representing transient elements.
| no_new_dataset | 0.95018 |
2503.06182 | Antonio Alliegro | Antonio Alliegro, Francesca Pistilli, Tatiana Tommasi, Giuseppe Averta | FORESCENE: FOREcasting human activity via latent SCENE graphs diffusion | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Forecasting human-environment interactions in daily activities is challenging
due to the high variability of human behavior. While predicting directly from
videos is possible, it is limited by confounding factors like irrelevant
objects or background noise that do not contribute to the interaction. A
promising alternative is using Scene Graphs (SGs) to track only the relevant
elements. However, current methods for forecasting future SGs face significant
challenges and often rely on unrealistic assumptions, such as fixed objects
over time, limiting their applicability to long-term activities where
interacted objects may appear or disappear. In this paper, we introduce
FORESCENE, a novel framework for Scene Graph Anticipation (SGA) that predicts
both object and relationship evolution over time. FORESCENE encodes observed
video segments into a latent representation using a tailored Graph Auto-Encoder
and forecasts future SGs using a Latent Diffusion Model (LDM). Our approach
enables continuous prediction of interaction dynamics without making
assumptions on the graph's content or structure. We evaluate FORESCENE on the
Action Genome dataset, where it outperforms existing SGA methods while solving
a significantly more complex task.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 11:56:00 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Alliegro",
"Antonio",
""
],
[
"Pistilli",
"Francesca",
""
],
[
"Tommasi",
"Tatiana",
""
],
[
"Averta",
"Giuseppe",
""
]
]
| TITLE: FORESCENE: FOREcasting human activity via latent SCENE graphs diffusion
ABSTRACT: Forecasting human-environment interactions in daily activities is challenging
due to the high variability of human behavior. While predicting directly from
videos is possible, it is limited by confounding factors like irrelevant
objects or background noise that do not contribute to the interaction. A
promising alternative is using Scene Graphs (SGs) to track only the relevant
elements. However, current methods for forecasting future SGs face significant
challenges and often rely on unrealistic assumptions, such as fixed objects
over time, limiting their applicability to long-term activities where
interacted objects may appear or disappear. In this paper, we introduce
FORESCENE, a novel framework for Scene Graph Anticipation (SGA) that predicts
both object and relationship evolution over time. FORESCENE encodes observed
video segments into a latent representation using a tailored Graph Auto-Encoder
and forecasts future SGs using a Latent Diffusion Model (LDM). Our approach
enables continuous prediction of interaction dynamics without making
assumptions on the graph's content or structure. We evaluate FORESCENE on the
Action Genome dataset, where it outperforms existing SGA methods while solving
a significantly more complex task.
| no_new_dataset | 0.948537 |
2503.06196 | Shashata Sawmya | Shashata Sawmya, Thomas L. Athey, Gwyneth Liu, Nir Shavit | NeuroADDA: Active Discriminative Domain Adaptation in Connectomic | 8 pages, 3 figures, 3 tables | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Training segmentation models from scratch has been the standard approach for
new electron microscopy connectomics datasets. However, leveraging pretrained
models from existing datasets could improve efficiency and performance in
constrained annotation budget. In this study, we investigate domain adaptation
in connectomics by analyzing six major datasets spanning different organisms.
We show that, Maximum Mean Discrepancy (MMD) between neuron image distributions
serves as a reliable indicator of transferability, and identifies the optimal
source domain for transfer learning. Building on this, we introduce NeuroADDA,
a method that combines optimal domain selection with source-free active
learning to effectively adapt pretrained backbones to a new dataset. NeuroADDA
consistently outperforms training from scratch across diverse datasets and
fine-tuning sample sizes, with the largest gain observed at $n=4$ samples with
a 25-67\% reduction in Variation of Information. Finally, we show that our
analysis of distributional differences among neuron images from multiple
species in a learned feature space reveals that these domain "distances"
correlate with phylogenetic distance among those species.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 12:40:30 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Sawmya",
"Shashata",
""
],
[
"Athey",
"Thomas L.",
""
],
[
"Liu",
"Gwyneth",
""
],
[
"Shavit",
"Nir",
""
]
]
| TITLE: NeuroADDA: Active Discriminative Domain Adaptation in Connectomic
ABSTRACT: Training segmentation models from scratch has been the standard approach for
new electron microscopy connectomics datasets. However, leveraging pretrained
models from existing datasets could improve efficiency and performance in
constrained annotation budget. In this study, we investigate domain adaptation
in connectomics by analyzing six major datasets spanning different organisms.
We show that, Maximum Mean Discrepancy (MMD) between neuron image distributions
serves as a reliable indicator of transferability, and identifies the optimal
source domain for transfer learning. Building on this, we introduce NeuroADDA,
a method that combines optimal domain selection with source-free active
learning to effectively adapt pretrained backbones to a new dataset. NeuroADDA
consistently outperforms training from scratch across diverse datasets and
fine-tuning sample sizes, with the largest gain observed at $n=4$ samples with
a 25-67\% reduction in Variation of Information. Finally, we show that our
analysis of distributional differences among neuron images from multiple
species in a learned feature space reveals that these domain "distances"
correlate with phylogenetic distance among those species.
| no_new_dataset | 0.945751 |
2503.06200 | YeCong Wan | Yecong Wan, Mingwen Shao, Yuanshuo Cheng, Jun Shu, Shuigen Wang | Removing Multiple Hybrid Adverse Weather in Video via a Unified Model | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Videos captured under real-world adverse weather conditions typically suffer
from uncertain hybrid weather artifacts with heterogeneous degradation
distributions. However, existing algorithms only excel at specific single
degradation distributions due to limited adaption capacity and have to deal
with different weather degradations with separately trained models, thus may
fail to handle real-world stochastic weather scenarios. Besides, the model
training is also infeasible due to the lack of paired video data to
characterize the coexistence of multiple weather. To ameliorate the
aforementioned issue, we propose a novel unified model, dubbed UniWRV, to
remove multiple heterogeneous video weather degradations in an all-in-one
fashion. Specifically, to tackle degenerate spatial feature heterogeneity, we
propose a tailored weather prior guided module that queries exclusive priors
for different instances as prompts to steer spatial feature characterization.
To tackle degenerate temporal feature heterogeneity, we propose a dynamic
routing aggregation module that can automatically select optimal fusion paths
for different instances to dynamically integrate temporal features.
Additionally, we managed to construct a new synthetic video dataset, termed
HWVideo, for learning and benchmarking multiple hybrid adverse weather removal,
which contains 15 hybrid weather conditions with a total of 1500
adverse-weather/clean paired video clips. Real-world hybrid weather videos are
also collected for evaluating model generalizability. Comprehensive experiments
demonstrate that our UniWRV exhibits robust and superior adaptation capability
in multiple heterogeneous degradations learning scenarios, including various
generic video restoration tasks beyond weather removal.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:01:22 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wan",
"Yecong",
""
],
[
"Shao",
"Mingwen",
""
],
[
"Cheng",
"Yuanshuo",
""
],
[
"Shu",
"Jun",
""
],
[
"Wang",
"Shuigen",
""
]
]
| TITLE: Removing Multiple Hybrid Adverse Weather in Video via a Unified Model
ABSTRACT: Videos captured under real-world adverse weather conditions typically suffer
from uncertain hybrid weather artifacts with heterogeneous degradation
distributions. However, existing algorithms only excel at specific single
degradation distributions due to limited adaption capacity and have to deal
with different weather degradations with separately trained models, thus may
fail to handle real-world stochastic weather scenarios. Besides, the model
training is also infeasible due to the lack of paired video data to
characterize the coexistence of multiple weather. To ameliorate the
aforementioned issue, we propose a novel unified model, dubbed UniWRV, to
remove multiple heterogeneous video weather degradations in an all-in-one
fashion. Specifically, to tackle degenerate spatial feature heterogeneity, we
propose a tailored weather prior guided module that queries exclusive priors
for different instances as prompts to steer spatial feature characterization.
To tackle degenerate temporal feature heterogeneity, we propose a dynamic
routing aggregation module that can automatically select optimal fusion paths
for different instances to dynamically integrate temporal features.
Additionally, we managed to construct a new synthetic video dataset, termed
HWVideo, for learning and benchmarking multiple hybrid adverse weather removal,
which contains 15 hybrid weather conditions with a total of 1500
adverse-weather/clean paired video clips. Real-world hybrid weather videos are
also collected for evaluating model generalizability. Comprehensive experiments
demonstrate that our UniWRV exhibits robust and superior adaptation capability
in multiple heterogeneous degradations learning scenarios, including various
generic video restoration tasks beyond weather removal.
| new_dataset | 0.962391 |
2503.06201 | Yixin Wu | Yixin Wu, Feiran Zhang, Tianyuan Shi, Ruicheng Yin, Zhenghua Wang,
Zhenliang Gan, Xiaohua Wang, Changze Lv, Xiaoqing Zheng, Xuanjing Huang | Explainable Synthetic Image Detection through Diffusion Timestep
Ensembling | 13 pages, 5 figures | null | null | null | cs.CV cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in diffusion models have enabled the creation of deceptively
real images, posing significant security risks when misused. In this study, we
reveal that natural and synthetic images exhibit distinct differences in the
high-frequency domains of their Fourier power spectra after undergoing
iterative noise perturbations through an inverse multi-step denoising process,
suggesting that such noise can provide additional discriminative information
for identifying synthetic images. Based on this observation, we propose a novel
detection method that amplifies these differences by progressively adding noise
to the original images across multiple timesteps, and train an ensemble of
classifiers on these noised images. To enhance human comprehension, we
introduce an explanation generation and refinement module to identify flaws
located in AI-generated images. Additionally, we construct two new datasets,
GenHard and GenExplain, derived from the GenImage benchmark, providing
detection samples of greater difficulty and high-quality rationales for fake
images. Extensive experiments show that our method achieves state-of-the-art
performance with 98.91% and 95.89% detection accuracy on regular and harder
samples, increasing a minimal of 2.51% and 3.46% compared to baselines.
Furthermore, our method also generalizes effectively to images generated by
other diffusion models. Our code and datasets will be made publicly available.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:04:20 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wu",
"Yixin",
""
],
[
"Zhang",
"Feiran",
""
],
[
"Shi",
"Tianyuan",
""
],
[
"Yin",
"Ruicheng",
""
],
[
"Wang",
"Zhenghua",
""
],
[
"Gan",
"Zhenliang",
""
],
[
"Wang",
"Xiaohua",
""
],
[
"Lv",
"Changze",
""
],
[
"Zheng",
"Xiaoqing",
""
],
[
"Huang",
"Xuanjing",
""
]
]
| TITLE: Explainable Synthetic Image Detection through Diffusion Timestep
Ensembling
ABSTRACT: Recent advances in diffusion models have enabled the creation of deceptively
real images, posing significant security risks when misused. In this study, we
reveal that natural and synthetic images exhibit distinct differences in the
high-frequency domains of their Fourier power spectra after undergoing
iterative noise perturbations through an inverse multi-step denoising process,
suggesting that such noise can provide additional discriminative information
for identifying synthetic images. Based on this observation, we propose a novel
detection method that amplifies these differences by progressively adding noise
to the original images across multiple timesteps, and train an ensemble of
classifiers on these noised images. To enhance human comprehension, we
introduce an explanation generation and refinement module to identify flaws
located in AI-generated images. Additionally, we construct two new datasets,
GenHard and GenExplain, derived from the GenImage benchmark, providing
detection samples of greater difficulty and high-quality rationales for fake
images. Extensive experiments show that our method achieves state-of-the-art
performance with 98.91% and 95.89% detection accuracy on regular and harder
samples, increasing a minimal of 2.51% and 3.46% compared to baselines.
Furthermore, our method also generalizes effectively to images generated by
other diffusion models. Our code and datasets will be made publicly available.
| new_dataset | 0.935169 |
2503.06202 | Wei Liu | Wei Liu, Zhiying Deng, Zhongyu Niu, Jun Wang, Haozhao Wang, Zhigang
Zeng, Ruixuan Li | Breaking Free from MMI: A New Frontier in Rationalization by Probing
Input Utilization | null | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extracting a small subset of crucial rationales from the full input is a key
problem in explainability research. The most widely used fundamental criterion
for rationale extraction is the maximum mutual information (MMI) criterion. In
this paper, we first demonstrate that MMI suffers from diminishing marginal
returns. Once part of the rationale has been identified, finding the remaining
portions contributes only marginally to increasing the mutual information,
making it difficult to use MMI to locate the rest. In contrast to MMI that aims
to reproduce the prediction, we seek to identify the parts of the input that
the network can actually utilize.
This is achieved by comparing how different rationale candidates match the
capability space of the weight matrix. The weight matrix of a neural network is
typically low-rank, meaning that the linear combinations of its column vectors
can only cover part of the directions in a high-dimensional space
(high-dimension: the dimensions of an input vector). If an input is fully
utilized by the network, {it generally matches these directions (e.g., a
portion of a hypersphere), resulting in a representation with a high norm.
Conversely, if an input primarily falls outside (orthogonal to) these
directions}, its representation norm will approach zero, behaving like noise
that the network cannot effectively utilize. Building on this, we propose using
the norms of rationale candidates as an alternative objective to MMI. Through
experiments on four text classification datasets and one graph classification
dataset using three network architectures (GRUs, BERT, and GCN), we show that
our method outperforms MMI and its improved variants in identifying better
rationales. We also compare our method with a representative LLM
(llama-3.1-8b-instruct) and find that our simple method gets comparable results
to it and can sometimes even outperform it.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:08:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Liu",
"Wei",
""
],
[
"Deng",
"Zhiying",
""
],
[
"Niu",
"Zhongyu",
""
],
[
"Wang",
"Jun",
""
],
[
"Wang",
"Haozhao",
""
],
[
"Zeng",
"Zhigang",
""
],
[
"Li",
"Ruixuan",
""
]
]
| TITLE: Breaking Free from MMI: A New Frontier in Rationalization by Probing
Input Utilization
ABSTRACT: Extracting a small subset of crucial rationales from the full input is a key
problem in explainability research. The most widely used fundamental criterion
for rationale extraction is the maximum mutual information (MMI) criterion. In
this paper, we first demonstrate that MMI suffers from diminishing marginal
returns. Once part of the rationale has been identified, finding the remaining
portions contributes only marginally to increasing the mutual information,
making it difficult to use MMI to locate the rest. In contrast to MMI that aims
to reproduce the prediction, we seek to identify the parts of the input that
the network can actually utilize.
This is achieved by comparing how different rationale candidates match the
capability space of the weight matrix. The weight matrix of a neural network is
typically low-rank, meaning that the linear combinations of its column vectors
can only cover part of the directions in a high-dimensional space
(high-dimension: the dimensions of an input vector). If an input is fully
utilized by the network, {it generally matches these directions (e.g., a
portion of a hypersphere), resulting in a representation with a high norm.
Conversely, if an input primarily falls outside (orthogonal to) these
directions}, its representation norm will approach zero, behaving like noise
that the network cannot effectively utilize. Building on this, we propose using
the norms of rationale candidates as an alternative objective to MMI. Through
experiments on four text classification datasets and one graph classification
dataset using three network architectures (GRUs, BERT, and GCN), we show that
our method outperforms MMI and its improved variants in identifying better
rationales. We also compare our method with a representative LLM
(llama-3.1-8b-instruct) and find that our simple method gets comparable results
to it and can sometimes even outperform it.
| no_new_dataset | 0.948394 |
2503.06204 | Ofir Ben Shoham | Oriel Perets, Ofir Ben Shoham, Nir Grinberg, Nadav Rappoport | CUPCase: Clinically Uncommon Patient Cases and Diagnoses Dataset | Accepted to AAAI 2025 | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Medical benchmark datasets significantly contribute to developing Large
Language Models (LLMs) for medical knowledge extraction, diagnosis,
summarization, and other uses. Yet, current benchmarks are mainly derived from
exam questions given to medical students or cases described in the medical
literature, lacking the complexity of real-world patient cases that deviate
from classic textbook abstractions. These include rare diseases, uncommon
presentations of common diseases, and unexpected treatment responses. Here, we
construct Clinically Uncommon Patient Cases and Diagnosis Dataset (CUPCase)
based on 3,562 real-world case reports from BMC, including diagnoses in
open-ended textual format and as multiple-choice options with distractors.
Using this dataset, we evaluate the ability of state-of-the-art LLMs, including
both general-purpose and Clinical LLMs, to identify and correctly diagnose a
patient case, and test models' performance when only partial information about
cases is available. Our findings show that general-purpose GPT-4o attains the
best performance in both the multiple-choice task (average accuracy of 87.9%)
and the open-ended task (BERTScore F1 of 0.764), outperforming several LLMs
with a focus on the medical domain such as Meditron-70B and MedLM-Large.
Moreover, GPT-4o was able to maintain 87% and 88% of its performance with only
the first 20% of tokens of the case presentation in multiple-choice and free
text, respectively, highlighting the potential of LLMs to aid in early
diagnosis in real-world cases. CUPCase expands our ability to evaluate LLMs for
clinical decision support in an open and reproducible manner.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:21:44 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Perets",
"Oriel",
""
],
[
"Shoham",
"Ofir Ben",
""
],
[
"Grinberg",
"Nir",
""
],
[
"Rappoport",
"Nadav",
""
]
]
| TITLE: CUPCase: Clinically Uncommon Patient Cases and Diagnoses Dataset
ABSTRACT: Medical benchmark datasets significantly contribute to developing Large
Language Models (LLMs) for medical knowledge extraction, diagnosis,
summarization, and other uses. Yet, current benchmarks are mainly derived from
exam questions given to medical students or cases described in the medical
literature, lacking the complexity of real-world patient cases that deviate
from classic textbook abstractions. These include rare diseases, uncommon
presentations of common diseases, and unexpected treatment responses. Here, we
construct Clinically Uncommon Patient Cases and Diagnosis Dataset (CUPCase)
based on 3,562 real-world case reports from BMC, including diagnoses in
open-ended textual format and as multiple-choice options with distractors.
Using this dataset, we evaluate the ability of state-of-the-art LLMs, including
both general-purpose and Clinical LLMs, to identify and correctly diagnose a
patient case, and test models' performance when only partial information about
cases is available. Our findings show that general-purpose GPT-4o attains the
best performance in both the multiple-choice task (average accuracy of 87.9%)
and the open-ended task (BERTScore F1 of 0.764), outperforming several LLMs
with a focus on the medical domain such as Meditron-70B and MedLM-Large.
Moreover, GPT-4o was able to maintain 87% and 88% of its performance with only
the first 20% of tokens of the case presentation in multiple-choice and free
text, respectively, highlighting the potential of LLMs to aid in early
diagnosis in real-world cases. CUPCase expands our ability to evaluate LLMs for
clinical decision support in an open and reproducible manner.
| new_dataset | 0.872565 |
2503.06213 | Ruiyu Wang | Ruiyu Wang, Sen Wang, Xinxin Zuo, Qiang Sun | Lifelong Learning with Task-Specific Adaptation: Addressing the
Stability-Plasticity Dilemma | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lifelong learning (LL) aims to continuously acquire new knowledge while
retaining previously learned knowledge. A central challenge in LL is the
stability-plasticity dilemma, which requires models to balance the preservation
of previous knowledge (stability) with the ability to learn new tasks
(plasticity). While parameter-efficient fine-tuning (PEFT) has been widely
adopted in large language models, its application to lifelong learning remains
underexplored. To bridge this gap, this paper proposes AdaLL, an adapter-based
framework designed to address the dilemma through a simple, universal, and
effective strategy. AdaLL co-trains the backbone network and adapters under
regularization constraints, enabling the backbone to capture task-invariant
features while allowing the adapters to specialize in task-specific
information. Unlike methods that freeze the backbone network, AdaLL
incrementally enhances the backbone's capabilities across tasks while
minimizing interference through backbone regularization. This architectural
design significantly improves both stability and plasticity, effectively
eliminating the stability-plasticity dilemma. Extensive experiments demonstrate
that AdaLL consistently outperforms existing methods across various
configurations, including dataset choices, task sequences, and task scales.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:33:38 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Ruiyu",
""
],
[
"Wang",
"Sen",
""
],
[
"Zuo",
"Xinxin",
""
],
[
"Sun",
"Qiang",
""
]
]
| TITLE: Lifelong Learning with Task-Specific Adaptation: Addressing the
Stability-Plasticity Dilemma
ABSTRACT: Lifelong learning (LL) aims to continuously acquire new knowledge while
retaining previously learned knowledge. A central challenge in LL is the
stability-plasticity dilemma, which requires models to balance the preservation
of previous knowledge (stability) with the ability to learn new tasks
(plasticity). While parameter-efficient fine-tuning (PEFT) has been widely
adopted in large language models, its application to lifelong learning remains
underexplored. To bridge this gap, this paper proposes AdaLL, an adapter-based
framework designed to address the dilemma through a simple, universal, and
effective strategy. AdaLL co-trains the backbone network and adapters under
regularization constraints, enabling the backbone to capture task-invariant
features while allowing the adapters to specialize in task-specific
information. Unlike methods that freeze the backbone network, AdaLL
incrementally enhances the backbone's capabilities across tasks while
minimizing interference through backbone regularization. This architectural
design significantly improves both stability and plasticity, effectively
eliminating the stability-plasticity dilemma. Extensive experiments demonstrate
that AdaLL consistently outperforms existing methods across various
configurations, including dataset choices, task sequences, and task scales.
| no_new_dataset | 0.945651 |
2503.06222 | Meng Wang | Meng Wang, Fan Wu, Yunchuan Qin, Ruihui Li, Zhuo Tang, Kenli Li | Vision-based 3D Semantic Scene Completion via Capture Dynamic
Representations | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The vision-based semantic scene completion task aims to predict dense
geometric and semantic 3D scene representations from 2D images. However, the
presence of dynamic objects in the scene seriously affects the accuracy of the
model inferring 3D structures from 2D images. Existing methods simply stack
multiple frames of image input to increase dense scene semantic information,
but ignore the fact that dynamic objects and non-texture areas violate
multi-view consistency and matching reliability. To address these issues, we
propose a novel method, CDScene: Vision-based Robust Semantic Scene Completion
via Capturing Dynamic Representations. First, we leverage a multimodal
large-scale model to extract 2D explicit semantics and align them into 3D
space. Second, we exploit the characteristics of monocular and stereo depth to
decouple scene information into dynamic and static features. The dynamic
features contain structural relationships around dynamic objects, and the
static features contain dense contextual spatial information. Finally, we
design a dynamic-static adaptive fusion module to effectively extract and
aggregate complementary features, achieving robust and accurate semantic scene
completion in autonomous driving scenarios. Extensive experimental results on
the SemanticKITTI, SSCBench-KITTI360, and SemanticKITTI-C datasets demonstrate
the superiority and robustness of CDScene over existing state-of-the-art
methods.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 13:49:43 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Meng",
""
],
[
"Wu",
"Fan",
""
],
[
"Qin",
"Yunchuan",
""
],
[
"Li",
"Ruihui",
""
],
[
"Tang",
"Zhuo",
""
],
[
"Li",
"Kenli",
""
]
]
| TITLE: Vision-based 3D Semantic Scene Completion via Capture Dynamic
Representations
ABSTRACT: The vision-based semantic scene completion task aims to predict dense
geometric and semantic 3D scene representations from 2D images. However, the
presence of dynamic objects in the scene seriously affects the accuracy of the
model inferring 3D structures from 2D images. Existing methods simply stack
multiple frames of image input to increase dense scene semantic information,
but ignore the fact that dynamic objects and non-texture areas violate
multi-view consistency and matching reliability. To address these issues, we
propose a novel method, CDScene: Vision-based Robust Semantic Scene Completion
via Capturing Dynamic Representations. First, we leverage a multimodal
large-scale model to extract 2D explicit semantics and align them into 3D
space. Second, we exploit the characteristics of monocular and stereo depth to
decouple scene information into dynamic and static features. The dynamic
features contain structural relationships around dynamic objects, and the
static features contain dense contextual spatial information. Finally, we
design a dynamic-static adaptive fusion module to effectively extract and
aggregate complementary features, achieving robust and accurate semantic scene
completion in autonomous driving scenarios. Extensive experimental results on
the SemanticKITTI, SSCBench-KITTI360, and SemanticKITTI-C datasets demonstrate
the superiority and robustness of CDScene over existing state-of-the-art
methods.
| no_new_dataset | 0.941493 |
2503.06229 | Federico Mazzoni | Federico Mazzoni, Riccardo Guidotti, Alessio Malizia | A Frank System for Co-Evolutionary Hybrid Decision-Making | 13 pages | Advances in Intelligent Data Analysis XXII, Lecture Notes in
Computer Science, vol. 14642, Springer, pp. 236-248, 2024 | 10.1007/978-3-031-58553-1_19 | null | cs.HC cs.AI cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | We introduce Frank, a human-in-the-loop system for co-evolutionary hybrid
decision-making aiding the user to label records from an un-labeled dataset.
Frank employs incremental learning to ``evolve'' in parallel with the user's
decisions, by training an interpretable machine learning model on the records
labeled by the user. Furthermore, Frank advances state-of-the-art approaches by
offering inconsistency controls, explanations, fairness checks, and bad-faith
safeguards simultaneously. We evaluate our proposal by simulating the users'
behavior with various levels of expertise and reliance on Frank's suggestions.
The experiments show that Frank's intervention leads to improvements in the
accuracy and the fairness of the decisions.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 14:06:16 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Mazzoni",
"Federico",
""
],
[
"Guidotti",
"Riccardo",
""
],
[
"Malizia",
"Alessio",
""
]
]
| TITLE: A Frank System for Co-Evolutionary Hybrid Decision-Making
ABSTRACT: We introduce Frank, a human-in-the-loop system for co-evolutionary hybrid
decision-making aiding the user to label records from an un-labeled dataset.
Frank employs incremental learning to ``evolve'' in parallel with the user's
decisions, by training an interpretable machine learning model on the records
labeled by the user. Furthermore, Frank advances state-of-the-art approaches by
offering inconsistency controls, explanations, fairness checks, and bad-faith
safeguards simultaneously. We evaluate our proposal by simulating the users'
behavior with various levels of expertise and reliance on Frank's suggestions.
The experiments show that Frank's intervention leads to improvements in the
accuracy and the fairness of the decisions.
| no_new_dataset | 0.952442 |
2503.06231 | Aditya Shankar | Aditya Shankar, Lydia Y. Chen, Arie van Deursen, Rihan Hai | WaveStitch: Flexible and Fast Conditional Time Series Generation with
Diffusion Models | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Generating temporal data under constraints is critical for forecasting,
imputation, and synthesis. These datasets often include auxiliary conditions
that influence the values within the time series signal. Existing methods face
three key challenges: (1) they fail to adapt to conditions at inference time;
(2) they rely on sequential generation, which slows the generation speed; and
(3) they inefficiently encode categorical features, leading to increased
sparsity and input sizes. We propose WaveStitch, a novel method that addresses
these challenges by leveraging denoising diffusion probabilistic models to
efficiently generate accurate temporal data under given auxiliary constraints.
WaveStitch overcomes these limitations by: (1) modeling interactions between
constraints and signals to generalize to new, unseen conditions; (2) enabling
the parallel synthesis of sequential segments with a novel "stitching"
mechanism to enforce coherence across segments; and (3) encoding categorical
features as compact periodic signals while preserving temporal patterns.
Extensive evaluations across diverse datasets highlight WaveStitch's ability to
generalize to unseen conditions during inference, achieving up to a 10x lower
mean-squared-error compared to the state-of-the-art methods. Moreover,
WaveStitch generates data up to 460x faster than autoregressive methods while
maintaining comparable accuracy. By efficiently encoding categorical features,
WaveStitch provides a robust and efficient solution for temporal data
generation. Our code is open-sourced: https://github.com/adis98/HierarchicalTS
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 14:12:10 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Shankar",
"Aditya",
""
],
[
"Chen",
"Lydia Y.",
""
],
[
"van Deursen",
"Arie",
""
],
[
"Hai",
"Rihan",
""
]
]
| TITLE: WaveStitch: Flexible and Fast Conditional Time Series Generation with
Diffusion Models
ABSTRACT: Generating temporal data under constraints is critical for forecasting,
imputation, and synthesis. These datasets often include auxiliary conditions
that influence the values within the time series signal. Existing methods face
three key challenges: (1) they fail to adapt to conditions at inference time;
(2) they rely on sequential generation, which slows the generation speed; and
(3) they inefficiently encode categorical features, leading to increased
sparsity and input sizes. We propose WaveStitch, a novel method that addresses
these challenges by leveraging denoising diffusion probabilistic models to
efficiently generate accurate temporal data under given auxiliary constraints.
WaveStitch overcomes these limitations by: (1) modeling interactions between
constraints and signals to generalize to new, unseen conditions; (2) enabling
the parallel synthesis of sequential segments with a novel "stitching"
mechanism to enforce coherence across segments; and (3) encoding categorical
features as compact periodic signals while preserving temporal patterns.
Extensive evaluations across diverse datasets highlight WaveStitch's ability to
generalize to unseen conditions during inference, achieving up to a 10x lower
mean-squared-error compared to the state-of-the-art methods. Moreover,
WaveStitch generates data up to 460x faster than autoregressive methods while
maintaining comparable accuracy. By efficiently encoding categorical features,
WaveStitch provides a robust and efficient solution for temporal data
generation. Our code is open-sourced: https://github.com/adis98/HierarchicalTS
| no_new_dataset | 0.946695 |
2503.06247 | Minghao Fu | Minghao Fu, Danning Li, Aryan Gadhiya, Benjamin Lambright, Mohamed
Alowais, Mohab Bahnassy, Saad El Dine Elletter, Hawau Olamide Toyin, Haiyan
Jiang, Kun Zhang, Hanan Aldarmaki | Infant Cry Detection Using Causal Temporal Representation | Accepted to ICASSP 2025 | null | null | null | cs.SD cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses a major challenge in acoustic event detection, in
particular infant cry detection in the presence of other sounds and background
noises: the lack of precise annotated data. We present two contributions for
supervised and unsupervised infant cry detection. The first is an annotated
dataset for cry segmentation, which enables supervised models to achieve
state-of-the-art performance. Additionally, we propose a novel unsupervised
method, Causal Representation Spare Transition Clustering (CRSTC), based on
causal temporal representation, which helps address the issue of data scarcity
more generally. By integrating the detected cry segments, we significantly
improve the performance of downstream infant cry classification, highlighting
the potential of this approach for infant care applications.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 15:15:23 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Fu",
"Minghao",
""
],
[
"Li",
"Danning",
""
],
[
"Gadhiya",
"Aryan",
""
],
[
"Lambright",
"Benjamin",
""
],
[
"Alowais",
"Mohamed",
""
],
[
"Bahnassy",
"Mohab",
""
],
[
"Elletter",
"Saad El Dine",
""
],
[
"Toyin",
"Hawau Olamide",
""
],
[
"Jiang",
"Haiyan",
""
],
[
"Zhang",
"Kun",
""
],
[
"Aldarmaki",
"Hanan",
""
]
]
| TITLE: Infant Cry Detection Using Causal Temporal Representation
ABSTRACT: This paper addresses a major challenge in acoustic event detection, in
particular infant cry detection in the presence of other sounds and background
noises: the lack of precise annotated data. We present two contributions for
supervised and unsupervised infant cry detection. The first is an annotated
dataset for cry segmentation, which enables supervised models to achieve
state-of-the-art performance. Additionally, we propose a novel unsupervised
method, Causal Representation Spare Transition Clustering (CRSTC), based on
causal temporal representation, which helps address the issue of data scarcity
more generally. By integrating the detected cry segments, we significantly
improve the performance of downstream infant cry classification, highlighting
the potential of this approach for infant care applications.
| no_new_dataset | 0.95418 |
2503.06261 | Wei-En Tai | Wei-En Tai, Yu-Lin Shih, Cheng Sun, Yu-Chiang Frank Wang, Hwann-Tzong
Chen | Segment Anything, Even Occluded | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Amodal instance segmentation, which aims to detect and segment both visible
and invisible parts of objects in images, plays a crucial role in various
applications including autonomous driving, robotic manipulation, and scene
understanding. While existing methods require training both front-end detectors
and mask decoders jointly, this approach lacks flexibility and fails to
leverage the strengths of pre-existing modal detectors. To address this
limitation, we propose SAMEO, a novel framework that adapts the Segment
Anything Model (SAM) as a versatile mask decoder capable of interfacing with
various front-end detectors to enable mask prediction even for partially
occluded objects. Acknowledging the constraints of limited amodal segmentation
datasets, we introduce Amodal-LVIS, a large-scale synthetic dataset comprising
300K images derived from the modal LVIS and LVVIS datasets. This dataset
significantly expands the training data available for amodal segmentation
research. Our experimental results demonstrate that our approach, when trained
on the newly extended dataset, including Amodal-LVIS, achieves remarkable
zero-shot performance on both COCOA-cls and D2SA benchmarks, highlighting its
potential for generalization to unseen scenarios.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 16:14:57 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Tai",
"Wei-En",
""
],
[
"Shih",
"Yu-Lin",
""
],
[
"Sun",
"Cheng",
""
],
[
"Wang",
"Yu-Chiang Frank",
""
],
[
"Chen",
"Hwann-Tzong",
""
]
]
| TITLE: Segment Anything, Even Occluded
ABSTRACT: Amodal instance segmentation, which aims to detect and segment both visible
and invisible parts of objects in images, plays a crucial role in various
applications including autonomous driving, robotic manipulation, and scene
understanding. While existing methods require training both front-end detectors
and mask decoders jointly, this approach lacks flexibility and fails to
leverage the strengths of pre-existing modal detectors. To address this
limitation, we propose SAMEO, a novel framework that adapts the Segment
Anything Model (SAM) as a versatile mask decoder capable of interfacing with
various front-end detectors to enable mask prediction even for partially
occluded objects. Acknowledging the constraints of limited amodal segmentation
datasets, we introduce Amodal-LVIS, a large-scale synthetic dataset comprising
300K images derived from the modal LVIS and LVVIS datasets. This dataset
significantly expands the training data available for amodal segmentation
research. Our experimental results demonstrate that our approach, when trained
on the newly extended dataset, including Amodal-LVIS, achieves remarkable
zero-shot performance on both COCOA-cls and D2SA benchmarks, highlighting its
potential for generalization to unseen scenarios.
| new_dataset | 0.961678 |
2503.06268 | Shaobin Zhuang | Shaobin Zhuang, Zhipeng Huang, Binxin Yang, Ying Zhang, Fangyikang
Wang, Canmiao Fu, Chong Sun, Zheng-Jun Zha, Chen Li, Yali Wang | Get In Video: Add Anything You Want to the Video | Project page:https://zhuangshaobin.github.io/GetInVideo-project/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video editing increasingly demands the ability to incorporate specific
real-world instances into existing footage, yet current approaches
fundamentally fail to capture the unique visual characteristics of particular
subjects and ensure natural instance/scene interactions. We formalize this
overlooked yet critical editing paradigm as "Get-In-Video Editing", where users
provide reference images to precisely specify visual elements they wish to
incorporate into videos. Addressing this task's dual challenges, severe
training data scarcity and technical challenges in maintaining spatiotemporal
coherence, we introduce three key contributions. First, we develop GetIn-1M
dataset created through our automated Recognize-Track-Erase pipeline, which
sequentially performs video captioning, salient instance identification, object
detection, temporal tracking, and instance removal to generate high-quality
video editing pairs with comprehensive annotations (reference image, tracking
mask, instance prompt). Second, we present GetInVideo, a novel end-to-end
framework that leverages a diffusion transformer architecture with 3D full
attention to process reference images, condition videos, and masks
simultaneously, maintaining temporal coherence, preserving visual identity, and
ensuring natural scene interactions when integrating reference objects into
videos. Finally, we establish GetInBench, the first comprehensive benchmark for
Get-In-Video Editing scenario, demonstrating our approach's superior
performance through extensive evaluations. Our work enables accessible,
high-quality incorporation of specific real-world subjects into videos,
significantly advancing personalized video editing capabilities.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 16:27:53 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhuang",
"Shaobin",
""
],
[
"Huang",
"Zhipeng",
""
],
[
"Yang",
"Binxin",
""
],
[
"Zhang",
"Ying",
""
],
[
"Wang",
"Fangyikang",
""
],
[
"Fu",
"Canmiao",
""
],
[
"Sun",
"Chong",
""
],
[
"Zha",
"Zheng-Jun",
""
],
[
"Li",
"Chen",
""
],
[
"Wang",
"Yali",
""
]
]
| TITLE: Get In Video: Add Anything You Want to the Video
ABSTRACT: Video editing increasingly demands the ability to incorporate specific
real-world instances into existing footage, yet current approaches
fundamentally fail to capture the unique visual characteristics of particular
subjects and ensure natural instance/scene interactions. We formalize this
overlooked yet critical editing paradigm as "Get-In-Video Editing", where users
provide reference images to precisely specify visual elements they wish to
incorporate into videos. Addressing this task's dual challenges, severe
training data scarcity and technical challenges in maintaining spatiotemporal
coherence, we introduce three key contributions. First, we develop GetIn-1M
dataset created through our automated Recognize-Track-Erase pipeline, which
sequentially performs video captioning, salient instance identification, object
detection, temporal tracking, and instance removal to generate high-quality
video editing pairs with comprehensive annotations (reference image, tracking
mask, instance prompt). Second, we present GetInVideo, a novel end-to-end
framework that leverages a diffusion transformer architecture with 3D full
attention to process reference images, condition videos, and masks
simultaneously, maintaining temporal coherence, preserving visual identity, and
ensuring natural scene interactions when integrating reference objects into
videos. Finally, we establish GetInBench, the first comprehensive benchmark for
Get-In-Video Editing scenario, demonstrating our approach's superior
performance through extensive evaluations. Our work enables accessible,
high-quality incorporation of specific real-world subjects into videos,
significantly advancing personalized video editing capabilities.
| new_dataset | 0.966092 |
2503.06269 | Thomas Winninger | Thomas Winninger, Boussad Addad, Katarzyna Kapusta | Using Mechanistic Interpretability to Craft Adversarial Attacks against
Large Language Models | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Traditional white-box methods for creating adversarial perturbations against
LLMs typically rely only on gradient computation from the targeted model,
ignoring the internal mechanisms responsible for attack success or failure.
Conversely, interpretability studies that analyze these internal mechanisms
lack practical applications beyond runtime interventions. We bridge this gap by
introducing a novel white-box approach that leverages mechanistic
interpretability techniques to craft practical adversarial inputs.
Specifically, we first identify acceptance subspaces - sets of feature vectors
that do not trigger the model's refusal mechanisms - then use gradient-based
optimization to reroute embeddings from refusal subspaces to acceptance
subspaces, effectively achieving jailbreaks. This targeted approach
significantly reduces computation cost, achieving attack success rates of
80-95\% on state-of-the-art models including Gemma2, Llama3.2, and Qwen2.5
within minutes or even seconds, compared to existing techniques that often fail
or require hours of computation. We believe this approach opens a new direction
for both attack research and defense development. Furthermore, it showcases a
practical application of mechanistic interpretability where other methods are
less efficient, which highlights its utility. The code and generated datasets
are available at https://github.com/Sckathach/subspace-rerouting.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 16:29:45 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Winninger",
"Thomas",
""
],
[
"Addad",
"Boussad",
""
],
[
"Kapusta",
"Katarzyna",
""
]
]
| TITLE: Using Mechanistic Interpretability to Craft Adversarial Attacks against
Large Language Models
ABSTRACT: Traditional white-box methods for creating adversarial perturbations against
LLMs typically rely only on gradient computation from the targeted model,
ignoring the internal mechanisms responsible for attack success or failure.
Conversely, interpretability studies that analyze these internal mechanisms
lack practical applications beyond runtime interventions. We bridge this gap by
introducing a novel white-box approach that leverages mechanistic
interpretability techniques to craft practical adversarial inputs.
Specifically, we first identify acceptance subspaces - sets of feature vectors
that do not trigger the model's refusal mechanisms - then use gradient-based
optimization to reroute embeddings from refusal subspaces to acceptance
subspaces, effectively achieving jailbreaks. This targeted approach
significantly reduces computation cost, achieving attack success rates of
80-95\% on state-of-the-art models including Gemma2, Llama3.2, and Qwen2.5
within minutes or even seconds, compared to existing techniques that often fail
or require hours of computation. We believe this approach opens a new direction
for both attack research and defense development. Furthermore, it showcases a
practical application of mechanistic interpretability where other methods are
less efficient, which highlights its utility. The code and generated datasets
are available at https://github.com/Sckathach/subspace-rerouting.
| no_new_dataset | 0.943712 |
2503.06272 | Gubio Gome De Lima | Gubio Lima and Gustavo Miranda and Tiago de Souza Farias | Introdu\c{c}\~ao a rede neural para F\'isicos | 26 pages, 19 figures | null | null | null | physics.class-ph physics.comp-ph | http://creativecommons.org/publicdomain/zero/1.0/ | As t\'ecnicas de aprendizado de m\'aquina emergiram como ferramentas
poderosas para enfrentar uma ampla gama de desafios. A integra\c{c}\~ao dessas
t\'ecnicas com a f\'isica tem conduzido a abordagens inovadoras na
compreens\~ao, controle e simula\c{c}\~ao de fen\^omenos f\'isicos. Este artigo
visa proporcionar uma introdu\c{c}\~ao pr\'atica \`as redes neurais e seus
conceitos fundamentais, destacando perspectivas recentes dos avan\c{c}os na
interse\c{c}\~ao entre modelos de aprendizado de m\'aquina e sistemas
f\'isicos. Al\'em disso, apresentamos um material pr\'atico para orientar o
leitor em seus primeiros passos na aplica\c{c}\~ao de redes neurais para
resolver problemas f\'isicos. Como exemplo ilustrativo, fornecemos quatro
aplica\c{c}\~oes de complexidades crescentes para o problema de um p\^endulo
simples, a saber: fit de par\^ametros de EDO do p\^endulo para
aproxima\c{c}\~ao de \^angulo pequeno; Physics Inspired Neural Networks (PINNs)
para encontrar solu\c{c}\~oes da EDO do p\^endulo em angulo pequeno;
Autoencoders em dataset de imagens do p\^endulo para estima\c{c}\~ao de
dimensionalidade de espa\c{c}o de par\^ametros do problema f\'isico; Uso de
arquiteturas Sparse Identification of Non-Linear Dynamics (SINDy) para
descoberta de modelos e express\~oes anal\'iticas para o problema do p\^endulo
n\~ao linear (angulos grandes) .
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 16:33:52 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Lima",
"Gubio",
""
],
[
"Miranda",
"Gustavo",
""
],
[
"Farias",
"Tiago de Souza",
""
]
]
| TITLE: Introdu\c{c}\~ao a rede neural para F\'isicos
ABSTRACT: As t\'ecnicas de aprendizado de m\'aquina emergiram como ferramentas
poderosas para enfrentar uma ampla gama de desafios. A integra\c{c}\~ao dessas
t\'ecnicas com a f\'isica tem conduzido a abordagens inovadoras na
compreens\~ao, controle e simula\c{c}\~ao de fen\^omenos f\'isicos. Este artigo
visa proporcionar uma introdu\c{c}\~ao pr\'atica \`as redes neurais e seus
conceitos fundamentais, destacando perspectivas recentes dos avan\c{c}os na
interse\c{c}\~ao entre modelos de aprendizado de m\'aquina e sistemas
f\'isicos. Al\'em disso, apresentamos um material pr\'atico para orientar o
leitor em seus primeiros passos na aplica\c{c}\~ao de redes neurais para
resolver problemas f\'isicos. Como exemplo ilustrativo, fornecemos quatro
aplica\c{c}\~oes de complexidades crescentes para o problema de um p\^endulo
simples, a saber: fit de par\^ametros de EDO do p\^endulo para
aproxima\c{c}\~ao de \^angulo pequeno; Physics Inspired Neural Networks (PINNs)
para encontrar solu\c{c}\~oes da EDO do p\^endulo em angulo pequeno;
Autoencoders em dataset de imagens do p\^endulo para estima\c{c}\~ao de
dimensionalidade de espa\c{c}o de par\^ametros do problema f\'isico; Uso de
arquiteturas Sparse Identification of Non-Linear Dynamics (SINDy) para
descoberta de modelos e express\~oes anal\'iticas para o problema do p\^endulo
n\~ao linear (angulos grandes) .
| no_new_dataset | 0.947672 |
2503.06276 | Songping Wang | Songping Wang, Xinquan Yue, Yueming Lyu, Caifeng Shan | Exploring Adversarial Transferability between Kolmogorov-arnold Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Kolmogorov-Arnold Networks (KANs) have emerged as a transformative model
paradigm, significantly impacting various fields. However, their adversarial
robustness remains less underexplored, especially across different KAN
architectures. To explore this critical safety issue, we conduct an analysis
and find that due to overfitting to the specific basis functions of KANs, they
possess poor adversarial transferability among different KANs. To tackle this
challenge, we propose AdvKAN, the first transfer attack method for KANs. AdvKAN
integrates two key components: 1) a Breakthrough-Defense Surrogate Model
(BDSM), which employs a breakthrough-defense training strategy to mitigate
overfitting to the specific structures of KANs. 2) a Global-Local Interaction
(GLI) technique, which promotes sufficient interaction between adversarial
gradients of hierarchical levels, further smoothing out loss surfaces of KANs.
Both of them work together to enhance the strength of transfer attack among
different KANs. Extensive experimental results on various KANs and datasets
demonstrate the effectiveness of AdvKAN, which possesses notably superior
attack capabilities and deeply reveals the vulnerabilities of KANs. Code will
be released upon acceptance.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 16:48:05 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Songping",
""
],
[
"Yue",
"Xinquan",
""
],
[
"Lyu",
"Yueming",
""
],
[
"Shan",
"Caifeng",
""
]
]
| TITLE: Exploring Adversarial Transferability between Kolmogorov-arnold Networks
ABSTRACT: Kolmogorov-Arnold Networks (KANs) have emerged as a transformative model
paradigm, significantly impacting various fields. However, their adversarial
robustness remains less underexplored, especially across different KAN
architectures. To explore this critical safety issue, we conduct an analysis
and find that due to overfitting to the specific basis functions of KANs, they
possess poor adversarial transferability among different KANs. To tackle this
challenge, we propose AdvKAN, the first transfer attack method for KANs. AdvKAN
integrates two key components: 1) a Breakthrough-Defense Surrogate Model
(BDSM), which employs a breakthrough-defense training strategy to mitigate
overfitting to the specific structures of KANs. 2) a Global-Local Interaction
(GLI) technique, which promotes sufficient interaction between adversarial
gradients of hierarchical levels, further smoothing out loss surfaces of KANs.
Both of them work together to enhance the strength of transfer attack among
different KANs. Extensive experimental results on various KANs and datasets
demonstrate the effectiveness of AdvKAN, which possesses notably superior
attack capabilities and deeply reveals the vulnerabilities of KANs. Code will
be released upon acceptance.
| no_new_dataset | 0.945096 |
2503.06282 | Shuangzhi Li | Shuangzhi Li, Junlong Shen, Lei Ma, and Xingyu Li | From Dataset to Real-world: General 3D Object Detection via Generalized
Cross-domain Few-shot Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | LiDAR-based 3D object detection datasets have been pivotal for autonomous
driving, yet they cover a limited range of objects, restricting the model's
generalization across diverse deployment environments. To address this, we
introduce the first generalized cross-domain few-shot (GCFS) task in 3D object
detection, which focuses on adapting a source-pretrained model for high
performance on both common and novel classes in a target domain with few-shot
samples. Our solution integrates multi-modal fusion and contrastive-enhanced
prototype learning within one framework, holistically overcoming challenges
related to data scarcity and domain adaptation in the GCFS setting. The
multi-modal fusion module utilizes 2D vision-language models to extract rich,
open-set semantic knowledge. To address biases in point distributions across
varying structural complexities, we particularly introduce a physically-aware
box searching strategy that leverages laser imaging principles to generate
high-quality 3D box proposals from 2D insights, enhancing object recall. To
effectively capture domain-specific representations for each class from limited
target data, we further propose a contrastive-enhanced prototype learning,
which strengthens the model's adaptability. We evaluate our approach with three
GCFS benchmark settings, and extensive experiments demonstrate the
effectiveness of our solution for GCFS tasks. The code will be publicly
available.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 17:05:21 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Shuangzhi",
""
],
[
"Shen",
"Junlong",
""
],
[
"Ma",
"Lei",
""
],
[
"Li",
"Xingyu",
""
]
]
| TITLE: From Dataset to Real-world: General 3D Object Detection via Generalized
Cross-domain Few-shot Learning
ABSTRACT: LiDAR-based 3D object detection datasets have been pivotal for autonomous
driving, yet they cover a limited range of objects, restricting the model's
generalization across diverse deployment environments. To address this, we
introduce the first generalized cross-domain few-shot (GCFS) task in 3D object
detection, which focuses on adapting a source-pretrained model for high
performance on both common and novel classes in a target domain with few-shot
samples. Our solution integrates multi-modal fusion and contrastive-enhanced
prototype learning within one framework, holistically overcoming challenges
related to data scarcity and domain adaptation in the GCFS setting. The
multi-modal fusion module utilizes 2D vision-language models to extract rich,
open-set semantic knowledge. To address biases in point distributions across
varying structural complexities, we particularly introduce a physically-aware
box searching strategy that leverages laser imaging principles to generate
high-quality 3D box proposals from 2D insights, enhancing object recall. To
effectively capture domain-specific representations for each class from limited
target data, we further propose a contrastive-enhanced prototype learning,
which strengthens the model's adaptability. We evaluate our approach with three
GCFS benchmark settings, and extensive experiments demonstrate the
effectiveness of our solution for GCFS tasks. The code will be publicly
available.
| no_new_dataset | 0.944536 |
2503.06296 | Vinay Verma Kumar | Vinay Kumar Verma, Shreyas Sunil Kulkarni, Happy Mittal and Deepak
Gupta | MoEMoE: Question Guided Dense and Scalable Sparse Mixture-of-Expert for
Multi-source Multi-modal Answering | To appear at NAACL Industry Track | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Question Answering (QA) and Visual Question Answering (VQA) are well-studied
problems in the language and vision domain. One challenging scenario involves
multiple sources of information, each of a different modality, where the answer
to the question may exist in one or more sources. This scenario contains richer
information but is highly complex to handle. In this work, we formulate a novel
question-answer generation (QAG) framework in an environment containing
multi-source, multimodal information. The answer may belong to any or all
sources; therefore, selecting the most prominent answer source or an optimal
combination of all sources for a given question is challenging. To address this
issue, we propose a question-guided attention mechanism that learns attention
across multiple sources and decodes this information for robust and unbiased
answer generation. To learn attention within each source, we introduce an
explicit alignment between questions and various information sources, which
facilitates identifying the most pertinent parts of the source information
relative to the question. Scalability in handling diverse questions poses a
challenge. We address this by extending our model to a sparse
mixture-of-experts (sparse-MoE) framework, enabling it to handle thousands of
question types. Experiments on T5 and Flan-T5 using three datasets demonstrate
the model's efficacy, supported by ablation studies.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 18:09:13 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Verma",
"Vinay Kumar",
""
],
[
"Kulkarni",
"Shreyas Sunil",
""
],
[
"Mittal",
"Happy",
""
],
[
"Gupta",
"Deepak",
""
]
]
| TITLE: MoEMoE: Question Guided Dense and Scalable Sparse Mixture-of-Expert for
Multi-source Multi-modal Answering
ABSTRACT: Question Answering (QA) and Visual Question Answering (VQA) are well-studied
problems in the language and vision domain. One challenging scenario involves
multiple sources of information, each of a different modality, where the answer
to the question may exist in one or more sources. This scenario contains richer
information but is highly complex to handle. In this work, we formulate a novel
question-answer generation (QAG) framework in an environment containing
multi-source, multimodal information. The answer may belong to any or all
sources; therefore, selecting the most prominent answer source or an optimal
combination of all sources for a given question is challenging. To address this
issue, we propose a question-guided attention mechanism that learns attention
across multiple sources and decodes this information for robust and unbiased
answer generation. To learn attention within each source, we introduce an
explicit alignment between questions and various information sources, which
facilitates identifying the most pertinent parts of the source information
relative to the question. Scalability in handling diverse questions poses a
challenge. We address this by extending our model to a sparse
mixture-of-experts (sparse-MoE) framework, enabling it to handle thousands of
question types. Experiments on T5 and Flan-T5 using three datasets demonstrate
the model's efficacy, supported by ablation studies.
| no_new_dataset | 0.941708 |
2503.06312 | Zhitong Xiong | Zhitong Xiong, Yi Wang, Weikang Yu, Adam J Stewart, Jie Zhao, Nils
Lehmann, Thomas Dujardin, Zhenghang Yuan, Pedram Ghamisi, Xiao Xiang Zhu | GeoLangBind: Unifying Earth Observation with Agglomerative
Vision-Language Foundation Models | code & weights: https://github.com/xiong-zhitong/GeoLB-SigLIP | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Earth observation (EO) data, collected from diverse sensors with varying
imaging principles, present significant challenges in creating unified
analytical frameworks. We present GeoLangBind, a novel agglomerative
vision--language foundation model that bridges the gap between heterogeneous EO
data modalities using language as a unifying medium. Our approach aligns
different EO data types into a shared language embedding space, enabling
seamless integration and complementary feature learning from diverse sensor
data. To achieve this, we construct a large-scale multimodal image--text
dataset, GeoLangBind-2M, encompassing six data modalities. GeoLangBind
leverages this dataset to develop a zero-shot foundation model capable of
processing arbitrary numbers of EO data channels as input. Through our designed
Modality-aware Knowledge Agglomeration (MaKA) module and progressive multimodal
weight merging strategy, we create a powerful agglomerative foundation model
that excels in both zero-shot vision--language comprehension and fine-grained
visual understanding. Extensive evaluation across 23 datasets covering multiple
tasks demonstrates GeoLangBind's superior performance and versatility in EO
applications, offering a robust framework for various environmental monitoring
and analysis tasks. The dataset and pretrained models will be publicly
available.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 19:10:04 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Xiong",
"Zhitong",
""
],
[
"Wang",
"Yi",
""
],
[
"Yu",
"Weikang",
""
],
[
"Stewart",
"Adam J",
""
],
[
"Zhao",
"Jie",
""
],
[
"Lehmann",
"Nils",
""
],
[
"Dujardin",
"Thomas",
""
],
[
"Yuan",
"Zhenghang",
""
],
[
"Ghamisi",
"Pedram",
""
],
[
"Zhu",
"Xiao Xiang",
""
]
]
| TITLE: GeoLangBind: Unifying Earth Observation with Agglomerative
Vision-Language Foundation Models
ABSTRACT: Earth observation (EO) data, collected from diverse sensors with varying
imaging principles, present significant challenges in creating unified
analytical frameworks. We present GeoLangBind, a novel agglomerative
vision--language foundation model that bridges the gap between heterogeneous EO
data modalities using language as a unifying medium. Our approach aligns
different EO data types into a shared language embedding space, enabling
seamless integration and complementary feature learning from diverse sensor
data. To achieve this, we construct a large-scale multimodal image--text
dataset, GeoLangBind-2M, encompassing six data modalities. GeoLangBind
leverages this dataset to develop a zero-shot foundation model capable of
processing arbitrary numbers of EO data channels as input. Through our designed
Modality-aware Knowledge Agglomeration (MaKA) module and progressive multimodal
weight merging strategy, we create a powerful agglomerative foundation model
that excels in both zero-shot vision--language comprehension and fine-grained
visual understanding. Extensive evaluation across 23 datasets covering multiple
tasks demonstrates GeoLangBind's superior performance and versatility in EO
applications, offering a robust framework for various environmental monitoring
and analysis tasks. The dataset and pretrained models will be publicly
available.
| new_dataset | 0.958693 |
2503.06313 | Chandan Kumar Sah | Chandan Kumar Sah, Ankit Kumar Shaw, Xiaoli Lian, Arsalan Shahid Baig,
Tuopu Wen, Kun Jiang, Mengmeng Yang, Diange Yang | Advancing Autonomous Vehicle Intelligence: Deep Learning and Multimodal
LLM for Traffic Sign Recognition and Robust Lane Detection | 11 pages, 9 figures | null | null | null | cs.CV cs.AI cs.CL cs.LG cs.RO | http://creativecommons.org/licenses/by/4.0/ | Autonomous vehicles (AVs) require reliable traffic sign recognition and
robust lane detection capabilities to ensure safe navigation in complex and
dynamic environments. This paper introduces an integrated approach combining
advanced deep learning techniques and Multimodal Large Language Models (MLLMs)
for comprehensive road perception. For traffic sign recognition, we
systematically evaluate ResNet-50, YOLOv8, and RT-DETR, achieving
state-of-the-art performance of 99.8% with ResNet-50, 98.0% accuracy with
YOLOv8, and achieved 96.6% accuracy in RT-DETR despite its higher computational
complexity. For lane detection, we propose a CNN-based segmentation method
enhanced by polynomial curve fitting, which delivers high accuracy under
favorable conditions. Furthermore, we introduce a lightweight, Multimodal,
LLM-based framework that directly undergoes instruction tuning using small yet
diverse datasets, eliminating the need for initial pretraining. This framework
effectively handles various lane types, complex intersections, and merging
zones, significantly enhancing lane detection reliability by reasoning under
adverse conditions. Despite constraints in available training resources, our
multimodal approach demonstrates advanced reasoning capabilities, achieving a
Frame Overall Accuracy (FRM) of 53.87%, a Question Overall Accuracy (QNS) of
82.83%, lane detection accuracies of 99.6% in clear conditions and 93.0% at
night, and robust performance in reasoning about lane invisibility due to rain
(88.4%) or road degradation (95.6%). The proposed comprehensive framework
markedly enhances AV perception reliability, thus contributing significantly to
safer autonomous driving across diverse and challenging road scenarios.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 19:12:36 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Sah",
"Chandan Kumar",
""
],
[
"Shaw",
"Ankit Kumar",
""
],
[
"Lian",
"Xiaoli",
""
],
[
"Baig",
"Arsalan Shahid",
""
],
[
"Wen",
"Tuopu",
""
],
[
"Jiang",
"Kun",
""
],
[
"Yang",
"Mengmeng",
""
],
[
"Yang",
"Diange",
""
]
]
| TITLE: Advancing Autonomous Vehicle Intelligence: Deep Learning and Multimodal
LLM for Traffic Sign Recognition and Robust Lane Detection
ABSTRACT: Autonomous vehicles (AVs) require reliable traffic sign recognition and
robust lane detection capabilities to ensure safe navigation in complex and
dynamic environments. This paper introduces an integrated approach combining
advanced deep learning techniques and Multimodal Large Language Models (MLLMs)
for comprehensive road perception. For traffic sign recognition, we
systematically evaluate ResNet-50, YOLOv8, and RT-DETR, achieving
state-of-the-art performance of 99.8% with ResNet-50, 98.0% accuracy with
YOLOv8, and achieved 96.6% accuracy in RT-DETR despite its higher computational
complexity. For lane detection, we propose a CNN-based segmentation method
enhanced by polynomial curve fitting, which delivers high accuracy under
favorable conditions. Furthermore, we introduce a lightweight, Multimodal,
LLM-based framework that directly undergoes instruction tuning using small yet
diverse datasets, eliminating the need for initial pretraining. This framework
effectively handles various lane types, complex intersections, and merging
zones, significantly enhancing lane detection reliability by reasoning under
adverse conditions. Despite constraints in available training resources, our
multimodal approach demonstrates advanced reasoning capabilities, achieving a
Frame Overall Accuracy (FRM) of 53.87%, a Question Overall Accuracy (QNS) of
82.83%, lane detection accuracies of 99.6% in clear conditions and 93.0% at
night, and robust performance in reasoning about lane invisibility due to rain
(88.4%) or road degradation (95.6%). The proposed comprehensive framework
markedly enhances AV perception reliability, thus contributing significantly to
safer autonomous driving across diverse and challenging road scenarios.
| no_new_dataset | 0.951188 |
2503.06317 | Badhan Chandra Das | Badhan Chandra Das, M. Hadi Amini, Yanzhao Wu | Accurate and Efficient Two-Stage Gun Detection in Video | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Object detection in videos plays a crucial role in advancing applications
such as public safety and anomaly detection. Existing methods have explored
different techniques, including CNN, deep learning, and Transformers, for
object detection and video classification. However, detecting tiny objects,
e.g., guns, in videos remains challenging due to their small scale and varying
appearances in complex scenes. Moreover, existing video analysis models for
classification or detection often perform poorly in real-world gun detection
scenarios due to limited labeled video datasets for training. Thus, developing
efficient methods for effectively capturing tiny object features and designing
models capable of accurate gun detection in real-world videos is imperative. To
address these challenges, we make three original contributions in this paper.
First, we conduct an empirical study of several existing video classification
and object detection methods to identify guns in videos. Our extensive analysis
shows that these methods may not accurately detect guns in videos. Second, we
propose a novel two-stage gun detection method. In stage 1, we train an
image-augmented model to effectively classify ``Gun'' videos. To make the
detection more precise and efficient, stage 2 employs an object detection model
to locate the exact region of the gun within video frames for videos classified
as ``Gun'' by stage 1. Third, our experimental results demonstrate that the
proposed domain-specific method achieves significant performance improvements
and enhances efficiency compared with existing techniques. We also discuss
challenges and future research directions in gun detection tasks in computer
vision.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 19:26:23 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Das",
"Badhan Chandra",
""
],
[
"Amini",
"M. Hadi",
""
],
[
"Wu",
"Yanzhao",
""
]
]
| TITLE: Accurate and Efficient Two-Stage Gun Detection in Video
ABSTRACT: Object detection in videos plays a crucial role in advancing applications
such as public safety and anomaly detection. Existing methods have explored
different techniques, including CNN, deep learning, and Transformers, for
object detection and video classification. However, detecting tiny objects,
e.g., guns, in videos remains challenging due to their small scale and varying
appearances in complex scenes. Moreover, existing video analysis models for
classification or detection often perform poorly in real-world gun detection
scenarios due to limited labeled video datasets for training. Thus, developing
efficient methods for effectively capturing tiny object features and designing
models capable of accurate gun detection in real-world videos is imperative. To
address these challenges, we make three original contributions in this paper.
First, we conduct an empirical study of several existing video classification
and object detection methods to identify guns in videos. Our extensive analysis
shows that these methods may not accurately detect guns in videos. Second, we
propose a novel two-stage gun detection method. In stage 1, we train an
image-augmented model to effectively classify ``Gun'' videos. To make the
detection more precise and efficient, stage 2 employs an object detection model
to locate the exact region of the gun within video frames for videos classified
as ``Gun'' by stage 1. Third, our experimental results demonstrate that the
proposed domain-specific method achieves significant performance improvements
and enhances efficiency compared with existing techniques. We also discuss
challenges and future research directions in gun detection tasks in computer
vision.
| no_new_dataset | 0.949153 |
2503.06321 | Rumman Ahmed Prodhan | Md Ohiduzzaman Ovi, Maliha Sanjana, Fahad Fahad, Mahjabin Runa, Zarin
Tasnim Rothy, Tanmoy Sarkar Pias, A.M. Tayeful Islam, and Rumman Ahmed
Prodhan | Enhanced Pediatric Dental Segmentation Using a Custom SegUNet with VGG19
Backbone on Panoramic Radiographs | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pediatric dental segmentation is critical in dental diagnostics, presenting
unique challenges due to variations in dental structures and the lower number
of pediatric X-ray images. This study proposes a custom SegUNet model with a
VGG19 backbone, designed explicitly for pediatric dental segmentation and
applied to the Children's Dental Panoramic Radiographs dataset. The SegUNet
architecture with a VGG19 backbone has been employed on this dataset for the
first time, achieving state-of-the-art performance. The model reached an
accuracy of 97.53%, a dice coefficient of 92.49%, and an intersection over
union (IOU) of 91.46%, setting a new benchmark for this dataset. These results
demonstrate the effectiveness of the VGG19 backbone in enhancing feature
extraction and improving segmentation precision. Comprehensive evaluations
across metrics, including precision, recall, and specificity, indicate the
robustness of this approach. The model's ability to generalize across diverse
dental structures makes it a valuable tool for clinical applications in
pediatric dental care. It offers a reliable and efficient solution for
automated dental diagnostics.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 19:32:25 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Ovi",
"Md Ohiduzzaman",
""
],
[
"Sanjana",
"Maliha",
""
],
[
"Fahad",
"Fahad",
""
],
[
"Runa",
"Mahjabin",
""
],
[
"Rothy",
"Zarin Tasnim",
""
],
[
"Pias",
"Tanmoy Sarkar",
""
],
[
"Islam",
"A. M. Tayeful",
""
],
[
"Prodhan",
"Rumman Ahmed",
""
]
]
| TITLE: Enhanced Pediatric Dental Segmentation Using a Custom SegUNet with VGG19
Backbone on Panoramic Radiographs
ABSTRACT: Pediatric dental segmentation is critical in dental diagnostics, presenting
unique challenges due to variations in dental structures and the lower number
of pediatric X-ray images. This study proposes a custom SegUNet model with a
VGG19 backbone, designed explicitly for pediatric dental segmentation and
applied to the Children's Dental Panoramic Radiographs dataset. The SegUNet
architecture with a VGG19 backbone has been employed on this dataset for the
first time, achieving state-of-the-art performance. The model reached an
accuracy of 97.53%, a dice coefficient of 92.49%, and an intersection over
union (IOU) of 91.46%, setting a new benchmark for this dataset. These results
demonstrate the effectiveness of the VGG19 backbone in enhancing feature
extraction and improving segmentation precision. Comprehensive evaluations
across metrics, including precision, recall, and specificity, indicate the
robustness of this approach. The model's ability to generalize across diverse
dental structures makes it a valuable tool for clinical applications in
pediatric dental care. It offers a reliable and efficient solution for
automated dental diagnostics.
| new_dataset | 0.964422 |
2503.06352 | Xiao Yue | Xiao Yue, Guangzhi Qu, Lige Gan | GIN-Graph: A Generative Interpretation Network for Model-Level
Explanation of Graph Neural Networks | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One significant challenge of exploiting Graph neural networks (GNNs) in
real-life scenarios is that they are always treated as black boxes, therefore
leading to the requirement of interpretability. Model-level interpretations
explain what patterns maximize probability of predicting to a certain class.
However, existing model-level interpretation methods pose several limitations
such as generating invalid explanation graphs and requiring extreme fine-tuning
on hyperparameters manually. In this paper, we propose a new Generative
Interpretation Network for Model-Level Explanation of Graph Neural Networks
(GIN-Graph), to generate reliable model-level explanation graphs. The implicit
and likelihood-free generative adversarial networks are exploited to construct
explanation graphs similar to original graphs, meanwhile maximizing the
prediction probability for a certain class by adopting a novel objective
function. Experimental results indicate that GIN-Graph can be easily applied to
GNN models trained on a variety of graph datasets to create meaningful
explanation graphs without requiring extensive fine-tuning on hyperparameters.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 22:39:36 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Yue",
"Xiao",
""
],
[
"Qu",
"Guangzhi",
""
],
[
"Gan",
"Lige",
""
]
]
| TITLE: GIN-Graph: A Generative Interpretation Network for Model-Level
Explanation of Graph Neural Networks
ABSTRACT: One significant challenge of exploiting Graph neural networks (GNNs) in
real-life scenarios is that they are always treated as black boxes, therefore
leading to the requirement of interpretability. Model-level interpretations
explain what patterns maximize probability of predicting to a certain class.
However, existing model-level interpretation methods pose several limitations
such as generating invalid explanation graphs and requiring extreme fine-tuning
on hyperparameters manually. In this paper, we propose a new Generative
Interpretation Network for Model-Level Explanation of Graph Neural Networks
(GIN-Graph), to generate reliable model-level explanation graphs. The implicit
and likelihood-free generative adversarial networks are exploited to construct
explanation graphs similar to original graphs, meanwhile maximizing the
prediction probability for a certain class by adopting a novel objective
function. Experimental results indicate that GIN-Graph can be easily applied to
GNN models trained on a variety of graph datasets to create meaningful
explanation graphs without requiring extensive fine-tuning on hyperparameters.
| no_new_dataset | 0.951188 |
2503.06362 | Umberto Cappellazzo | Umberto Cappellazzo, Minsu Kim, Stavros Petridis | Adaptive Audio-Visual Speech Recognition via Matryoshka-Based Multimodal
LLMs | null | null | null | null | cs.CV cs.MM cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | Audio-Visual Speech Recognition (AVSR) leverages both audio and visual
modalities to enhance speech recognition robustness, particularly in noisy
environments. Recent advancements in Large Language Models (LLMs) have
demonstrated their effectiveness in speech recognition, including AVSR.
However, due to the significant length of speech representations, direct
integration with LLMs imposes substantial computational costs. Prior approaches
address this by compressing speech representations before feeding them into
LLMs. However, higher compression ratios often lead to performance degradation,
necessitating a trade-off between computational efficiency and recognition
accuracy. To address this challenge, we propose Llama-MTSK, the first
Matryoshka-based Multimodal LLM for AVSR, which enables flexible adaptation of
the audio-visual token allocation based on specific computational constraints
while preserving high performance. Our approach, inspired by Matryoshka
Representation Learning, encodes audio-visual representations at multiple
granularities within a single model, eliminating the need to train separate
models for different compression levels. Moreover, to efficiently fine-tune the
LLM, we introduce three LoRA-based Matryoshka strategies using global and
scale-specific LoRA modules. Extensive evaluations on the two largest AVSR
datasets demonstrate that Llama-MTSK achieves state-of-the-art results,
matching or surpassing models trained independently at fixed compression
levels.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 00:02:10 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Cappellazzo",
"Umberto",
""
],
[
"Kim",
"Minsu",
""
],
[
"Petridis",
"Stavros",
""
]
]
| TITLE: Adaptive Audio-Visual Speech Recognition via Matryoshka-Based Multimodal
LLMs
ABSTRACT: Audio-Visual Speech Recognition (AVSR) leverages both audio and visual
modalities to enhance speech recognition robustness, particularly in noisy
environments. Recent advancements in Large Language Models (LLMs) have
demonstrated their effectiveness in speech recognition, including AVSR.
However, due to the significant length of speech representations, direct
integration with LLMs imposes substantial computational costs. Prior approaches
address this by compressing speech representations before feeding them into
LLMs. However, higher compression ratios often lead to performance degradation,
necessitating a trade-off between computational efficiency and recognition
accuracy. To address this challenge, we propose Llama-MTSK, the first
Matryoshka-based Multimodal LLM for AVSR, which enables flexible adaptation of
the audio-visual token allocation based on specific computational constraints
while preserving high performance. Our approach, inspired by Matryoshka
Representation Learning, encodes audio-visual representations at multiple
granularities within a single model, eliminating the need to train separate
models for different compression levels. Moreover, to efficiently fine-tune the
LLM, we introduce three LoRA-based Matryoshka strategies using global and
scale-specific LoRA modules. Extensive evaluations on the two largest AVSR
datasets demonstrate that Llama-MTSK achieves state-of-the-art results,
matching or surpassing models trained independently at fixed compression
levels.
| no_new_dataset | 0.943348 |
2503.06366 | Henry Kvinge | Herman Chau, Helen Jenne, Davis Brown, Jesse He, Mark Raugas, Sara
Billey, Henry Kvinge | Machine Learning meets Algebraic Combinatorics: A Suite of Datasets
Capturing Research-level Conjecturing Ability in Pure Mathematics | 26 pages, comments welcome | null | null | null | cs.LG cs.AI math.CO math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With recent dramatic increases in AI system capabilities, there has been
growing interest in utilizing machine learning for reasoning-heavy,
quantitative tasks, particularly mathematics. While there are many resources
capturing mathematics at the high-school, undergraduate, and graduate level,
there are far fewer resources available that align with the level of difficulty
and open endedness encountered by professional mathematicians working on open
problems. To address this, we introduce a new collection of datasets, the
Algebraic Combinatorics Dataset Repository (ACD Repo), representing either
foundational results or open problems in algebraic combinatorics, a subfield of
mathematics that studies discrete structures arising from abstract algebra.
Further differentiating our dataset collection is the fact that it aims at the
conjecturing process. Each dataset includes an open-ended research-level
question and a large collection of examples (up to 10M in some cases) from
which conjectures should be generated. We describe all nine datasets, the
different ways machine learning models can be applied to them (e.g., training
with narrow models followed by interpretability analysis or program synthesis
with LLMs), and discuss some of the challenges involved in designing datasets
like these.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 00:11:40 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Chau",
"Herman",
""
],
[
"Jenne",
"Helen",
""
],
[
"Brown",
"Davis",
""
],
[
"He",
"Jesse",
""
],
[
"Raugas",
"Mark",
""
],
[
"Billey",
"Sara",
""
],
[
"Kvinge",
"Henry",
""
]
]
| TITLE: Machine Learning meets Algebraic Combinatorics: A Suite of Datasets
Capturing Research-level Conjecturing Ability in Pure Mathematics
ABSTRACT: With recent dramatic increases in AI system capabilities, there has been
growing interest in utilizing machine learning for reasoning-heavy,
quantitative tasks, particularly mathematics. While there are many resources
capturing mathematics at the high-school, undergraduate, and graduate level,
there are far fewer resources available that align with the level of difficulty
and open endedness encountered by professional mathematicians working on open
problems. To address this, we introduce a new collection of datasets, the
Algebraic Combinatorics Dataset Repository (ACD Repo), representing either
foundational results or open problems in algebraic combinatorics, a subfield of
mathematics that studies discrete structures arising from abstract algebra.
Further differentiating our dataset collection is the fact that it aims at the
conjecturing process. Each dataset includes an open-ended research-level
question and a large collection of examples (up to 10M in some cases) from
which conjectures should be generated. We describe all nine datasets, the
different ways machine learning models can be applied to them (e.g., training
with narrow models followed by interpretability analysis or program synthesis
with LLMs), and discuss some of the challenges involved in designing datasets
like these.
| new_dataset | 0.969179 |
2503.06368 | Leonardo Scabini | Leonardo Scabini, Kallil M. Zielinski, Emir Konuk, Ricardo T. Fares,
Lucas C. Ribas, Kevin Smith, and Odemir M. Bruno | VORTEX: Challenging CNNs at Texture Recognition by using Vision
Transformers with Orderless and Randomized Token Encodings | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Texture recognition has recently been dominated by ImageNet-pre-trained deep
Convolutional Neural Networks (CNNs), with specialized modifications and
feature engineering required to achieve state-of-the-art (SOTA) performance.
However, although Vision Transformers (ViTs) were introduced a few years ago,
little is known about their texture recognition ability. Therefore, in this
work, we introduce VORTEX (ViTs with Orderless and Randomized Token Encodings
for Texture Recognition), a novel method that enables the effective use of ViTs
for texture analysis. VORTEX extracts multi-depth token embeddings from
pre-trained ViT backbones and employs a lightweight module to aggregate
hierarchical features and perform orderless encoding, obtaining a better image
representation for texture recognition tasks. This approach allows seamless
integration with any ViT with the common transformer architecture. Moreover, no
fine-tuning of the backbone is performed, since they are used only as frozen
feature extractors, and the features are fed to a linear SVM. We evaluate
VORTEX on nine diverse texture datasets, demonstrating its ability to achieve
or surpass SOTA performance in a variety of texture analysis scenarios. By
bridging the gap between texture recognition with CNNs and transformer-based
architectures, VORTEX paves the way for adopting emerging transformer
foundation models. Furthermore, VORTEX demonstrates robust computational
efficiency when coupled with ViT backbones compared to CNNs with similar costs.
The method implementation and experimental scripts are publicly available in
our online repository.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 00:36:02 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Scabini",
"Leonardo",
""
],
[
"Zielinski",
"Kallil M.",
""
],
[
"Konuk",
"Emir",
""
],
[
"Fares",
"Ricardo T.",
""
],
[
"Ribas",
"Lucas C.",
""
],
[
"Smith",
"Kevin",
""
],
[
"Bruno",
"Odemir M.",
""
]
]
| TITLE: VORTEX: Challenging CNNs at Texture Recognition by using Vision
Transformers with Orderless and Randomized Token Encodings
ABSTRACT: Texture recognition has recently been dominated by ImageNet-pre-trained deep
Convolutional Neural Networks (CNNs), with specialized modifications and
feature engineering required to achieve state-of-the-art (SOTA) performance.
However, although Vision Transformers (ViTs) were introduced a few years ago,
little is known about their texture recognition ability. Therefore, in this
work, we introduce VORTEX (ViTs with Orderless and Randomized Token Encodings
for Texture Recognition), a novel method that enables the effective use of ViTs
for texture analysis. VORTEX extracts multi-depth token embeddings from
pre-trained ViT backbones and employs a lightweight module to aggregate
hierarchical features and perform orderless encoding, obtaining a better image
representation for texture recognition tasks. This approach allows seamless
integration with any ViT with the common transformer architecture. Moreover, no
fine-tuning of the backbone is performed, since they are used only as frozen
feature extractors, and the features are fed to a linear SVM. We evaluate
VORTEX on nine diverse texture datasets, demonstrating its ability to achieve
or surpass SOTA performance in a variety of texture analysis scenarios. By
bridging the gap between texture recognition with CNNs and transformer-based
architectures, VORTEX paves the way for adopting emerging transformer
foundation models. Furthermore, VORTEX demonstrates robust computational
efficiency when coupled with ViT backbones compared to CNNs with similar costs.
The method implementation and experimental scripts are publicly available in
our online repository.
| no_new_dataset | 0.949529 |
2503.06370 | Saba Sanami | Hesam Mosalli, Saba Sanami, Yu Yang, Hen-Geul Yeh, Amir G. Aghdam | Dynamic Load Balancing for EV Charging Stations Using Reinforcement
Learning and Demand Prediction | 19th Annual IEEE International Systems Conference (SysCon 2025) | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper presents a method for load balancing and dynamic pricing in
electric vehicle (EV) charging networks, utilizing reinforcement learning (RL)
to enhance network performance. The proposed framework integrates a pre-trained
graph neural network to predict demand elasticity and inform pricing decisions.
The spatio-temporal EV charging demand prediction (EVCDP) dataset from Shenzhen
is utilized to capture the geographic and temporal characteristics of the
charging stations. The RL model dynamically adjusts prices at individual
stations based on occupancy, maximum station capacity, and demand forecasts,
ensuring an equitable network load distribution while preventing station
overloads. By leveraging spatially-aware demand predictions and a carefully
designed reward function, the framework achieves efficient load balancing and
adaptive pricing strategies that respond to localized demand and global network
dynamics, ensuring improved network stability and user satisfaction. The
efficacy of the approach is validated through simulations on the dataset,
showing significant improvements in load balancing and reduced overload as the
RL agent iteratively interacts with the environment and learns to dynamically
adjust pricing strategies based on real-time demand patterns and station
constraints. The findings highlight the potential of adaptive pricing and
load-balancing strategies to address the complexities of EV infrastructure,
paving the way for scalable and user-centric solutions.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 00:40:35 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Mosalli",
"Hesam",
""
],
[
"Sanami",
"Saba",
""
],
[
"Yang",
"Yu",
""
],
[
"Yeh",
"Hen-Geul",
""
],
[
"Aghdam",
"Amir G.",
""
]
]
| TITLE: Dynamic Load Balancing for EV Charging Stations Using Reinforcement
Learning and Demand Prediction
ABSTRACT: This paper presents a method for load balancing and dynamic pricing in
electric vehicle (EV) charging networks, utilizing reinforcement learning (RL)
to enhance network performance. The proposed framework integrates a pre-trained
graph neural network to predict demand elasticity and inform pricing decisions.
The spatio-temporal EV charging demand prediction (EVCDP) dataset from Shenzhen
is utilized to capture the geographic and temporal characteristics of the
charging stations. The RL model dynamically adjusts prices at individual
stations based on occupancy, maximum station capacity, and demand forecasts,
ensuring an equitable network load distribution while preventing station
overloads. By leveraging spatially-aware demand predictions and a carefully
designed reward function, the framework achieves efficient load balancing and
adaptive pricing strategies that respond to localized demand and global network
dynamics, ensuring improved network stability and user satisfaction. The
efficacy of the approach is validated through simulations on the dataset,
showing significant improvements in load balancing and reduced overload as the
RL agent iteratively interacts with the environment and learns to dynamically
adjust pricing strategies based on real-time demand patterns and station
constraints. The findings highlight the potential of adaptive pricing and
load-balancing strategies to address the complexities of EV infrastructure,
paving the way for scalable and user-centric solutions.
| no_new_dataset | 0.951729 |
2503.06382 | Yuanhao Cai | Guofeng Zhang, Ruyi Zha, Hao He, Yixun Liang, Alan Yuille, Hongdong
Li, Yuanhao Cai | X-LRM: X-ray Large Reconstruction Model for Extremely Sparse-View
Computed Tomography Recovery in One Second | A large reconstruction model and the largest dataset (16K samples)
for sparse-view CT recovery | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sparse-view 3D CT reconstruction aims to recover volumetric structures from a
limited number of 2D X-ray projections. Existing feedforward methods are
constrained by the limited capacity of CNN-based architectures and the scarcity
of large-scale training datasets. In this paper, we propose an X-ray Large
Reconstruction Model (X-LRM) for extremely sparse-view (<10 views) CT
reconstruction. X-LRM consists of two key components: X-former and X-triplane.
Our X-former can handle an arbitrary number of input views using an MLP-based
image tokenizer and a Transformer-based encoder. The output tokens are then
upsampled into our X-triplane representation, which models the 3D radiodensity
as an implicit neural field. To support the training of X-LRM, we introduce
Torso-16K, a large-scale dataset comprising over 16K volume-projection pairs of
various torso organs. Extensive experiments demonstrate that X-LRM outperforms
the state-of-the-art method by 1.5 dB and achieves 27x faster speed and better
flexibility. Furthermore, the downstream evaluation of lung segmentation tasks
also suggests the practical value of our approach. Our code, pre-trained
models, and dataset will be released at https://github.com/caiyuanhao1998/X-LRM
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 01:39:59 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhang",
"Guofeng",
""
],
[
"Zha",
"Ruyi",
""
],
[
"He",
"Hao",
""
],
[
"Liang",
"Yixun",
""
],
[
"Yuille",
"Alan",
""
],
[
"Li",
"Hongdong",
""
],
[
"Cai",
"Yuanhao",
""
]
]
| TITLE: X-LRM: X-ray Large Reconstruction Model for Extremely Sparse-View
Computed Tomography Recovery in One Second
ABSTRACT: Sparse-view 3D CT reconstruction aims to recover volumetric structures from a
limited number of 2D X-ray projections. Existing feedforward methods are
constrained by the limited capacity of CNN-based architectures and the scarcity
of large-scale training datasets. In this paper, we propose an X-ray Large
Reconstruction Model (X-LRM) for extremely sparse-view (<10 views) CT
reconstruction. X-LRM consists of two key components: X-former and X-triplane.
Our X-former can handle an arbitrary number of input views using an MLP-based
image tokenizer and a Transformer-based encoder. The output tokens are then
upsampled into our X-triplane representation, which models the 3D radiodensity
as an implicit neural field. To support the training of X-LRM, we introduce
Torso-16K, a large-scale dataset comprising over 16K volume-projection pairs of
various torso organs. Extensive experiments demonstrate that X-LRM outperforms
the state-of-the-art method by 1.5 dB and achieves 27x faster speed and better
flexibility. Furthermore, the downstream evaluation of lung segmentation tasks
also suggests the practical value of our approach. Our code, pre-trained
models, and dataset will be released at https://github.com/caiyuanhao1998/X-LRM
| new_dataset | 0.962036 |
2503.06387 | Joseph Wilson | Anurag Swarnim Yadav and Joseph N. Wilson | R+R: Security Vulnerability Dataset Quality Is Critical | 15 pages, 1 figure, 35 tables. To be published in Proceedings of the
2024 Annual Computer Security Applications Conference (ACSAC) | null | null | null | cs.SE cs.CR | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) are of great interest in vulnerability detection
and repair. The effectiveness of these models hinges on the quality of the
datasets used for both training and evaluation. Our investigation reveals that
a number of studies featured in prominent software engineering conferences have
employed datasets that are plagued by high duplication rates, questionable
label accuracy, and incomplete samples. Using these datasets for
experimentation will yield incorrect results that are significantly different
from actual expected behavior. For example, the state-of-the-art VulRepair
Model, which is reported to have 44% accuracy, on average yielded 9% accuracy
when test-set duplicates were removed from its training set and 13% accuracy
when training-set duplicates were removed from its test set. In an effort to
tackle these data quality concerns, we have retrained models from several
papers without duplicates and conducted an accuracy assessment of labels for
the top ten most hazardous Common Weakness Enumerations (CWEs). Our findings
indicate that 56% of the samples had incorrect labels and 44% comprised
incomplete samples--only 31% were both accurate and complete. Finally, we
employ transfer learning using a large deduplicated bugfix corpus to show that
these models can exhibit better performance if given larger amounts of
high-quality pre-training data, leading us to conclude that while previous
studies have over-estimated performance due to poor dataset quality, this does
not demonstrate that better performance is not possible.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 01:49:30 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Yadav",
"Anurag Swarnim",
""
],
[
"Wilson",
"Joseph N.",
""
]
]
| TITLE: R+R: Security Vulnerability Dataset Quality Is Critical
ABSTRACT: Large Language Models (LLMs) are of great interest in vulnerability detection
and repair. The effectiveness of these models hinges on the quality of the
datasets used for both training and evaluation. Our investigation reveals that
a number of studies featured in prominent software engineering conferences have
employed datasets that are plagued by high duplication rates, questionable
label accuracy, and incomplete samples. Using these datasets for
experimentation will yield incorrect results that are significantly different
from actual expected behavior. For example, the state-of-the-art VulRepair
Model, which is reported to have 44% accuracy, on average yielded 9% accuracy
when test-set duplicates were removed from its training set and 13% accuracy
when training-set duplicates were removed from its test set. In an effort to
tackle these data quality concerns, we have retrained models from several
papers without duplicates and conducted an accuracy assessment of labels for
the top ten most hazardous Common Weakness Enumerations (CWEs). Our findings
indicate that 56% of the samples had incorrect labels and 44% comprised
incomplete samples--only 31% were both accurate and complete. Finally, we
employ transfer learning using a large deduplicated bugfix corpus to show that
these models can exhibit better performance if given larger amounts of
high-quality pre-training data, leading us to conclude that while previous
studies have over-estimated performance due to poor dataset quality, this does
not demonstrate that better performance is not possible.
| no_new_dataset | 0.946349 |
2503.06392 | Tao Feng | Tao Feng, Yunke Zhang, Huandong Wang, Yong Li | EPR-GAIL: An EPR-Enhanced Hierarchical Imitation Learning Framework to
Simulate Complex User Consumption Behaviors | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | User consumption behavior data, which records individuals' online spending
history at various types of stores, has been widely used in various
applications, such as store recommendation, site selection, and sale
forecasting. However, its high worth is limited due to deficiencies in data
comprehensiveness and changes of application scenarios. Thus, generating
high-quality sequential consumption data by simulating complex user consumption
behaviors is of great importance to real-world applications. Two branches of
existing sequence generation methods are both limited in quality. Model-based
methods with simplified assumptions fail to model the complex decision process
of user consumption, while data-driven methods that emulate real-world data are
prone to noises, unobserved behaviors, and dynamic decision space. In this
work, we propose to enhance the fidelity and trustworthiness of the data-driven
Generative Adversarial Imitation Learning (GAIL) method by blending it with the
Exploration and Preferential Return EPR model . The core idea of our EPR-GAIL
framework is to model user consumption behaviors as a complex EPR decision
process, which consists of purchase, exploration, and preference decisions.
Specifically, we design the hierarchical policy function in the generator as a
realization of the EPR decision process and employ the probability
distributions of the EPR model to guide the reward function in the
discriminator. Extensive experiments on two real-world datasets of user
consumption behaviors on an online platform demonstrate that the EPR-GAIL
framework outperforms the best state-of-the-art baseline by over 19\% in terms
of data fidelity. Furthermore, the generated consumption behavior data can
improve the performance of sale prediction and location recommendation by up to
35.29% and 11.19%, respectively, validating its advantage for practical
applications.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 01:56:42 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Feng",
"Tao",
""
],
[
"Zhang",
"Yunke",
""
],
[
"Wang",
"Huandong",
""
],
[
"Li",
"Yong",
""
]
]
| TITLE: EPR-GAIL: An EPR-Enhanced Hierarchical Imitation Learning Framework to
Simulate Complex User Consumption Behaviors
ABSTRACT: User consumption behavior data, which records individuals' online spending
history at various types of stores, has been widely used in various
applications, such as store recommendation, site selection, and sale
forecasting. However, its high worth is limited due to deficiencies in data
comprehensiveness and changes of application scenarios. Thus, generating
high-quality sequential consumption data by simulating complex user consumption
behaviors is of great importance to real-world applications. Two branches of
existing sequence generation methods are both limited in quality. Model-based
methods with simplified assumptions fail to model the complex decision process
of user consumption, while data-driven methods that emulate real-world data are
prone to noises, unobserved behaviors, and dynamic decision space. In this
work, we propose to enhance the fidelity and trustworthiness of the data-driven
Generative Adversarial Imitation Learning (GAIL) method by blending it with the
Exploration and Preferential Return EPR model . The core idea of our EPR-GAIL
framework is to model user consumption behaviors as a complex EPR decision
process, which consists of purchase, exploration, and preference decisions.
Specifically, we design the hierarchical policy function in the generator as a
realization of the EPR decision process and employ the probability
distributions of the EPR model to guide the reward function in the
discriminator. Extensive experiments on two real-world datasets of user
consumption behaviors on an online platform demonstrate that the EPR-GAIL
framework outperforms the best state-of-the-art baseline by over 19\% in terms
of data fidelity. Furthermore, the generated consumption behavior data can
improve the performance of sale prediction and location recommendation by up to
35.29% and 11.19%, respectively, validating its advantage for practical
applications.
| no_new_dataset | 0.946597 |
2503.06395 | Tao Feng | Tao Feng, Yunke Zhang, Xiaochen Fan, Huandong Wang, Yong Li | Causal Discovery and Inference towards Urban Elements and Associated
Factors | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | To uncover the city's fundamental functioning mechanisms, it is important to
acquire a deep understanding of complicated relationships among citizens,
location, and mobility behaviors. Previous research studies have applied direct
correlation analysis to investigate such relationships. Nevertheless, due to
the ubiquitous confounding effects, empirical correlation analysis may not
accurately reflect underlying causal relationships among basic urban elements.
In this paper, we propose a novel urban causal computing framework to
comprehensively explore causalities and confounding effects among a variety of
factors across different types of urban elements. In particular, we design a
reinforcement learning algorithm to discover the potential causal graph, which
depicts the causal relations between urban factors. The causal graph further
serves as the guidance for estimating causal effects between pair-wise urban
factors by propensity score matching. After removing the confounding effects
from correlations, we leverage significance levels of causal effects in
downstream urban mobility prediction tasks. Experimental studies on open-source
urban datasets show that the discovered causal graph demonstrates a
hierarchical structure, where citizens affect locations, and they both cause
changes in urban mobility behaviors. Experimental results in urban mobility
prediction tasks further show that the proposed method can effectively reduce
confounding effects and enhance performance of urban computing tasks.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 02:15:04 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Feng",
"Tao",
""
],
[
"Zhang",
"Yunke",
""
],
[
"Fan",
"Xiaochen",
""
],
[
"Wang",
"Huandong",
""
],
[
"Li",
"Yong",
""
]
]
| TITLE: Causal Discovery and Inference towards Urban Elements and Associated
Factors
ABSTRACT: To uncover the city's fundamental functioning mechanisms, it is important to
acquire a deep understanding of complicated relationships among citizens,
location, and mobility behaviors. Previous research studies have applied direct
correlation analysis to investigate such relationships. Nevertheless, due to
the ubiquitous confounding effects, empirical correlation analysis may not
accurately reflect underlying causal relationships among basic urban elements.
In this paper, we propose a novel urban causal computing framework to
comprehensively explore causalities and confounding effects among a variety of
factors across different types of urban elements. In particular, we design a
reinforcement learning algorithm to discover the potential causal graph, which
depicts the causal relations between urban factors. The causal graph further
serves as the guidance for estimating causal effects between pair-wise urban
factors by propensity score matching. After removing the confounding effects
from correlations, we leverage significance levels of causal effects in
downstream urban mobility prediction tasks. Experimental studies on open-source
urban datasets show that the discovered causal graph demonstrates a
hierarchical structure, where citizens affect locations, and they both cause
changes in urban mobility behaviors. Experimental results in urban mobility
prediction tasks further show that the proposed method can effectively reduce
confounding effects and enhance performance of urban computing tasks.
| no_new_dataset | 0.946941 |
2503.06396 | Qiqi Bao | Enqiang Zhu and Qiqi Bao and Yu Zhang and Chanjuan Liu | Optimizing Minimum Vertex Cover Solving via a GCN-assisted Heuristic
Algorithm | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of finding a minimum vertex cover (MVC) in a graph is a
well-known NP-hard problem with significant practical applications in
optimization and scheduling. Its complexity, combined with the increasing scale
of problems, underscores the need for efficient and effective algorithms.
However, existing heuristic algorithms for MVC often rely on simplistic
initialization strategies and overlook the impact of edge attributes and
neighborhood information on vertex selection. In this paper, we introduce
GCNIVC, a novel heuristic search algorithm designed to address the limitations
of existing methods for solving MVC problems in large-scale graphs. Our
approach features two main innovations. First, it utilizes a Graph
Convolutional Network (GCN) to capture the global structure of graphs, which
enables the generation of high-quality initial solutions that enhance the
efficiency of the subsequent search process. Second, GCNIVC introduces a new
heuristic that employs three containers and the concept of double-covered edges
(dc-edges), improving search efficiency and providing greater flexibility for
adding and removing operations based on edge attributes. Through extensive
experiments on benchmark datasets, we demonstrate that GCNIVC outperforms
state-of-the-art MVC algorithms in terms of both accuracy and efficiency. Our
results highlight the effectiveness of GCNIVC's GCN-assisted initialization and
its edge-informed search strategy. This study not only advances the
understanding of MVC problem-solving but also contributes a new tool for
addressing large-scale graph optimization challenges.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 02:31:03 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhu",
"Enqiang",
""
],
[
"Bao",
"Qiqi",
""
],
[
"Zhang",
"Yu",
""
],
[
"Liu",
"Chanjuan",
""
]
]
| TITLE: Optimizing Minimum Vertex Cover Solving via a GCN-assisted Heuristic
Algorithm
ABSTRACT: The problem of finding a minimum vertex cover (MVC) in a graph is a
well-known NP-hard problem with significant practical applications in
optimization and scheduling. Its complexity, combined with the increasing scale
of problems, underscores the need for efficient and effective algorithms.
However, existing heuristic algorithms for MVC often rely on simplistic
initialization strategies and overlook the impact of edge attributes and
neighborhood information on vertex selection. In this paper, we introduce
GCNIVC, a novel heuristic search algorithm designed to address the limitations
of existing methods for solving MVC problems in large-scale graphs. Our
approach features two main innovations. First, it utilizes a Graph
Convolutional Network (GCN) to capture the global structure of graphs, which
enables the generation of high-quality initial solutions that enhance the
efficiency of the subsequent search process. Second, GCNIVC introduces a new
heuristic that employs three containers and the concept of double-covered edges
(dc-edges), improving search efficiency and providing greater flexibility for
adding and removing operations based on edge attributes. Through extensive
experiments on benchmark datasets, we demonstrate that GCNIVC outperforms
state-of-the-art MVC algorithms in terms of both accuracy and efficiency. Our
results highlight the effectiveness of GCNIVC's GCN-assisted initialization and
its edge-informed search strategy. This study not only advances the
understanding of MVC problem-solving but also contributes a new tool for
addressing large-scale graph optimization challenges.
| no_new_dataset | 0.944995 |
2503.06397 | Yanyu Zhu | Yanyu Zhu, Licheng Bai, Jintao Xu, Jiwei Tang, Hai-tao Zheng | Removing Averaging: Personalized Lip-Sync Driven Characters Based on
Identity Adapter | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in diffusion-based lip-syncing generative models have
demonstrated their ability to produce highly synchronized talking face videos
for visual dubbing. Although these models excel at lip synchronization, they
often struggle to maintain fine-grained control over facial details in
generated images. In this work, we identify "lip averaging" phenomenon where
the model fails to preserve subtle facial details when dubbing unseen
in-the-wild videos. This issue arises because the commonly used UNet backbone
primarily integrates audio features into visual representations in the latent
space via cross-attention mechanisms and multi-scale fusion, but it struggles
to retain fine-grained lip details in the generated faces. To address this
issue, we propose UnAvgLip, which extracts identity embeddings from reference
videos to generate highly faithful facial sequences while maintaining accurate
lip synchronization. Specifically, our method comprises two primary components:
(1) an Identity Perceiver module that encodes facial embeddings to align with
conditioned audio features; and (2) an ID-CrossAttn module that injects facial
embeddings into the generation process, enhancing model's capability of
identity retention. Extensive experiments demonstrate that, at a modest
training and inference cost, UnAvgLip effectively mitigates the "averaging"
phenomenon in lip inpainting, significantly preserving unique facial
characteristics while maintaining precise lip synchronization. Compared with
the original approach, our method demonstrates significant improvements of 5%
on the identity consistency metric and 2% on the SSIM metric across two
benchmark datasets (HDTF and LRW).
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 02:36:31 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhu",
"Yanyu",
""
],
[
"Bai",
"Licheng",
""
],
[
"Xu",
"Jintao",
""
],
[
"Tang",
"Jiwei",
""
],
[
"Zheng",
"Hai-tao",
""
]
]
| TITLE: Removing Averaging: Personalized Lip-Sync Driven Characters Based on
Identity Adapter
ABSTRACT: Recent advances in diffusion-based lip-syncing generative models have
demonstrated their ability to produce highly synchronized talking face videos
for visual dubbing. Although these models excel at lip synchronization, they
often struggle to maintain fine-grained control over facial details in
generated images. In this work, we identify "lip averaging" phenomenon where
the model fails to preserve subtle facial details when dubbing unseen
in-the-wild videos. This issue arises because the commonly used UNet backbone
primarily integrates audio features into visual representations in the latent
space via cross-attention mechanisms and multi-scale fusion, but it struggles
to retain fine-grained lip details in the generated faces. To address this
issue, we propose UnAvgLip, which extracts identity embeddings from reference
videos to generate highly faithful facial sequences while maintaining accurate
lip synchronization. Specifically, our method comprises two primary components:
(1) an Identity Perceiver module that encodes facial embeddings to align with
conditioned audio features; and (2) an ID-CrossAttn module that injects facial
embeddings into the generation process, enhancing model's capability of
identity retention. Extensive experiments demonstrate that, at a modest
training and inference cost, UnAvgLip effectively mitigates the "averaging"
phenomenon in lip inpainting, significantly preserving unique facial
characteristics while maintaining precise lip synchronization. Compared with
the original approach, our method demonstrates significant improvements of 5%
on the identity consistency metric and 2% on the SSIM metric across two
benchmark datasets (HDTF and LRW).
| no_new_dataset | 0.946745 |
2503.06398 | Tao Feng | Tao Feng, Yunke Zhang, Huandong Wang, Yong Li | Causality Enhanced Origin-Destination Flow Prediction in Data-Scarce
Cities | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Accurate origin-destination (OD) flow prediction is of great importance to
developing cities, as it can contribute to optimize urban structures and
layouts. However, with the common issues of missing regional features and
lacking OD flow data, it is quite daunting to predict OD flow in developing
cities. To address this challenge, we propose a novel Causality-Enhanced OD
Flow Prediction (CE-OFP), a unified framework that aims to transfer urban
knowledge between cities and achieve accuracy improvements in OD flow
predictions across data-scarce cities. In specific, we propose a novel
reinforcement learning model to discover universal causalities among urban
features in data-rich cities and build corresponding causal graphs. Then, we
further build Causality-Enhanced Variational Auto-Encoder (CE-VAE) to
incorporate causal graphs for effective feature reconstruction in data-scarce
cities. Finally, with the reconstructed features, we devise a knowledge
distillation method with a graph attention network to migrate the OD prediction
model from data-rich cities to data-scare cities. Extensive experiments on two
pairs of real-world datasets validate that the proposed CE-OFP remarkably
outperforms state-of-the-art baselines, which can reduce the RMSE of OD flow
prediction for data-scarce cities by up to 11%.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 02:36:36 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Feng",
"Tao",
""
],
[
"Zhang",
"Yunke",
""
],
[
"Wang",
"Huandong",
""
],
[
"Li",
"Yong",
""
]
]
| TITLE: Causality Enhanced Origin-Destination Flow Prediction in Data-Scarce
Cities
ABSTRACT: Accurate origin-destination (OD) flow prediction is of great importance to
developing cities, as it can contribute to optimize urban structures and
layouts. However, with the common issues of missing regional features and
lacking OD flow data, it is quite daunting to predict OD flow in developing
cities. To address this challenge, we propose a novel Causality-Enhanced OD
Flow Prediction (CE-OFP), a unified framework that aims to transfer urban
knowledge between cities and achieve accuracy improvements in OD flow
predictions across data-scarce cities. In specific, we propose a novel
reinforcement learning model to discover universal causalities among urban
features in data-rich cities and build corresponding causal graphs. Then, we
further build Causality-Enhanced Variational Auto-Encoder (CE-VAE) to
incorporate causal graphs for effective feature reconstruction in data-scarce
cities. Finally, with the reconstructed features, we devise a knowledge
distillation method with a graph attention network to migrate the OD prediction
model from data-rich cities to data-scare cities. Extensive experiments on two
pairs of real-world datasets validate that the proposed CE-OFP remarkably
outperforms state-of-the-art baselines, which can reduce the RMSE of OD flow
prediction for data-scarce cities by up to 11%.
| no_new_dataset | 0.946843 |
2503.06422 | Yuteng Zhang | Lin Zhang, Yuteng Zhang, Dusit Niyato, Lei Ren, Pengfei Gu, Zhen Chen,
Yuanjun Laili, Wentong Cai and Agostino Bruzzone | GenAI for Simulation Model in Model-Based Systems Engineering | This work has been submitted to the IEEE for possible publication | null | null | null | cs.SE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generative AI (GenAI) has demonstrated remarkable capabilities in code
generation, and its integration into complex product modeling and simulation
code generation can significantly enhance the efficiency of the system design
phase in Model-Based Systems Engineering (MBSE). In this study, we introduce a
generative system design methodology framework for MBSE, offering a practical
approach for the intelligent generation of simulation models for system
physical properties. First, we employ inference techniques, generative models,
and integrated modeling and simulation languages to construct simulation models
for system physical properties based on product design documents. Subsequently,
we fine-tune the language model used for simulation model generation on an
existing library of simulation models and additional datasets generated through
generative modeling. Finally, we introduce evaluation metrics for the generated
simulation models for system physical properties. Our proposed approach to
simulation model generation presents the innovative concept of scalable
templates for simulation models. Using these templates, GenAI generates
simulation models for system physical properties through code completion. The
experimental results demonstrate that, for mainstream open-source
Transformer-based models, the quality of the simulation model is significantly
improved using the simulation model generation method proposed in this paper.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 03:33:25 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhang",
"Lin",
""
],
[
"Zhang",
"Yuteng",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Ren",
"Lei",
""
],
[
"Gu",
"Pengfei",
""
],
[
"Chen",
"Zhen",
""
],
[
"Laili",
"Yuanjun",
""
],
[
"Cai",
"Wentong",
""
],
[
"Bruzzone",
"Agostino",
""
]
]
| TITLE: GenAI for Simulation Model in Model-Based Systems Engineering
ABSTRACT: Generative AI (GenAI) has demonstrated remarkable capabilities in code
generation, and its integration into complex product modeling and simulation
code generation can significantly enhance the efficiency of the system design
phase in Model-Based Systems Engineering (MBSE). In this study, we introduce a
generative system design methodology framework for MBSE, offering a practical
approach for the intelligent generation of simulation models for system
physical properties. First, we employ inference techniques, generative models,
and integrated modeling and simulation languages to construct simulation models
for system physical properties based on product design documents. Subsequently,
we fine-tune the language model used for simulation model generation on an
existing library of simulation models and additional datasets generated through
generative modeling. Finally, we introduce evaluation metrics for the generated
simulation models for system physical properties. Our proposed approach to
simulation model generation presents the innovative concept of scalable
templates for simulation models. Using these templates, GenAI generates
simulation models for system physical properties through code completion. The
experimental results demonstrate that, for mainstream open-source
Transformer-based models, the quality of the simulation model is significantly
improved using the simulation model generation method proposed in this paper.
| no_new_dataset | 0.948489 |
2503.06428 | Tianshu Huang | Tianshu Huang, Arjun Ramesh, Emily Ruppel, Nuno Pereira, Anthony Rowe,
Carlee Joe-Wong | Interference-Aware Edge Runtime Prediction with Conformal Matrix
Completion | To appear at MLSys 2025 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurately estimating workload runtime is a longstanding goal in computer
systems, and plays a key role in efficient resource provisioning, latency
minimization, and various other system management tasks. Runtime prediction is
particularly important for managing increasingly complex distributed systems in
which more sophisticated processing is pushed to the edge in search of better
latency. Previous approaches for runtime prediction in edge systems suffer from
poor data efficiency or require intensive instrumentation; these challenges are
compounded in heterogeneous edge computing environments, where historical
runtime data may be sparsely available and instrumentation is often
challenging. Moreover, edge computing environments often feature multi-tenancy
due to limited resources at the network edge, potentially leading to
interference between workloads and further complicating the runtime prediction
problem. Drawing from insights across machine learning and computer systems, we
design a matrix factorization-inspired method that generates accurate
interference-aware predictions with tight provably-guaranteed uncertainty
bounds. We validate our method on a novel WebAssembly runtime dataset collected
from 24 unique devices, achieving a prediction error of 5.2% -- 2x better than
a naive application of existing methods.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 03:41:32 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Huang",
"Tianshu",
""
],
[
"Ramesh",
"Arjun",
""
],
[
"Ruppel",
"Emily",
""
],
[
"Pereira",
"Nuno",
""
],
[
"Rowe",
"Anthony",
""
],
[
"Joe-Wong",
"Carlee",
""
]
]
| TITLE: Interference-Aware Edge Runtime Prediction with Conformal Matrix
Completion
ABSTRACT: Accurately estimating workload runtime is a longstanding goal in computer
systems, and plays a key role in efficient resource provisioning, latency
minimization, and various other system management tasks. Runtime prediction is
particularly important for managing increasingly complex distributed systems in
which more sophisticated processing is pushed to the edge in search of better
latency. Previous approaches for runtime prediction in edge systems suffer from
poor data efficiency or require intensive instrumentation; these challenges are
compounded in heterogeneous edge computing environments, where historical
runtime data may be sparsely available and instrumentation is often
challenging. Moreover, edge computing environments often feature multi-tenancy
due to limited resources at the network edge, potentially leading to
interference between workloads and further complicating the runtime prediction
problem. Drawing from insights across machine learning and computer systems, we
design a matrix factorization-inspired method that generates accurate
interference-aware predictions with tight provably-guaranteed uncertainty
bounds. We validate our method on a novel WebAssembly runtime dataset collected
from 24 unique devices, achieving a prediction error of 5.2% -- 2x better than
a naive application of existing methods.
| new_dataset | 0.962356 |
2503.06430 | Zhangchi Qiu | Zhangchi Qiu, Linhao Luo, Zicheng Zhao, Shirui Pan and Alan Wee-Chung
Liew | Graph Retrieval-Augmented LLM for Conversational Recommendation Systems | Accepted by PAKDD 2025 | null | null | null | cs.CL cs.AI cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conversational Recommender Systems (CRSs) have emerged as a transformative
paradigm for offering personalized recommendations through natural language
dialogue. However, they face challenges with knowledge sparsity, as users often
provide brief, incomplete preference statements. While recent methods have
integrated external knowledge sources to mitigate this, they still struggle
with semantic understanding and complex preference reasoning. Recent Large
Language Models (LLMs) demonstrate promising capabilities in natural language
understanding and reasoning, showing significant potential for CRSs.
Nevertheless, due to the lack of domain knowledge, existing LLM-based CRSs
either produce hallucinated recommendations or demand expensive domain-specific
training, which largely limits their applicability. In this work, we present
G-CRS (Graph Retrieval-Augmented Large Language Model for Conversational
Recommender Systems), a novel training-free framework that combines graph
retrieval-augmented generation and in-context learning to enhance LLMs'
recommendation capabilities. Specifically, G-CRS employs a two-stage
retrieve-and-recommend architecture, where a GNN-based graph reasoner first
identifies candidate items, followed by Personalized PageRank exploration to
jointly discover potential items and similar user interactions. These retrieved
contexts are then transformed into structured prompts for LLM reasoning,
enabling contextually grounded recommendations without task-specific training.
Extensive experiments on two public datasets show that G-CRS achieves superior
recommendation performance compared to existing methods without requiring
task-specific training.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 03:56:22 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Qiu",
"Zhangchi",
""
],
[
"Luo",
"Linhao",
""
],
[
"Zhao",
"Zicheng",
""
],
[
"Pan",
"Shirui",
""
],
[
"Liew",
"Alan Wee-Chung",
""
]
]
| TITLE: Graph Retrieval-Augmented LLM for Conversational Recommendation Systems
ABSTRACT: Conversational Recommender Systems (CRSs) have emerged as a transformative
paradigm for offering personalized recommendations through natural language
dialogue. However, they face challenges with knowledge sparsity, as users often
provide brief, incomplete preference statements. While recent methods have
integrated external knowledge sources to mitigate this, they still struggle
with semantic understanding and complex preference reasoning. Recent Large
Language Models (LLMs) demonstrate promising capabilities in natural language
understanding and reasoning, showing significant potential for CRSs.
Nevertheless, due to the lack of domain knowledge, existing LLM-based CRSs
either produce hallucinated recommendations or demand expensive domain-specific
training, which largely limits their applicability. In this work, we present
G-CRS (Graph Retrieval-Augmented Large Language Model for Conversational
Recommender Systems), a novel training-free framework that combines graph
retrieval-augmented generation and in-context learning to enhance LLMs'
recommendation capabilities. Specifically, G-CRS employs a two-stage
retrieve-and-recommend architecture, where a GNN-based graph reasoner first
identifies candidate items, followed by Personalized PageRank exploration to
jointly discover potential items and similar user interactions. These retrieved
contexts are then transformed into structured prompts for LLM reasoning,
enabling contextually grounded recommendations without task-specific training.
Extensive experiments on two public datasets show that G-CRS achieves superior
recommendation performance compared to existing methods without requiring
task-specific training.
| no_new_dataset | 0.949153 |
2503.06435 | Adrian Chow | Adrian Chow, Evelien Riddell, Yimu Wang, Sean Sedwards, Krzysztof
Czarnecki | OV-SCAN: Semantically Consistent Alignment for Novel Object Discovery in
Open-Vocabulary 3D Object Detection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Open-vocabulary 3D object detection for autonomous driving aims to detect
novel objects beyond the predefined training label sets in point cloud scenes.
Existing approaches achieve this by connecting traditional 3D object detectors
with vision-language models (VLMs) to regress 3D bounding boxes for novel
objects and perform open-vocabulary classification through cross-modal
alignment between 3D and 2D features. However, achieving robust cross-modal
alignment remains a challenge due to semantic inconsistencies when generating
corresponding 3D and 2D feature pairs. To overcome this challenge, we present
OV-SCAN, an Open-Vocabulary 3D framework that enforces Semantically Consistent
Alignment for Novel object discovery. OV-SCAN employs two core strategies:
discovering precise 3D annotations and filtering out low-quality or corrupted
alignment pairs (arising from 3D annotation, occlusion-induced, or
resolution-induced noise). Extensive experiments on the nuScenes dataset
demonstrate that OV-SCAN achieves state-of-the-art performance.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 04:22:08 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Chow",
"Adrian",
""
],
[
"Riddell",
"Evelien",
""
],
[
"Wang",
"Yimu",
""
],
[
"Sedwards",
"Sean",
""
],
[
"Czarnecki",
"Krzysztof",
""
]
]
| TITLE: OV-SCAN: Semantically Consistent Alignment for Novel Object Discovery in
Open-Vocabulary 3D Object Detection
ABSTRACT: Open-vocabulary 3D object detection for autonomous driving aims to detect
novel objects beyond the predefined training label sets in point cloud scenes.
Existing approaches achieve this by connecting traditional 3D object detectors
with vision-language models (VLMs) to regress 3D bounding boxes for novel
objects and perform open-vocabulary classification through cross-modal
alignment between 3D and 2D features. However, achieving robust cross-modal
alignment remains a challenge due to semantic inconsistencies when generating
corresponding 3D and 2D feature pairs. To overcome this challenge, we present
OV-SCAN, an Open-Vocabulary 3D framework that enforces Semantically Consistent
Alignment for Novel object discovery. OV-SCAN employs two core strategies:
discovering precise 3D annotations and filtering out low-quality or corrupted
alignment pairs (arising from 3D annotation, occlusion-induced, or
resolution-induced noise). Extensive experiments on the nuScenes dataset
demonstrate that OV-SCAN achieves state-of-the-art performance.
| no_new_dataset | 0.942401 |
2503.06436 | Fan Meng | Fan Meng | Physics-Informed Residual Neural Ordinary Differential Equations for
Enhanced Tropical Cyclone Intensity Forecasting | 14 pages, 9 figures | null | null | null | physics.ao-ph cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Accurate tropical cyclone (TC) intensity prediction is crucial for mitigating
storm hazards, yet its complex dynamics pose challenges to traditional methods.
Here, we introduce a Physics-Informed Residual Neural Ordinary Differential
Equation (PIR-NODE) model to precisely forecast TC intensity evolution. This
model leverages the powerful non-linear fitting capabilities of deep learning,
integrates residual connections to enhance model depth and training stability,
and explicitly models the continuous temporal evolution of TC intensity using
Neural ODEs. Experimental results in the SHIPS dataset demonstrate that the
PIR-NODE model achieves a significant improvement in 24-hour intensity
prediction accuracy compared to traditional statistical models and benchmark
deep learning methods, with a 25. 2\% reduction in the root mean square error
(RMSE) and a 19.5\% increase in R-square (R2) relative to a baseline of neural
network. Crucially, the residual structure effectively preserves initial state
information, and the model exhibits robust generalization capabilities. This
study details the PIR-NODE model architecture, physics-informed integration
strategies, and comprehensive experimental validation, revealing the
substantial potential of deep learning techniques in predicting complex
geophysical systems and laying the foundation for future refined TC forecasting
research.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 04:23:07 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Meng",
"Fan",
""
]
]
| TITLE: Physics-Informed Residual Neural Ordinary Differential Equations for
Enhanced Tropical Cyclone Intensity Forecasting
ABSTRACT: Accurate tropical cyclone (TC) intensity prediction is crucial for mitigating
storm hazards, yet its complex dynamics pose challenges to traditional methods.
Here, we introduce a Physics-Informed Residual Neural Ordinary Differential
Equation (PIR-NODE) model to precisely forecast TC intensity evolution. This
model leverages the powerful non-linear fitting capabilities of deep learning,
integrates residual connections to enhance model depth and training stability,
and explicitly models the continuous temporal evolution of TC intensity using
Neural ODEs. Experimental results in the SHIPS dataset demonstrate that the
PIR-NODE model achieves a significant improvement in 24-hour intensity
prediction accuracy compared to traditional statistical models and benchmark
deep learning methods, with a 25. 2\% reduction in the root mean square error
(RMSE) and a 19.5\% increase in R-square (R2) relative to a baseline of neural
network. Crucially, the residual structure effectively preserves initial state
information, and the model exhibits robust generalization capabilities. This
study details the PIR-NODE model architecture, physics-informed integration
strategies, and comprehensive experimental validation, revealing the
substantial potential of deep learning techniques in predicting complex
geophysical systems and laying the foundation for future refined TC forecasting
research.
| no_new_dataset | 0.951549 |
2503.06439 | Nuoa Lei | Nuoa Lei, Arman Shehabi, Jun Lu, Zhi Cao, Jonathan Koomey, Sarah
Smith, Eric Masanet | Generalizable Machine Learning Models for Predicting Data Center Server
Power, Efficiency, and Throughput | null | null | null | null | cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | In the rapidly evolving digital era, comprehending the intricate dynamics
influencing server power consumption, efficiency, and performance is crucial
for sustainable data center operations. However, existing models lack the
ability to provide a detailed and reliable understanding of these intricate
relationships. This study employs a machine learning-based approach, using the
SPECPower_ssj2008 database, to facilitate user-friendly and generalizable
server modeling. The resulting models demonstrate high accuracy, with errors
falling within approximately 10% on the testing dataset, showcasing their
practical utility and generalizability. Through meticulous analysis, predictive
features related to hardware availability date, server workload level, and
specifications are identified, providing insights into optimizing energy
conservation, efficiency, and performance in server deployment and operation.
By systematically measuring biases and uncertainties, the study underscores the
need for caution when employing historical data for prospective server
modeling, considering the dynamic nature of technology landscapes.
Collectively, this work offers valuable insights into the sustainable
deployment and operation of servers in data centers, paving the way for
enhanced resource use efficiency and more environmentally conscious practices.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 04:39:53 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Lei",
"Nuoa",
""
],
[
"Shehabi",
"Arman",
""
],
[
"Lu",
"Jun",
""
],
[
"Cao",
"Zhi",
""
],
[
"Koomey",
"Jonathan",
""
],
[
"Smith",
"Sarah",
""
],
[
"Masanet",
"Eric",
""
]
]
| TITLE: Generalizable Machine Learning Models for Predicting Data Center Server
Power, Efficiency, and Throughput
ABSTRACT: In the rapidly evolving digital era, comprehending the intricate dynamics
influencing server power consumption, efficiency, and performance is crucial
for sustainable data center operations. However, existing models lack the
ability to provide a detailed and reliable understanding of these intricate
relationships. This study employs a machine learning-based approach, using the
SPECPower_ssj2008 database, to facilitate user-friendly and generalizable
server modeling. The resulting models demonstrate high accuracy, with errors
falling within approximately 10% on the testing dataset, showcasing their
practical utility and generalizability. Through meticulous analysis, predictive
features related to hardware availability date, server workload level, and
specifications are identified, providing insights into optimizing energy
conservation, efficiency, and performance in server deployment and operation.
By systematically measuring biases and uncertainties, the study underscores the
need for caution when employing historical data for prospective server
modeling, considering the dynamic nature of technology landscapes.
Collectively, this work offers valuable insights into the sustainable
deployment and operation of servers in data centers, paving the way for
enhanced resource use efficiency and more environmentally conscious practices.
| no_new_dataset | 0.948394 |
2503.06441 | Huaming Du | Huaming Du, Lei Yuan, Qing Yang, Xingyan Chen, Yu Zhao, Han Ji, Fuzhen
Zhuang, Carl Yang, Gang Kou | Identifying Evidence Subgraphs for Financial Risk Detection via Graph
Counterfactual and Factual Reasoning | null | null | null | null | cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Company financial risks pose a significant threat to personal wealth and
national economic stability, stimulating increasing attention towards the
development of efficient andtimely methods for monitoring them. Current
approaches tend to use graph neural networks (GNNs) to model the momentum
spillover effect of risks. However, due to the black-box nature of GNNs, these
methods leave much to be improved for precise and reliable explanations towards
company risks. In this paper, we propose CF3, a novel Counterfactual and
Factual learning method for company Financial risk detection, which generates
evidence subgraphs on company knowledge graphs to reliably detect and explain
company financial risks. Specifically, we first propose a meta-path attribution
process based on Granger causality, selecting the meta-paths most relevant to
the target node labels to construct an attribution subgraph. Subsequently, we
propose anedge-type-aware graph generator to identify important edges, and we
also devise a layer-based feature masker to recognize crucial node features.
Finally, we utilize counterfactual-factual reasoning and a loss function based
on attribution subgraphs to jointly guide the learning of the graph generator
and feature masker. Extensive experiments on three real-world datasets
demonstrate the superior performance of our method compared to state-of-the-art
approaches in the field of financial risk detection.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 04:45:39 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Du",
"Huaming",
""
],
[
"Yuan",
"Lei",
""
],
[
"Yang",
"Qing",
""
],
[
"Chen",
"Xingyan",
""
],
[
"Zhao",
"Yu",
""
],
[
"Ji",
"Han",
""
],
[
"Zhuang",
"Fuzhen",
""
],
[
"Yang",
"Carl",
""
],
[
"Kou",
"Gang",
""
]
]
| TITLE: Identifying Evidence Subgraphs for Financial Risk Detection via Graph
Counterfactual and Factual Reasoning
ABSTRACT: Company financial risks pose a significant threat to personal wealth and
national economic stability, stimulating increasing attention towards the
development of efficient andtimely methods for monitoring them. Current
approaches tend to use graph neural networks (GNNs) to model the momentum
spillover effect of risks. However, due to the black-box nature of GNNs, these
methods leave much to be improved for precise and reliable explanations towards
company risks. In this paper, we propose CF3, a novel Counterfactual and
Factual learning method for company Financial risk detection, which generates
evidence subgraphs on company knowledge graphs to reliably detect and explain
company financial risks. Specifically, we first propose a meta-path attribution
process based on Granger causality, selecting the meta-paths most relevant to
the target node labels to construct an attribution subgraph. Subsequently, we
propose anedge-type-aware graph generator to identify important edges, and we
also devise a layer-based feature masker to recognize crucial node features.
Finally, we utilize counterfactual-factual reasoning and a loss function based
on attribution subgraphs to jointly guide the learning of the graph generator
and feature masker. Extensive experiments on three real-world datasets
demonstrate the superior performance of our method compared to state-of-the-art
approaches in the field of financial risk detection.
| no_new_dataset | 0.944944 |
2503.06444 | Zuqing Li | Zuqing Li, Jianzhong Qi, Junhao Gan | CtrTab: Tabular Data Synthesis with High-Dimensional and Limited Data | null | null | null | null | cs.LG cs.AI cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion-based tabular data synthesis models have yielded promising results.
However, we observe that when the data dimensionality increases, existing
models tend to degenerate and may perform even worse than simpler,
non-diffusion-based models. This is because limited training samples in
high-dimensional space often hinder generative models from capturing the
distribution accurately. To address this issue, we propose CtrTab-a condition
controlled diffusion model for tabular data synthesis-to improve the
performance of diffusion-based generative models in high-dimensional, low-data
scenarios. Through CtrTab, we inject samples with added Laplace noise as
control signals to improve data diversity and show its resemblance to L2
regularization, which enhances model robustness. Experimental results across
multiple datasets show that CtrTab outperforms state-of-the-art models, with
performance gap in accuracy over 80% on average. Our source code will be
released upon paper publication.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 05:01:56 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Zuqing",
""
],
[
"Qi",
"Jianzhong",
""
],
[
"Gan",
"Junhao",
""
]
]
| TITLE: CtrTab: Tabular Data Synthesis with High-Dimensional and Limited Data
ABSTRACT: Diffusion-based tabular data synthesis models have yielded promising results.
However, we observe that when the data dimensionality increases, existing
models tend to degenerate and may perform even worse than simpler,
non-diffusion-based models. This is because limited training samples in
high-dimensional space often hinder generative models from capturing the
distribution accurately. To address this issue, we propose CtrTab-a condition
controlled diffusion model for tabular data synthesis-to improve the
performance of diffusion-based generative models in high-dimensional, low-data
scenarios. Through CtrTab, we inject samples with added Laplace noise as
control signals to improve data diversity and show its resemblance to L2
regularization, which enhances model robustness. Experimental results across
multiple datasets show that CtrTab outperforms state-of-the-art models, with
performance gap in accuracy over 80% on average. Our source code will be
released upon paper publication.
| no_new_dataset | 0.949576 |
2503.06445 | Junlong Zhang | Shuai Pang, Junlong Zhang, and Lei Dong | Socioeconomic centers in cities worldwide | null | null | null | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Urban centers serve as engines of regional development, yet accurately
defining and identifying the socioeconomic centers of cities globally remains a
big challenge. Existing mapping efforts are often limited to large cities in
developed regions and rely on data sources that are unavailable in many
developing countries. This data scarcity hinders the establishment of
consistent urban indicators, such as accessibility, to assess progress towards
the United Nations Sustainable Development Goals (SDGs). Here, we develop and
validate a global map of the socioeconomic centers of cities for 2020 by
integrating nighttime light and population density data within an advanced
geospatial modeling framework. Our analysis reveals that monocentric cities --
the standard urban model -- still dominate our planet, accounting for over 80%
of cities worldwide. However, these monocentric cities encompass only
approximately 20% of the total urbanized area, urban population, and nighttime
light intensity; this 80/20 pattern underscores significant disparities in
urban development. Further analysis, combined with socioeconomic datasets,
reveals a marked difference between developed and developing regions:
high-income countries exhibit greater polycentricity than low-income countries,
demonstrating a positive correlation between urban sprawl and economic growth.
Our global dataset and findings provide critical insights into urban structure
and development, with important implications for urban planning, policymaking,
and the formulation of indicators for urban sustainability assessment.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 05:02:28 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Pang",
"Shuai",
""
],
[
"Zhang",
"Junlong",
""
],
[
"Dong",
"Lei",
""
]
]
| TITLE: Socioeconomic centers in cities worldwide
ABSTRACT: Urban centers serve as engines of regional development, yet accurately
defining and identifying the socioeconomic centers of cities globally remains a
big challenge. Existing mapping efforts are often limited to large cities in
developed regions and rely on data sources that are unavailable in many
developing countries. This data scarcity hinders the establishment of
consistent urban indicators, such as accessibility, to assess progress towards
the United Nations Sustainable Development Goals (SDGs). Here, we develop and
validate a global map of the socioeconomic centers of cities for 2020 by
integrating nighttime light and population density data within an advanced
geospatial modeling framework. Our analysis reveals that monocentric cities --
the standard urban model -- still dominate our planet, accounting for over 80%
of cities worldwide. However, these monocentric cities encompass only
approximately 20% of the total urbanized area, urban population, and nighttime
light intensity; this 80/20 pattern underscores significant disparities in
urban development. Further analysis, combined with socioeconomic datasets,
reveals a marked difference between developed and developing regions:
high-income countries exhibit greater polycentricity than low-income countries,
demonstrating a positive correlation between urban sprawl and economic growth.
Our global dataset and findings provide critical insights into urban structure
and development, with important implications for urban planning, policymaking,
and the formulation of indicators for urban sustainability assessment.
| no_new_dataset | 0.538377 |
2503.06461 | Hongsin Lee | Seungju Cho, Hongsin Lee, Changick Kim | Long-tailed Adversarial Training with Self-Distillation | ICLR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Adversarial training significantly enhances adversarial robustness, yet
superior performance is predominantly achieved on balanced datasets.
Addressing adversarial robustness in the context of unbalanced or long-tailed
distributions is considerably more challenging, mainly due to the scarcity of
tail data instances.
Previous research on adversarial robustness within long-tailed distributions
has primarily focused on combining traditional long-tailed natural training
with existing adversarial robustness methods.
In this study, we provide an in-depth analysis for the challenge that
adversarial training struggles to achieve high performance on tail classes in
long-tailed distributions.
Furthermore, we propose a simple yet effective solution to advance
adversarial robustness on long-tailed distributions through a novel
self-distillation technique.
Specifically, this approach leverages a balanced self-teacher model, which is
trained using a balanced dataset sampled from the original long-tailed dataset.
Our extensive experiments demonstrate state-of-the-art performance in both
clean and robust accuracy for long-tailed adversarial robustness, with
significant improvements in tail class performance on various datasets. We
improve the accuracy against PGD attacks for tail classes by 20.3, 7.1, and 3.8
percentage points on CIFAR-10, CIFAR-100, and Tiny-ImageNet, respectively,
while achieving the highest robust accuracy.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 05:39:36 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Cho",
"Seungju",
""
],
[
"Lee",
"Hongsin",
""
],
[
"Kim",
"Changick",
""
]
]
| TITLE: Long-tailed Adversarial Training with Self-Distillation
ABSTRACT: Adversarial training significantly enhances adversarial robustness, yet
superior performance is predominantly achieved on balanced datasets.
Addressing adversarial robustness in the context of unbalanced or long-tailed
distributions is considerably more challenging, mainly due to the scarcity of
tail data instances.
Previous research on adversarial robustness within long-tailed distributions
has primarily focused on combining traditional long-tailed natural training
with existing adversarial robustness methods.
In this study, we provide an in-depth analysis for the challenge that
adversarial training struggles to achieve high performance on tail classes in
long-tailed distributions.
Furthermore, we propose a simple yet effective solution to advance
adversarial robustness on long-tailed distributions through a novel
self-distillation technique.
Specifically, this approach leverages a balanced self-teacher model, which is
trained using a balanced dataset sampled from the original long-tailed dataset.
Our extensive experiments demonstrate state-of-the-art performance in both
clean and robust accuracy for long-tailed adversarial robustness, with
significant improvements in tail class performance on various datasets. We
improve the accuracy against PGD attacks for tail classes by 20.3, 7.1, and 3.8
percentage points on CIFAR-10, CIFAR-100, and Tiny-ImageNet, respectively,
while achieving the highest robust accuracy.
| no_new_dataset | 0.947769 |
2503.06467 | Qiming Xia | Shijia Zhao, Qiming Xia, Xusheng Guo, Pufan Zou, Maoji Zheng, Hai Wu,
Chenglu Wen, Cheng Wang | SP3D: Boosting Sparsely-Supervised 3D Object Detection via Accurate
Cross-Modal Semantic Prompts | 11 pages, 3 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, sparsely-supervised 3D object detection has gained great attention,
achieving performance close to fully-supervised 3D objectors while requiring
only a few annotated instances. Nevertheless, these methods suffer challenges
when accurate labels are extremely absent. In this paper, we propose a boosting
strategy, termed SP3D, explicitly utilizing the cross-modal semantic prompts
generated from Large Multimodal Models (LMMs) to boost the 3D detector with
robust feature discrimination capability under sparse annotation settings.
Specifically, we first develop a Confident Points Semantic Transfer (CPST)
module that generates accurate cross-modal semantic prompts through
boundary-constrained center cluster selection. Based on these accurate semantic
prompts, which we treat as seed points, we introduce a Dynamic Cluster
Pseudo-label Generation (DCPG) module to yield pseudo-supervision signals from
the geometry shape of multi-scale neighbor points. Additionally, we design a
Distribution Shape score (DS score) that chooses high-quality supervision
signals for the initial training of the 3D detector. Experiments on the KITTI
dataset and Waymo Open Dataset (WOD) have validated that SP3D can enhance the
performance of sparsely supervised detectors by a large margin under meager
labeling conditions. Moreover, we verified SP3D in the zero-shot setting, where
its performance exceeded that of the state-of-the-art methods. The code is
available at https://github.com/xmuqimingxia/SP3D.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:08:04 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhao",
"Shijia",
""
],
[
"Xia",
"Qiming",
""
],
[
"Guo",
"Xusheng",
""
],
[
"Zou",
"Pufan",
""
],
[
"Zheng",
"Maoji",
""
],
[
"Wu",
"Hai",
""
],
[
"Wen",
"Chenglu",
""
],
[
"Wang",
"Cheng",
""
]
]
| TITLE: SP3D: Boosting Sparsely-Supervised 3D Object Detection via Accurate
Cross-Modal Semantic Prompts
ABSTRACT: Recently, sparsely-supervised 3D object detection has gained great attention,
achieving performance close to fully-supervised 3D objectors while requiring
only a few annotated instances. Nevertheless, these methods suffer challenges
when accurate labels are extremely absent. In this paper, we propose a boosting
strategy, termed SP3D, explicitly utilizing the cross-modal semantic prompts
generated from Large Multimodal Models (LMMs) to boost the 3D detector with
robust feature discrimination capability under sparse annotation settings.
Specifically, we first develop a Confident Points Semantic Transfer (CPST)
module that generates accurate cross-modal semantic prompts through
boundary-constrained center cluster selection. Based on these accurate semantic
prompts, which we treat as seed points, we introduce a Dynamic Cluster
Pseudo-label Generation (DCPG) module to yield pseudo-supervision signals from
the geometry shape of multi-scale neighbor points. Additionally, we design a
Distribution Shape score (DS score) that chooses high-quality supervision
signals for the initial training of the 3D detector. Experiments on the KITTI
dataset and Waymo Open Dataset (WOD) have validated that SP3D can enhance the
performance of sparsely supervised detectors by a large margin under meager
labeling conditions. Moreover, we verified SP3D in the zero-shot setting, where
its performance exceeded that of the state-of-the-art methods. The code is
available at https://github.com/xmuqimingxia/SP3D.
| no_new_dataset | 0.947478 |
2503.06474 | Kong Huanjun | Huanjun Kong, Zhefan Wang, Chenyang Wang, Zhe Ma, Nanqing Dong | HuixiangDou2: A Robustly Optimized GraphRAG Approach | 11 pages | null | null | null | cs.IR cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) perform well on familiar queries but struggle
with specialized or emerging topics. Graph-based Retrieval-Augmented Generation
(GraphRAG) addresses this by structuring domain knowledge as a graph for
dynamic retrieval. However, existing pipelines involve complex engineering
workflows, making it difficult to isolate the impact of individual components.
Evaluating retrieval effectiveness is also challenging due to dataset overlap
with LLM pretraining data. In this work, we introduce HuixiangDou2, a robustly
optimized GraphRAG framework. Specifically, we leverage the effectiveness of
dual-level retrieval and optimize its performance in a 32k context for maximum
precision, and compare logic-based retrieval and dual-level retrieval to
enhance overall functionality. Our implementation includes comparative
experiments on a test set, where Qwen2.5-7B-Instruct initially underperformed.
With our approach, the score improved significantly from 60 to 74.5, as
illustrated in the Figure. Experiments on domain-specific datasets reveal that
dual-level retrieval enhances fuzzy matching, while logic-form retrieval
improves structured reasoning. Furthermore, we propose a multi-stage
verification mechanism to improve retrieval robustness without increasing
computational cost. Empirical results show significant accuracy gains over
baselines, highlighting the importance of adaptive retrieval. To support
research and adoption, we release HuixiangDou2 as an open-source resource
https://github.com/tpoisonooo/huixiangdou2.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:20:24 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Kong",
"Huanjun",
""
],
[
"Wang",
"Zhefan",
""
],
[
"Wang",
"Chenyang",
""
],
[
"Ma",
"Zhe",
""
],
[
"Dong",
"Nanqing",
""
]
]
| TITLE: HuixiangDou2: A Robustly Optimized GraphRAG Approach
ABSTRACT: Large Language Models (LLMs) perform well on familiar queries but struggle
with specialized or emerging topics. Graph-based Retrieval-Augmented Generation
(GraphRAG) addresses this by structuring domain knowledge as a graph for
dynamic retrieval. However, existing pipelines involve complex engineering
workflows, making it difficult to isolate the impact of individual components.
Evaluating retrieval effectiveness is also challenging due to dataset overlap
with LLM pretraining data. In this work, we introduce HuixiangDou2, a robustly
optimized GraphRAG framework. Specifically, we leverage the effectiveness of
dual-level retrieval and optimize its performance in a 32k context for maximum
precision, and compare logic-based retrieval and dual-level retrieval to
enhance overall functionality. Our implementation includes comparative
experiments on a test set, where Qwen2.5-7B-Instruct initially underperformed.
With our approach, the score improved significantly from 60 to 74.5, as
illustrated in the Figure. Experiments on domain-specific datasets reveal that
dual-level retrieval enhances fuzzy matching, while logic-form retrieval
improves structured reasoning. Furthermore, we propose a multi-stage
verification mechanism to improve retrieval robustness without increasing
computational cost. Empirical results show significant accuracy gains over
baselines, highlighting the importance of adaptive retrieval. To support
research and adoption, we release HuixiangDou2 as an open-source resource
https://github.com/tpoisonooo/huixiangdou2.
| no_new_dataset | 0.944638 |
2503.06477 | Chuheng Wei | Chuheng Wei, Ziye Qin, Siyan Li, Ziyan Zhang, Xuanpeng Zhao, Amr
Abdelraouf, Rohit Gupta, Kyungtae Han, Matthew J. Barth, Guoyuan Wu | PDB: Not All Drivers Are the Same -- A Personalized Dataset for
Understanding Driving Behavior | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Driving behavior is inherently personal, influenced by individual habits,
decision-making styles, and physiological states. However, most existing
datasets treat all drivers as homogeneous, overlooking driver-specific
variability. To address this gap, we introduce the Personalized Driving
Behavior (PDB) dataset, a multi-modal dataset designed to capture
personalization in driving behavior under naturalistic driving conditions.
Unlike conventional datasets, PDB minimizes external influences by maintaining
consistent routes, vehicles, and lighting conditions across sessions. It
includes sources from 128-line LiDAR, front-facing camera video, GNSS, 9-axis
IMU, CAN bus data (throttle, brake, steering angle), and driver-specific
signals such as facial video and heart rate. The dataset features 12
participants, approximately 270,000 LiDAR frames, 1.6 million images, and 6.6
TB of raw sensor data. The processed trajectory dataset consists of 1,669
segments, each spanning 10 seconds with a 0.2-second interval. By explicitly
capturing drivers' behavior, PDB serves as a unique resource for human factor
analysis, driver identification, and personalized mobility applications,
contributing to the development of human-centric intelligent transportation
systems.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:28:39 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wei",
"Chuheng",
""
],
[
"Qin",
"Ziye",
""
],
[
"Li",
"Siyan",
""
],
[
"Zhang",
"Ziyan",
""
],
[
"Zhao",
"Xuanpeng",
""
],
[
"Abdelraouf",
"Amr",
""
],
[
"Gupta",
"Rohit",
""
],
[
"Han",
"Kyungtae",
""
],
[
"Barth",
"Matthew J.",
""
],
[
"Wu",
"Guoyuan",
""
]
]
| TITLE: PDB: Not All Drivers Are the Same -- A Personalized Dataset for
Understanding Driving Behavior
ABSTRACT: Driving behavior is inherently personal, influenced by individual habits,
decision-making styles, and physiological states. However, most existing
datasets treat all drivers as homogeneous, overlooking driver-specific
variability. To address this gap, we introduce the Personalized Driving
Behavior (PDB) dataset, a multi-modal dataset designed to capture
personalization in driving behavior under naturalistic driving conditions.
Unlike conventional datasets, PDB minimizes external influences by maintaining
consistent routes, vehicles, and lighting conditions across sessions. It
includes sources from 128-line LiDAR, front-facing camera video, GNSS, 9-axis
IMU, CAN bus data (throttle, brake, steering angle), and driver-specific
signals such as facial video and heart rate. The dataset features 12
participants, approximately 270,000 LiDAR frames, 1.6 million images, and 6.6
TB of raw sensor data. The processed trajectory dataset consists of 1,669
segments, each spanning 10 seconds with a 0.2-second interval. By explicitly
capturing drivers' behavior, PDB serves as a unique resource for human factor
analysis, driver identification, and personalized mobility applications,
contributing to the development of human-centric intelligent transportation
systems.
| new_dataset | 0.965867 |
2503.06479 | Ali Sarabadani | Ali Sarabadani (1), Kheirolah Rahsepar Fard (2), and Hamid Dalvand (3)
((1) Department of Computer Engineering and Information Technology,
University of Qom, Iran, (2) Department of Computer Engineering and
Information Technology, University of Qom, Iran, (3) Department of
Occupational Therapy, School of Rehabilitation, Tehran University of Medical
Sciences, Iran) | ExKG-LLM: Leveraging Large Language Models for Automated Expansion of
Cognitive Neuroscience Knowledge Graphs | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | The paper introduces ExKG-LLM, a framework designed to automate the expansion
of cognitive neuroscience knowledge graphs (CNKG) using large language models
(LLMs). It addresses limitations in existing tools by enhancing accuracy,
completeness, and usefulness in CNKG. The framework leverages a large dataset
of scientific papers and clinical reports, applying state-of-the-art LLMs to
extract, optimize, and integrate new entities and relationships. Evaluation
metrics include precision, recall, and graph density. Results show significant
improvements: precision (0.80, +6.67%), recall (0.81, +15.71%), F1 score
(0.805, +11.81%), and increased edge nodes (21.13% and 31.92%). Graph density
slightly decreased, reflecting a broader but more fragmented structure.
Engagement rates rose by 20%, while CNKG diameter increased to 15, indicating a
more distributed structure. Time complexity improved to O(n log n), but space
complexity rose to O(n2), indicating higher memory usage. ExKG-LLM demonstrates
potential for enhancing knowledge generation, semantic search, and clinical
decision-making in cognitive neuroscience, adaptable to broader scientific
fields.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:32:56 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Sarabadani",
"Ali",
""
],
[
"Fard",
"Kheirolah Rahsepar",
""
],
[
"Dalvand",
"Hamid",
""
]
]
| TITLE: ExKG-LLM: Leveraging Large Language Models for Automated Expansion of
Cognitive Neuroscience Knowledge Graphs
ABSTRACT: The paper introduces ExKG-LLM, a framework designed to automate the expansion
of cognitive neuroscience knowledge graphs (CNKG) using large language models
(LLMs). It addresses limitations in existing tools by enhancing accuracy,
completeness, and usefulness in CNKG. The framework leverages a large dataset
of scientific papers and clinical reports, applying state-of-the-art LLMs to
extract, optimize, and integrate new entities and relationships. Evaluation
metrics include precision, recall, and graph density. Results show significant
improvements: precision (0.80, +6.67%), recall (0.81, +15.71%), F1 score
(0.805, +11.81%), and increased edge nodes (21.13% and 31.92%). Graph density
slightly decreased, reflecting a broader but more fragmented structure.
Engagement rates rose by 20%, while CNKG diameter increased to 15, indicating a
more distributed structure. Time complexity improved to O(n log n), but space
complexity rose to O(n2), indicating higher memory usage. ExKG-LLM demonstrates
potential for enhancing knowledge generation, semantic search, and clinical
decision-making in cognitive neuroscience, adaptable to broader scientific
fields.
| no_new_dataset | 0.952926 |
2503.06482 | Honglin Li | Honglin Li, Zhongyi Shui, Yunlong Zhang, Chenglu Zhu, Lin Yang | PathVQ: Reforming Computational Pathology Foundation Model for Whole
Slide Image Analysis via Vector Quantization | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computational pathology and whole-slide image (WSI) analysis are pivotal in
cancer diagnosis and prognosis. However, the ultra-high resolution of WSIs
presents significant modeling challenges. Recent advancements in pathology
foundation models have improved performance, yet most approaches rely on [CLS]
token representation of tile ViT as slide-level inputs (16x16 pixels is
refereed as patch and 224x224 pixels as tile). This discards critical spatial
details from patch tokens, limiting downstream WSI analysis tasks. We find that
leveraging all spatial patch tokens benefits WSI analysis but incurs nearly
200x higher storage and training costs (e.g., 196 tokens in ViT$_{224}$). To
address this, we introduce vector quantized (VQ) distillation on patch feature,
which efficiently compresses spatial patch tokens using discrete indices and a
decoder. Our method reduces token dimensionality from 1024 to 16, achieving a
64x compression rate while preserving reconstruction fidelity. Furthermore, we
employ a multi-scale VQ (MSVQ) strategy, which not only enhances VQ
reconstruction performance but also serves as a Self-supervised Learning (SSL)
supervision for a seamless slide-level pretraining objective. Built upon the
quantized patch features and supervision targets of tile via MSVQ, we develop a
progressive convolutional module and slide-level SSL to extract representations
with rich spatial-information for downstream WSI tasks. Extensive evaluations
on multiple datasets demonstrate the effectiveness of our approach, achieving
state-of-the-art performance in WSI analysis. Code will be available soon.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:51:08 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Honglin",
""
],
[
"Shui",
"Zhongyi",
""
],
[
"Zhang",
"Yunlong",
""
],
[
"Zhu",
"Chenglu",
""
],
[
"Yang",
"Lin",
""
]
]
| TITLE: PathVQ: Reforming Computational Pathology Foundation Model for Whole
Slide Image Analysis via Vector Quantization
ABSTRACT: Computational pathology and whole-slide image (WSI) analysis are pivotal in
cancer diagnosis and prognosis. However, the ultra-high resolution of WSIs
presents significant modeling challenges. Recent advancements in pathology
foundation models have improved performance, yet most approaches rely on [CLS]
token representation of tile ViT as slide-level inputs (16x16 pixels is
refereed as patch and 224x224 pixels as tile). This discards critical spatial
details from patch tokens, limiting downstream WSI analysis tasks. We find that
leveraging all spatial patch tokens benefits WSI analysis but incurs nearly
200x higher storage and training costs (e.g., 196 tokens in ViT$_{224}$). To
address this, we introduce vector quantized (VQ) distillation on patch feature,
which efficiently compresses spatial patch tokens using discrete indices and a
decoder. Our method reduces token dimensionality from 1024 to 16, achieving a
64x compression rate while preserving reconstruction fidelity. Furthermore, we
employ a multi-scale VQ (MSVQ) strategy, which not only enhances VQ
reconstruction performance but also serves as a Self-supervised Learning (SSL)
supervision for a seamless slide-level pretraining objective. Built upon the
quantized patch features and supervision targets of tile via MSVQ, we develop a
progressive convolutional module and slide-level SSL to extract representations
with rich spatial-information for downstream WSI tasks. Extensive evaluations
on multiple datasets demonstrate the effectiveness of our approach, achieving
state-of-the-art performance in WSI analysis. Code will be available soon.
| no_new_dataset | 0.947381 |
2503.06484 | Xiao Wang | Xiao Wang, Yuehang Li, Fuling Wang, Bo Jiang, Yaowei Wang, Yonghong
Tian, Jin Tang, Bin Luo | Sign Language Translation using Frame and Event Stream: Benchmark
Dataset and Algorithms | In Peer Review | null | null | null | cs.CV cs.AI cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate sign language understanding serves as a crucial communication
channel for individuals with disabilities. Current sign language translation
algorithms predominantly rely on RGB frames, which may be limited by fixed
frame rates, variable lighting conditions, and motion blur caused by rapid hand
movements. Inspired by the recent successful application of event cameras in
other fields, we propose to leverage event streams to assist RGB cameras in
capturing gesture data, addressing the various challenges mentioned above.
Specifically, we first collect a large-scale RGB-Event sign language
translation dataset using the DVS346 camera, termed VECSL, which contains
15,676 RGB-Event samples, 15,191 glosses, and covers 2,568 Chinese characters.
These samples were gathered across a diverse range of indoor and outdoor
environments, capturing multiple viewing angles, varying light intensities, and
different camera motions. Due to the absence of benchmark algorithms for
comparison in this new task, we retrained and evaluated multiple
state-of-the-art SLT algorithms, and believe that this benchmark can
effectively support subsequent related research. Additionally, we propose a
novel RGB-Event sign language translation framework (i.e., M$^2$-SLT) that
incorporates fine-grained micro-sign and coarse-grained macro-sign retrieval,
achieving state-of-the-art results on the proposed dataset. Both the source
code and dataset will be released on https://github.com/Event-AHU/OpenESL.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 06:55:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Xiao",
""
],
[
"Li",
"Yuehang",
""
],
[
"Wang",
"Fuling",
""
],
[
"Jiang",
"Bo",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Tian",
"Yonghong",
""
],
[
"Tang",
"Jin",
""
],
[
"Luo",
"Bin",
""
]
]
| TITLE: Sign Language Translation using Frame and Event Stream: Benchmark
Dataset and Algorithms
ABSTRACT: Accurate sign language understanding serves as a crucial communication
channel for individuals with disabilities. Current sign language translation
algorithms predominantly rely on RGB frames, which may be limited by fixed
frame rates, variable lighting conditions, and motion blur caused by rapid hand
movements. Inspired by the recent successful application of event cameras in
other fields, we propose to leverage event streams to assist RGB cameras in
capturing gesture data, addressing the various challenges mentioned above.
Specifically, we first collect a large-scale RGB-Event sign language
translation dataset using the DVS346 camera, termed VECSL, which contains
15,676 RGB-Event samples, 15,191 glosses, and covers 2,568 Chinese characters.
These samples were gathered across a diverse range of indoor and outdoor
environments, capturing multiple viewing angles, varying light intensities, and
different camera motions. Due to the absence of benchmark algorithms for
comparison in this new task, we retrained and evaluated multiple
state-of-the-art SLT algorithms, and believe that this benchmark can
effectively support subsequent related research. Additionally, we propose a
novel RGB-Event sign language translation framework (i.e., M$^2$-SLT) that
incorporates fine-grained micro-sign and coarse-grained macro-sign retrieval,
achieving state-of-the-art results on the proposed dataset. Both the source
code and dataset will be released on https://github.com/Event-AHU/OpenESL.
| new_dataset | 0.959269 |
2503.06487 | Rina Mishra | Rina Mishra and Gaurav Varshney | A Study of Effectiveness of Brand Domain Identification Features for
Phishing Detection in 2025 | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Phishing websites continue to pose a significant security challenge, making
the development of robust detection mechanisms essential. Brand Domain
Identification (BDI) serves as a crucial step in many phishing detection
approaches. This study systematically evaluates the effectiveness of features
employed over the past decade for BDI, focusing on their weighted importance in
phishing detection as of 2025. The primary objective is to determine whether
the identified brand domain matches the claimed domain, utilizing popular
features for phishing detection. To validate feature importance and evaluate
performance, we conducted two experiments on a dataset comprising 4,667
legitimate sites and 4,561 phishing sites. In Experiment 1, we used the Weka
tool to identify optimized and important feature sets out of 5: CN
Information(CN), Logo Domain(LD),Form Action Domain(FAD),Most Common Link in
Domain(MCLD) and Cookie Domain through its 4 Attribute Ranking Evaluator. The
results revealed that none of the features were redundant, and Random Forest
emerged as the best classifier, achieving an impressive accuracy of 99.7\% with
an average response time of 0.08 seconds. In Experiment 2, we trained five
machine learning models, including Random Forest, Decision Tree, Support Vector
Machine, Multilayer Perceptron, and XGBoost to assess the performance of
individual BDI features and their combinations. The results demonstrated an
accuracy of 99.8\%, achieved with feature combinations of only three features:
Most Common Link Domain, Logo Domain, Form Action and Most Common Link
Domain,CN Info,Logo Domain using Random Forest as the best classifier. This
study underscores the importance of leveraging key domain features for
efficient phishing detection and paves the way for the development of
real-time, scalable detection systems.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 07:14:04 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Mishra",
"Rina",
""
],
[
"Varshney",
"Gaurav",
""
]
]
| TITLE: A Study of Effectiveness of Brand Domain Identification Features for
Phishing Detection in 2025
ABSTRACT: Phishing websites continue to pose a significant security challenge, making
the development of robust detection mechanisms essential. Brand Domain
Identification (BDI) serves as a crucial step in many phishing detection
approaches. This study systematically evaluates the effectiveness of features
employed over the past decade for BDI, focusing on their weighted importance in
phishing detection as of 2025. The primary objective is to determine whether
the identified brand domain matches the claimed domain, utilizing popular
features for phishing detection. To validate feature importance and evaluate
performance, we conducted two experiments on a dataset comprising 4,667
legitimate sites and 4,561 phishing sites. In Experiment 1, we used the Weka
tool to identify optimized and important feature sets out of 5: CN
Information(CN), Logo Domain(LD),Form Action Domain(FAD),Most Common Link in
Domain(MCLD) and Cookie Domain through its 4 Attribute Ranking Evaluator. The
results revealed that none of the features were redundant, and Random Forest
emerged as the best classifier, achieving an impressive accuracy of 99.7\% with
an average response time of 0.08 seconds. In Experiment 2, we trained five
machine learning models, including Random Forest, Decision Tree, Support Vector
Machine, Multilayer Perceptron, and XGBoost to assess the performance of
individual BDI features and their combinations. The results demonstrated an
accuracy of 99.8\%, achieved with feature combinations of only three features:
Most Common Link Domain, Logo Domain, Form Action and Most Common Link
Domain,CN Info,Logo Domain using Random Forest as the best classifier. This
study underscores the importance of leveraging key domain features for
efficient phishing detection and paves the way for the development of
real-time, scalable detection systems.
| no_new_dataset | 0.950365 |
2503.06488 | Shijun Cheng | Shijun Cheng, Mohammad H. Taufik, and Tariq Alkhalifah | Seismic wavefield solutions via physics-guided generative neural
operator | null | null | null | null | physics.geo-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current neural operators often struggle to generalize to complex,
out-of-distribution conditions, limiting their ability in seismic wavefield
representation. To address this, we propose a generative neural operator (GNO)
that leverages generative diffusion models (GDMs) to learn the underlying
statistical distribution of scattered wavefields while incorporating a
physics-guided sampling process at each inference step. This physics guidance
enforces wave equation-based constraints corresponding to specific velocity
models, driving the iteratively generated wavefields toward physically
consistent solutions. By training the diffusion model on wavefields
corresponding to a diverse dataset of velocity models, frequencies, and source
positions, our GNO enables to rapidly synthesize high-fidelity wavefields at
inference time. Numerical experiments demonstrate that our GNO not only
produces accurate wavefields matching numerical reference solutions, but also
generalizes effectively to previously unseen velocity models and frequencies.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 07:19:12 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Cheng",
"Shijun",
""
],
[
"Taufik",
"Mohammad H.",
""
],
[
"Alkhalifah",
"Tariq",
""
]
]
| TITLE: Seismic wavefield solutions via physics-guided generative neural
operator
ABSTRACT: Current neural operators often struggle to generalize to complex,
out-of-distribution conditions, limiting their ability in seismic wavefield
representation. To address this, we propose a generative neural operator (GNO)
that leverages generative diffusion models (GDMs) to learn the underlying
statistical distribution of scattered wavefields while incorporating a
physics-guided sampling process at each inference step. This physics guidance
enforces wave equation-based constraints corresponding to specific velocity
models, driving the iteratively generated wavefields toward physically
consistent solutions. By training the diffusion model on wavefields
corresponding to a diverse dataset of velocity models, frequencies, and source
positions, our GNO enables to rapidly synthesize high-fidelity wavefields at
inference time. Numerical experiments demonstrate that our GNO not only
produces accurate wavefields matching numerical reference solutions, but also
generalizes effectively to previously unseen velocity models and frequencies.
| no_new_dataset | 0.955899 |
2503.06492 | Yanling Wang | Yanling Wang, Yihan Zhao, Xiaodong Chen, Shasha Guo, Lixin Liu,
Haoyang Li, Yong Xiao, Jing Zhang, Qi Li, Ke Xu | VisualSimpleQA: A Benchmark for Decoupled Evaluation of Large
Vision-Language Models in Fact-Seeking Question Answering | null | null | null | null | cs.CL cs.CV | http://creativecommons.org/licenses/by/4.0/ | Large vision-language models (LVLMs) have demonstrated remarkable
achievements, yet the generation of non-factual responses remains prevalent in
fact-seeking question answering (QA). Current multimodal fact-seeking
benchmarks primarily focus on comparing model outputs to ground truth answers,
providing limited insights into the performance of modality-specific modules.
To bridge this gap, we introduce VisualSimpleQA, a multimodal fact-seeking
benchmark with two key features. First, it enables streamlined and decoupled
evaluation of LVLMs in visual and linguistic modalities. Second, it
incorporates well-defined difficulty criteria to guide human annotation and
facilitates the extraction of a challenging subset, VisualSimpleQA-hard.
Experiments on 15 LVLMs show that even state-of-the-art models such as GPT-4o
achieve merely 60%+ correctness in multimodal fact-seeking QA on VisualSimpleQA
and 30%+ on VisualSimpleQA-hard. Furthermore, the decoupled evaluation across
these models highlights substantial opportunities for improvement in both
visual and linguistic modules. The dataset is available at
https://huggingface.co/datasets/WYLing/VisualSimpleQA.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 07:25:32 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Yanling",
""
],
[
"Zhao",
"Yihan",
""
],
[
"Chen",
"Xiaodong",
""
],
[
"Guo",
"Shasha",
""
],
[
"Liu",
"Lixin",
""
],
[
"Li",
"Haoyang",
""
],
[
"Xiao",
"Yong",
""
],
[
"Zhang",
"Jing",
""
],
[
"Li",
"Qi",
""
],
[
"Xu",
"Ke",
""
]
]
| TITLE: VisualSimpleQA: A Benchmark for Decoupled Evaluation of Large
Vision-Language Models in Fact-Seeking Question Answering
ABSTRACT: Large vision-language models (LVLMs) have demonstrated remarkable
achievements, yet the generation of non-factual responses remains prevalent in
fact-seeking question answering (QA). Current multimodal fact-seeking
benchmarks primarily focus on comparing model outputs to ground truth answers,
providing limited insights into the performance of modality-specific modules.
To bridge this gap, we introduce VisualSimpleQA, a multimodal fact-seeking
benchmark with two key features. First, it enables streamlined and decoupled
evaluation of LVLMs in visual and linguistic modalities. Second, it
incorporates well-defined difficulty criteria to guide human annotation and
facilitates the extraction of a challenging subset, VisualSimpleQA-hard.
Experiments on 15 LVLMs show that even state-of-the-art models such as GPT-4o
achieve merely 60%+ correctness in multimodal fact-seeking QA on VisualSimpleQA
and 30%+ on VisualSimpleQA-hard. Furthermore, the decoupled evaluation across
these models highlights substantial opportunities for improvement in both
visual and linguistic modules. The dataset is available at
https://huggingface.co/datasets/WYLing/VisualSimpleQA.
| new_dataset | 0.950549 |
2503.06497 | Enming Zhang | Enming Zhang, Peizhe Gong, Xingyuan Dai, Yisheng Lv, Qinghai Miao | Evaluation of Safety Cognition Capability in Vision-Language Models for
Autonomous Driving | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Assessing the safety of vision-language models (VLMs) in autonomous driving
is particularly important; however, existing work mainly focuses on traditional
benchmark evaluations. As interactive components within autonomous driving
systems, VLMs must maintain strong safety cognition during interactions. From
this perspective, we propose a novel evaluation method: Safety Cognitive
Driving Benchmark (SCD-Bench) . To address the large-scale annotation challenge
for SCD-Bench, we develop the Autonomous Driving Image-Text Annotation System
(ADA) . Additionally, to ensure data quality in SCD-Bench, our dataset
undergoes manual refinement by experts with professional knowledge in
autonomous driving. We further develop an automated evaluation method based on
large language models (LLMs). To verify its effectiveness, we compare its
evaluation results with those of expert human evaluations, achieving a
consistency rate of 99.74%. Preliminary experimental results indicate that
existing open-source models still lack sufficient safety cognition, showing a
significant gap compared to GPT-4o. Notably, lightweight models (1B-4B)
demonstrate minimal safety cognition. However, since lightweight models are
crucial for autonomous driving systems, this presents a significant challenge
for integrating VLMs into the field.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 07:53:19 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhang",
"Enming",
""
],
[
"Gong",
"Peizhe",
""
],
[
"Dai",
"Xingyuan",
""
],
[
"Lv",
"Yisheng",
""
],
[
"Miao",
"Qinghai",
""
]
]
| TITLE: Evaluation of Safety Cognition Capability in Vision-Language Models for
Autonomous Driving
ABSTRACT: Assessing the safety of vision-language models (VLMs) in autonomous driving
is particularly important; however, existing work mainly focuses on traditional
benchmark evaluations. As interactive components within autonomous driving
systems, VLMs must maintain strong safety cognition during interactions. From
this perspective, we propose a novel evaluation method: Safety Cognitive
Driving Benchmark (SCD-Bench) . To address the large-scale annotation challenge
for SCD-Bench, we develop the Autonomous Driving Image-Text Annotation System
(ADA) . Additionally, to ensure data quality in SCD-Bench, our dataset
undergoes manual refinement by experts with professional knowledge in
autonomous driving. We further develop an automated evaluation method based on
large language models (LLMs). To verify its effectiveness, we compare its
evaluation results with those of expert human evaluations, achieving a
consistency rate of 99.74%. Preliminary experimental results indicate that
existing open-source models still lack sufficient safety cognition, showing a
significant gap compared to GPT-4o. Notably, lightweight models (1B-4B)
demonstrate minimal safety cognition. However, since lightweight models are
crucial for autonomous driving systems, this presents a significant challenge
for integrating VLMs into the field.
| new_dataset | 0.608361 |
2503.06500 | Yanwei Huang | Yanwei Huang, Yan Miao, Di Weng, Adam Perer, Yingcai Wu | StructVizor: Interactive Profiling of Semi-Structured Textual Data | Accepted for CHI 2025 | null | 10.1145/3706598.3713484 | null | cs.HC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Data profiling plays a critical role in understanding the structure of
complex datasets and supporting numerous downstream tasks, such as social media
analytics and financial fraud detection. While existing research predominantly
focuses on structured data formats, a substantial portion of semi-structured
textual data still requires ad-hoc and arduous manual profiling to extract and
comprehend its internal structures. In this work, we propose StructVizor, an
interactive profiling system that facilitates sensemaking and transformation of
semi-structured textual data. Our tool mainly addresses two challenges: a)
extracting and visualizing the diverse structural patterns within data, such as
how information is organized or related, and b) enabling users to efficiently
perform various wrangling operations on textual data. Through automatic data
parsing and structure mining, StructVizor enables visual analytics of
structural patterns, while incorporating novel interactions to enable
profile-based data wrangling. A comparative user study involving 12
participants demonstrates the system's usability and its effectiveness in
supporting exploratory data analysis and transformation tasks.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:03:32 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Huang",
"Yanwei",
""
],
[
"Miao",
"Yan",
""
],
[
"Weng",
"Di",
""
],
[
"Perer",
"Adam",
""
],
[
"Wu",
"Yingcai",
""
]
]
| TITLE: StructVizor: Interactive Profiling of Semi-Structured Textual Data
ABSTRACT: Data profiling plays a critical role in understanding the structure of
complex datasets and supporting numerous downstream tasks, such as social media
analytics and financial fraud detection. While existing research predominantly
focuses on structured data formats, a substantial portion of semi-structured
textual data still requires ad-hoc and arduous manual profiling to extract and
comprehend its internal structures. In this work, we propose StructVizor, an
interactive profiling system that facilitates sensemaking and transformation of
semi-structured textual data. Our tool mainly addresses two challenges: a)
extracting and visualizing the diverse structural patterns within data, such as
how information is organized or related, and b) enabling users to efficiently
perform various wrangling operations on textual data. Through automatic data
parsing and structure mining, StructVizor enables visual analytics of
structural patterns, while incorporating novel interactions to enable
profile-based data wrangling. A comparative user study involving 12
participants demonstrates the system's usability and its effectiveness in
supporting exploratory data analysis and transformation tasks.
| no_new_dataset | 0.948346 |
2503.06501 | Huaqi Tao | Huaqi Tao, Bingxi Liu, Calvin Chen, Tingjun Huang, He Li, Jinqiang
Cui, Hong Zhang | TextInPlace: Indoor Visual Place Recognition in Repetitive Structures
with Scene Text Spotting and Verification | 8 pages,5 figures | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Visual Place Recognition (VPR) is a crucial capability for long-term
autonomous robots, enabling them to identify previously visited locations using
visual information. However, existing methods remain limited in indoor settings
due to the highly repetitive structures inherent in such environments. We
observe that scene text typically appears in indoor spaces, serving to
distinguish visually similar but different places. This inspires us to propose
TextInPlace, a simple yet effective VPR framework that integrates Scene Text
Spotting (STS) to mitigate visual perceptual ambiguity in repetitive indoor
environments. Specifically, TextInPlace adopts a dual-branch architecture
within a local parameter sharing network. The VPR branch employs
attention-based aggregation to extract global descriptors for coarse-grained
retrieval, while the STS branch utilizes a bridging text spotter to detect and
recognize scene text. Finally, the discriminative text is filtered to compute
text similarity and re-rank the top-K retrieved images. To bridge the gap
between current text-based repetitive indoor scene datasets and the typical
scenarios encountered in robot navigation, we establish an indoor VPR benchmark
dataset, called Maze-with-Text. Extensive experiments on both custom and public
datasets demonstrate that TextInPlace achieves superior performance over
existing methods that rely solely on appearance information. The dataset, code,
and trained models are publicly available at
https://github.com/HqiTao/TextInPlace.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:03:41 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Tao",
"Huaqi",
""
],
[
"Liu",
"Bingxi",
""
],
[
"Chen",
"Calvin",
""
],
[
"Huang",
"Tingjun",
""
],
[
"Li",
"He",
""
],
[
"Cui",
"Jinqiang",
""
],
[
"Zhang",
"Hong",
""
]
]
| TITLE: TextInPlace: Indoor Visual Place Recognition in Repetitive Structures
with Scene Text Spotting and Verification
ABSTRACT: Visual Place Recognition (VPR) is a crucial capability for long-term
autonomous robots, enabling them to identify previously visited locations using
visual information. However, existing methods remain limited in indoor settings
due to the highly repetitive structures inherent in such environments. We
observe that scene text typically appears in indoor spaces, serving to
distinguish visually similar but different places. This inspires us to propose
TextInPlace, a simple yet effective VPR framework that integrates Scene Text
Spotting (STS) to mitigate visual perceptual ambiguity in repetitive indoor
environments. Specifically, TextInPlace adopts a dual-branch architecture
within a local parameter sharing network. The VPR branch employs
attention-based aggregation to extract global descriptors for coarse-grained
retrieval, while the STS branch utilizes a bridging text spotter to detect and
recognize scene text. Finally, the discriminative text is filtered to compute
text similarity and re-rank the top-K retrieved images. To bridge the gap
between current text-based repetitive indoor scene datasets and the typical
scenarios encountered in robot navigation, we establish an indoor VPR benchmark
dataset, called Maze-with-Text. Extensive experiments on both custom and public
datasets demonstrate that TextInPlace achieves superior performance over
existing methods that rely solely on appearance information. The dataset, code,
and trained models are publicly available at
https://github.com/HqiTao/TextInPlace.
| new_dataset | 0.973418 |
2503.06505 | Xirui Hu | Xirui Hu, Jiahao Wang, Hao Chen, Weizhan Zhang, Benqi Wang, Yikun Li,
Haishun Nan | DynamicID: Zero-Shot Multi-ID Image Personalization with Flexible Facial
Editability | 17 pages, 16 figures | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in text-to-image generation have spurred interest in
personalized human image generation, which aims to create novel images
featuring specific human identities as reference images indicate. Although
existing methods achieve high-fidelity identity preservation, they often
struggle with limited multi-ID usability and inadequate facial editability. We
present DynamicID, a tuning-free framework supported by a dual-stage training
paradigm that inherently facilitates both single-ID and multi-ID personalized
generation with high fidelity and flexible facial editability. Our key
innovations include: 1) Semantic-Activated Attention (SAA), which employs
query-level activation gating to minimize disruption to the original model when
injecting ID features and achieve multi-ID personalization without requiring
multi-ID samples during training. 2) Identity-Motion Reconfigurator (IMR),
which leverages contrastive learning to effectively disentangle and re-entangle
facial motion and identity features, thereby enabling flexible facial editing.
Additionally, we have developed a curated VariFace-10k facial dataset,
comprising 10k unique individuals, each represented by 35 distinct facial
images. Experimental results demonstrate that DynamicID outperforms
state-of-the-art methods in identity fidelity, facial editability, and multi-ID
personalization capability.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:16:19 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Hu",
"Xirui",
""
],
[
"Wang",
"Jiahao",
""
],
[
"Chen",
"Hao",
""
],
[
"Zhang",
"Weizhan",
""
],
[
"Wang",
"Benqi",
""
],
[
"Li",
"Yikun",
""
],
[
"Nan",
"Haishun",
""
]
]
| TITLE: DynamicID: Zero-Shot Multi-ID Image Personalization with Flexible Facial
Editability
ABSTRACT: Recent advancements in text-to-image generation have spurred interest in
personalized human image generation, which aims to create novel images
featuring specific human identities as reference images indicate. Although
existing methods achieve high-fidelity identity preservation, they often
struggle with limited multi-ID usability and inadequate facial editability. We
present DynamicID, a tuning-free framework supported by a dual-stage training
paradigm that inherently facilitates both single-ID and multi-ID personalized
generation with high fidelity and flexible facial editability. Our key
innovations include: 1) Semantic-Activated Attention (SAA), which employs
query-level activation gating to minimize disruption to the original model when
injecting ID features and achieve multi-ID personalization without requiring
multi-ID samples during training. 2) Identity-Motion Reconfigurator (IMR),
which leverages contrastive learning to effectively disentangle and re-entangle
facial motion and identity features, thereby enabling flexible facial editing.
Additionally, we have developed a curated VariFace-10k facial dataset,
comprising 10k unique individuals, each represented by 35 distinct facial
images. Experimental results demonstrate that DynamicID outperforms
state-of-the-art methods in identity fidelity, facial editability, and multi-ID
personalization capability.
| new_dataset | 0.950503 |
2503.06511 | Bohan Lin | Yiting Zheng, Bohan Lin, Jinqian Chen, Jihua Zhu | HFedCKD: Toward Robust Heterogeneous Federated Learning via Data-free
Knowledge Distillation and Two-way Contrast | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most current federated learning frameworks are modeled as static processes,
ignoring the dynamic characteristics of the learning system. Under the limited
communication budget of the central server, the flexible model architecture of
a large number of clients participating in knowledge transfer requires a lower
participation rate, active clients have uneven contributions, and the client
scale seriously hinders the performance of FL. We consider a more general and
practical federation scenario and propose a system heterogeneous federation
method based on data-free knowledge distillation and two-way contrast
(HFedCKD). We apply the Inverse Probability Weighted Distillation (IPWD)
strategy to the data-free knowledge transfer framework. The generator completes
the data features of the nonparticipating clients. IPWD implements a dynamic
evaluation of the prediction contribution of each client under different data
distributions. Based on the antibiased weighting of its prediction loss, the
weight distribution of each client is effectively adjusted to fairly integrate
the knowledge of participating clients. At the same time, the local model is
split into a feature extractor and a classifier. Through differential contrast
learning, the feature extractor is aligned with the global model in the feature
space, while the classifier maintains personalized decision-making
capabilities. HFedCKD effectively alleviates the knowledge offset caused by a
low participation rate under data-free knowledge distillation and improves the
performance and stability of the model. We conduct extensive experiments on
image and IoT datasets to comprehensively evaluate and verify the
generalization and robustness of the proposed HFedCKD framework.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:32:57 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zheng",
"Yiting",
""
],
[
"Lin",
"Bohan",
""
],
[
"Chen",
"Jinqian",
""
],
[
"Zhu",
"Jihua",
""
]
]
| TITLE: HFedCKD: Toward Robust Heterogeneous Federated Learning via Data-free
Knowledge Distillation and Two-way Contrast
ABSTRACT: Most current federated learning frameworks are modeled as static processes,
ignoring the dynamic characteristics of the learning system. Under the limited
communication budget of the central server, the flexible model architecture of
a large number of clients participating in knowledge transfer requires a lower
participation rate, active clients have uneven contributions, and the client
scale seriously hinders the performance of FL. We consider a more general and
practical federation scenario and propose a system heterogeneous federation
method based on data-free knowledge distillation and two-way contrast
(HFedCKD). We apply the Inverse Probability Weighted Distillation (IPWD)
strategy to the data-free knowledge transfer framework. The generator completes
the data features of the nonparticipating clients. IPWD implements a dynamic
evaluation of the prediction contribution of each client under different data
distributions. Based on the antibiased weighting of its prediction loss, the
weight distribution of each client is effectively adjusted to fairly integrate
the knowledge of participating clients. At the same time, the local model is
split into a feature extractor and a classifier. Through differential contrast
learning, the feature extractor is aligned with the global model in the feature
space, while the classifier maintains personalized decision-making
capabilities. HFedCKD effectively alleviates the knowledge offset caused by a
low participation rate under data-free knowledge distillation and improves the
performance and stability of the model. We conduct extensive experiments on
image and IoT datasets to comprehensively evaluate and verify the
generalization and robustness of the proposed HFedCKD framework.
| no_new_dataset | 0.947527 |
2503.06519 | Wenhui Zhang | Wenhui Zhang, Huiyu Xu, Zhibo Wang, Zeqing He, Ziqi Zhu, Kui Ren | Can Small Language Models Reliably Resist Jailbreak Attacks? A
Comprehensive Evaluation | 19 pages, 12 figures | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Small language models (SLMs) have emerged as promising alternatives to large
language models (LLMs) due to their low computational demands, enhanced privacy
guarantees and comparable performance in specific domains through light-weight
fine-tuning. Deploying SLMs on edge devices, such as smartphones and smart
vehicles, has become a growing trend. However, the security implications of
SLMs have received less attention than LLMs, particularly regarding jailbreak
attacks, which is recognized as one of the top threats of LLMs by the OWASP. In
this paper, we conduct the first large-scale empirical study of SLMs'
vulnerabilities to jailbreak attacks. Through systematically evaluation on 63
SLMs from 15 mainstream SLM families against 8 state-of-the-art jailbreak
methods, we demonstrate that 47.6% of evaluated SLMs show high susceptibility
to jailbreak attacks (ASR > 40%) and 38.1% of them can not even resist direct
harmful query (ASR > 50%). We further analyze the reasons behind the
vulnerabilities and identify four key factors: model size, model architecture,
training datasets and training techniques. Moreover, we assess the
effectiveness of three prompt-level defense methods and find that none of them
achieve perfect performance, with detection accuracy varying across different
SLMs and attack methods. Notably, we point out that the inherent security
awareness play a critical role in SLM security, and models with strong security
awareness could timely terminate unsafe response with little reminder. Building
upon the findings, we highlight the urgent need for security-by-design
approaches in SLM development and provide valuable insights for building more
trustworthy SLM ecosystem.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 08:47:16 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhang",
"Wenhui",
""
],
[
"Xu",
"Huiyu",
""
],
[
"Wang",
"Zhibo",
""
],
[
"He",
"Zeqing",
""
],
[
"Zhu",
"Ziqi",
""
],
[
"Ren",
"Kui",
""
]
]
| TITLE: Can Small Language Models Reliably Resist Jailbreak Attacks? A
Comprehensive Evaluation
ABSTRACT: Small language models (SLMs) have emerged as promising alternatives to large
language models (LLMs) due to their low computational demands, enhanced privacy
guarantees and comparable performance in specific domains through light-weight
fine-tuning. Deploying SLMs on edge devices, such as smartphones and smart
vehicles, has become a growing trend. However, the security implications of
SLMs have received less attention than LLMs, particularly regarding jailbreak
attacks, which is recognized as one of the top threats of LLMs by the OWASP. In
this paper, we conduct the first large-scale empirical study of SLMs'
vulnerabilities to jailbreak attacks. Through systematically evaluation on 63
SLMs from 15 mainstream SLM families against 8 state-of-the-art jailbreak
methods, we demonstrate that 47.6% of evaluated SLMs show high susceptibility
to jailbreak attacks (ASR > 40%) and 38.1% of them can not even resist direct
harmful query (ASR > 50%). We further analyze the reasons behind the
vulnerabilities and identify four key factors: model size, model architecture,
training datasets and training techniques. Moreover, we assess the
effectiveness of three prompt-level defense methods and find that none of them
achieve perfect performance, with detection accuracy varying across different
SLMs and attack methods. Notably, we point out that the inherent security
awareness play a critical role in SLM security, and models with strong security
awareness could timely terminate unsafe response with little reminder. Building
upon the findings, we highlight the urgent need for security-by-design
approaches in SLM development and provide valuable insights for building more
trustworthy SLM ecosystem.
| no_new_dataset | 0.94256 |
2503.06531 | Jie He | Jie He, Yu Fu | MetaXCR: Reinforcement-Based Meta-Transfer Learning for Cross-Lingual
Commonsense Reasoning | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Commonsense reasoning (CR) has been studied in many pieces of domain and has
achieved great progress with the aid of large datasets. Unfortunately, most
existing CR datasets are built in English, so most previous work focus on
English. Furthermore, as the annotation of commonsense reasoning is costly, it
is impossible to build a large dataset for every novel task. Therefore, there
are growing appeals for Cross-lingual Low-Resource Commonsense Reasoning, which
aims to leverage diverse existed English datasets to help the model adapt to
new cross-lingual target datasets with limited labeled data. In this paper, we
propose a multi-source adapter for cross-lingual low-resource Commonsense
Reasoning (MetaXCR). In this framework, we first extend meta learning by
incorporating multiple training datasets to learn a generalized task adapters
across different tasks. Then, we further introduce a reinforcement-based
sampling strategy to help the model sample the source task that is the most
helpful to the target task. Finally, we introduce two types of cross-lingual
meta-adaption methods to enhance the performance of models on target languages.
Extensive experiments demonstrate MetaXCR is superior over state-of-the-arts,
while being trained with fewer parameters than other work.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 09:27:57 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"He",
"Jie",
""
],
[
"Fu",
"Yu",
""
]
]
| TITLE: MetaXCR: Reinforcement-Based Meta-Transfer Learning for Cross-Lingual
Commonsense Reasoning
ABSTRACT: Commonsense reasoning (CR) has been studied in many pieces of domain and has
achieved great progress with the aid of large datasets. Unfortunately, most
existing CR datasets are built in English, so most previous work focus on
English. Furthermore, as the annotation of commonsense reasoning is costly, it
is impossible to build a large dataset for every novel task. Therefore, there
are growing appeals for Cross-lingual Low-Resource Commonsense Reasoning, which
aims to leverage diverse existed English datasets to help the model adapt to
new cross-lingual target datasets with limited labeled data. In this paper, we
propose a multi-source adapter for cross-lingual low-resource Commonsense
Reasoning (MetaXCR). In this framework, we first extend meta learning by
incorporating multiple training datasets to learn a generalized task adapters
across different tasks. Then, we further introduce a reinforcement-based
sampling strategy to help the model sample the source task that is the most
helpful to the target task. Finally, we introduce two types of cross-lingual
meta-adaption methods to enhance the performance of models on target languages.
Extensive experiments demonstrate MetaXCR is superior over state-of-the-arts,
while being trained with fewer parameters than other work.
| no_new_dataset | 0.947817 |
2503.06534 | Xingwei Tan | Xingwei Tan, Chen Lyu, Hafiz Muhammad Umer, Sahrish Khan, Mahathi
Parvatham, Lois Arthurs, Simon Cullen, Shelley Wilson, Arshad Jhumka,
Gabriele Pergola | SafeSpeech: A Comprehensive and Interactive Tool for Analysing Sexist
and Abusive Language in Conversations | NAACL 2025 system demonstration camera-ready | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Detecting toxic language including sexism, harassment and abusive behaviour,
remains a critical challenge, particularly in its subtle and context-dependent
forms. Existing approaches largely focus on isolated message-level
classification, overlooking toxicity that emerges across conversational
contexts. To promote and enable future research in this direction, we introduce
SafeSpeech, a comprehensive platform for toxic content detection and analysis
that bridges message-level and conversation-level insights. The platform
integrates fine-tuned classifiers and large language models (LLMs) to enable
multi-granularity detection, toxic-aware conversation summarization, and
persona profiling. SafeSpeech also incorporates explainability mechanisms, such
as perplexity gain analysis, to highlight the linguistic elements driving
predictions. Evaluations on benchmark datasets, including EDOS, OffensEval, and
HatEval, demonstrate the reproduction of state-of-the-art performance across
multiple tasks, including fine-grained sexism detection.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 09:31:17 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Tan",
"Xingwei",
""
],
[
"Lyu",
"Chen",
""
],
[
"Umer",
"Hafiz Muhammad",
""
],
[
"Khan",
"Sahrish",
""
],
[
"Parvatham",
"Mahathi",
""
],
[
"Arthurs",
"Lois",
""
],
[
"Cullen",
"Simon",
""
],
[
"Wilson",
"Shelley",
""
],
[
"Jhumka",
"Arshad",
""
],
[
"Pergola",
"Gabriele",
""
]
]
| TITLE: SafeSpeech: A Comprehensive and Interactive Tool for Analysing Sexist
and Abusive Language in Conversations
ABSTRACT: Detecting toxic language including sexism, harassment and abusive behaviour,
remains a critical challenge, particularly in its subtle and context-dependent
forms. Existing approaches largely focus on isolated message-level
classification, overlooking toxicity that emerges across conversational
contexts. To promote and enable future research in this direction, we introduce
SafeSpeech, a comprehensive platform for toxic content detection and analysis
that bridges message-level and conversation-level insights. The platform
integrates fine-tuned classifiers and large language models (LLMs) to enable
multi-granularity detection, toxic-aware conversation summarization, and
persona profiling. SafeSpeech also incorporates explainability mechanisms, such
as perplexity gain analysis, to highlight the linguistic elements driving
predictions. Evaluations on benchmark datasets, including EDOS, OffensEval, and
HatEval, demonstrate the reproduction of state-of-the-art performance across
multiple tasks, including fine-grained sexism detection.
| no_new_dataset | 0.94428 |
2503.06537 | Xiaoyang Liu | Xiaoyang Liu, Yuquan Wang, Zheng Chen, Jiezhang Cao, He Zhang, Yulun
Zhang and Xiaokang Yang | One-Step Diffusion Model for Image Motion-Deblurring | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Currently, methods for single-image deblurring based on CNNs and transformers
have demonstrated promising performance. However, these methods often suffer
from perceptual limitations, poor generalization ability, and struggle with
heavy or complex blur. While diffusion-based methods can partially address
these shortcomings, their multi-step denoising process limits their practical
usage. In this paper, we conduct an in-depth exploration of diffusion models in
deblurring and propose a one-step diffusion model for deblurring (OSDD), a
novel framework that reduces the denoising process to a single step,
significantly improving inference efficiency while maintaining high fidelity.
To tackle fidelity loss in diffusion models, we introduce an enhanced
variational autoencoder (eVAE), which improves structural restoration.
Additionally, we construct a high-quality synthetic deblurring dataset to
mitigate perceptual collapse and design a dynamic dual-adapter (DDA) to enhance
perceptual quality while preserving fidelity. Extensive experiments demonstrate
that our method achieves strong performance on both full and no-reference
metrics. Our code and pre-trained model will be publicly available at
https://github.com/xyLiu339/OSDD.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 09:39:57 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Liu",
"Xiaoyang",
""
],
[
"Wang",
"Yuquan",
""
],
[
"Chen",
"Zheng",
""
],
[
"Cao",
"Jiezhang",
""
],
[
"Zhang",
"He",
""
],
[
"Zhang",
"Yulun",
""
],
[
"Yang",
"Xiaokang",
""
]
]
| TITLE: One-Step Diffusion Model for Image Motion-Deblurring
ABSTRACT: Currently, methods for single-image deblurring based on CNNs and transformers
have demonstrated promising performance. However, these methods often suffer
from perceptual limitations, poor generalization ability, and struggle with
heavy or complex blur. While diffusion-based methods can partially address
these shortcomings, their multi-step denoising process limits their practical
usage. In this paper, we conduct an in-depth exploration of diffusion models in
deblurring and propose a one-step diffusion model for deblurring (OSDD), a
novel framework that reduces the denoising process to a single step,
significantly improving inference efficiency while maintaining high fidelity.
To tackle fidelity loss in diffusion models, we introduce an enhanced
variational autoencoder (eVAE), which improves structural restoration.
Additionally, we construct a high-quality synthetic deblurring dataset to
mitigate perceptual collapse and design a dynamic dual-adapter (DDA) to enhance
perceptual quality while preserving fidelity. Extensive experiments demonstrate
that our method achieves strong performance on both full and no-reference
metrics. Our code and pre-trained model will be publicly available at
https://github.com/xyLiu339/OSDD.
| new_dataset | 0.961207 |
2503.06542 | Yukang Feng | Jianwen Sun, Yukang Feng, Chuanhao Li, Fanrui Zhang, Zizhen Li, Jiaxin
Ai, Sizhuo Zhou, Yu Dai, Shenglin Zhang, Kaipeng Zhang | ARMOR v0.1: Empowering Autoregressive Multimodal Understanding Model
with Interleaved Multimodal Generation via Asymmetric Synergy | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Unified models (UniMs) for multimodal understanding and generation have
recently received much attention in the area of vision and language. Existing
UniMs are designed to simultaneously learn both multimodal understanding and
generation capabilities, demanding substantial computational resources, and
often struggle to generate interleaved text-image. We present ARMOR, a
resource-efficient and pure autoregressive framework that achieves both
understanding and generation by fine-tuning existing multimodal large language
models (MLLMs). Specifically, ARMOR extends existing MLLMs from three
perspectives: (1) For model architecture, an asymmetric encoder-decoder
architecture with a forward-switching mechanism is introduced to unify
embedding space integrating textual and visual modalities for enabling natural
text-image interleaved generation with minimal computational overhead. (2) For
training data, a meticulously curated, high-quality interleaved dataset is
collected for fine-tuning MLLMs. (3) For the training algorithm, we propose a
``what or how to generate" algorithm to empower existing MLLMs with multimodal
generation capabilities while preserving their multimodal understanding
capabilities, through three progressive training stages based on the collected
dataset. Experimental results demonstrate that ARMOR upgrades existing MLLMs to
UniMs with promising image generation capabilities, using limited training
resources. Our code will be released soon at https://armor.github.io.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 10:15:39 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Sun",
"Jianwen",
""
],
[
"Feng",
"Yukang",
""
],
[
"Li",
"Chuanhao",
""
],
[
"Zhang",
"Fanrui",
""
],
[
"Li",
"Zizhen",
""
],
[
"Ai",
"Jiaxin",
""
],
[
"Zhou",
"Sizhuo",
""
],
[
"Dai",
"Yu",
""
],
[
"Zhang",
"Shenglin",
""
],
[
"Zhang",
"Kaipeng",
""
]
]
| TITLE: ARMOR v0.1: Empowering Autoregressive Multimodal Understanding Model
with Interleaved Multimodal Generation via Asymmetric Synergy
ABSTRACT: Unified models (UniMs) for multimodal understanding and generation have
recently received much attention in the area of vision and language. Existing
UniMs are designed to simultaneously learn both multimodal understanding and
generation capabilities, demanding substantial computational resources, and
often struggle to generate interleaved text-image. We present ARMOR, a
resource-efficient and pure autoregressive framework that achieves both
understanding and generation by fine-tuning existing multimodal large language
models (MLLMs). Specifically, ARMOR extends existing MLLMs from three
perspectives: (1) For model architecture, an asymmetric encoder-decoder
architecture with a forward-switching mechanism is introduced to unify
embedding space integrating textual and visual modalities for enabling natural
text-image interleaved generation with minimal computational overhead. (2) For
training data, a meticulously curated, high-quality interleaved dataset is
collected for fine-tuning MLLMs. (3) For the training algorithm, we propose a
``what or how to generate" algorithm to empower existing MLLMs with multimodal
generation capabilities while preserving their multimodal understanding
capabilities, through three progressive training stages based on the collected
dataset. Experimental results demonstrate that ARMOR upgrades existing MLLMs to
UniMs with promising image generation capabilities, using limited training
resources. Our code will be released soon at https://armor.github.io.
| no_new_dataset | 0.795817 |
2503.06550 | Fan Yin | Fan Yin and Philippe Laban and Xiangyu Peng and Yilun Zhou and Yixin
Mao and Vaibhav Vats and Linnea Ross and Divyansh Agarwal and Caiming Xiong
and Chien-Sheng Wu | BingoGuard: LLM Content Moderation Tools with Risk Levels | 10 pages, 4 figures, 4 tables. ICLR 2025 poster | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Malicious content generated by large language models (LLMs) can pose varying
degrees of harm. Although existing LLM-based moderators can detect harmful
content, they struggle to assess risk levels and may miss lower-risk outputs.
Accurate risk assessment allows platforms with different safety thresholds to
tailor content filtering and rejection. In this paper, we introduce per-topic
severity rubrics for 11 harmful topics and build BingoGuard, an LLM-based
moderation system designed to predict both binary safety labels and severity
levels. To address the lack of annotations on levels of severity, we propose a
scalable generate-then-filter framework that first generates responses across
different severity levels and then filters out low-quality responses. Using
this framework, we create BingoGuardTrain, a training dataset with 54,897
examples covering a variety of topics, response severity, styles, and
BingoGuardTest, a test set with 988 examples explicitly labeled based on our
severity rubrics that enables fine-grained analysis on model behaviors on
different severity levels. Our BingoGuard-8B, trained on BingoGuardTrain,
achieves the state-of-the-art performance on several moderation benchmarks,
including WildGuardTest and HarmBench, as well as BingoGuardTest, outperforming
best public models, WildGuard, by 4.3\%. Our analysis demonstrates that
incorporating severity levels into training significantly enhances detection
performance and enables the model to effectively gauge the severity of harmful
responses.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 10:43:09 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Yin",
"Fan",
""
],
[
"Laban",
"Philippe",
""
],
[
"Peng",
"Xiangyu",
""
],
[
"Zhou",
"Yilun",
""
],
[
"Mao",
"Yixin",
""
],
[
"Vats",
"Vaibhav",
""
],
[
"Ross",
"Linnea",
""
],
[
"Agarwal",
"Divyansh",
""
],
[
"Xiong",
"Caiming",
""
],
[
"Wu",
"Chien-Sheng",
""
]
]
| TITLE: BingoGuard: LLM Content Moderation Tools with Risk Levels
ABSTRACT: Malicious content generated by large language models (LLMs) can pose varying
degrees of harm. Although existing LLM-based moderators can detect harmful
content, they struggle to assess risk levels and may miss lower-risk outputs.
Accurate risk assessment allows platforms with different safety thresholds to
tailor content filtering and rejection. In this paper, we introduce per-topic
severity rubrics for 11 harmful topics and build BingoGuard, an LLM-based
moderation system designed to predict both binary safety labels and severity
levels. To address the lack of annotations on levels of severity, we propose a
scalable generate-then-filter framework that first generates responses across
different severity levels and then filters out low-quality responses. Using
this framework, we create BingoGuardTrain, a training dataset with 54,897
examples covering a variety of topics, response severity, styles, and
BingoGuardTest, a test set with 988 examples explicitly labeled based on our
severity rubrics that enables fine-grained analysis on model behaviors on
different severity levels. Our BingoGuard-8B, trained on BingoGuardTrain,
achieves the state-of-the-art performance on several moderation benchmarks,
including WildGuardTest and HarmBench, as well as BingoGuardTest, outperforming
best public models, WildGuard, by 4.3\%. Our analysis demonstrates that
incorporating severity levels into training significantly enhances detection
performance and enables the model to effectively gauge the severity of harmful
responses.
| new_dataset | 0.945901 |
2503.06553 | Jiaxin Ai | Jiaxin Ai, Pengfei Zhou, Zhaopan Xu, Ming Li, Fanrui Zhang, Zizhen Li,
Jianwen Sun, Yukang Feng, Baojin Huang, Zhongyuan Wang, Kaipeng Zhang | ProJudge: A Multi-Modal Multi-Discipline Benchmark and
Instruction-Tuning Dataset for MLLM-based Process Judges | null | null | null | null | cs.AI cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As multi-modal large language models (MLLMs) frequently exhibit errors when
solving scientific problems, evaluating the validity of their reasoning
processes is critical for ensuring reliability and uncovering fine-grained
model weaknesses. Since human evaluation is laborious and costly, prompting
MLLMs as automated process judges has become a common practice. However, the
reliability of these model-based judges remains uncertain. To address this, we
introduce ProJudgeBench, the first comprehensive benchmark specifically
designed for evaluating abilities of MLLM-based process judges. ProJudgeBench
comprises 2,400 test cases and 50,118 step-level labels, spanning four
scientific disciplines with diverse difficulty levels and multi-modal content.
In ProJudgeBench, each step is meticulously annotated by human experts for
correctness, error type, and explanation, enabling a systematic evaluation of
judges' capabilities to detect, classify and diagnose errors. Evaluation on
ProJudgeBench reveals a significant performance gap between open-source and
proprietary models. To bridge this gap, we further propose ProJudge-173k, a
large-scale instruction-tuning dataset, and a Dynamic Dual-Phase fine-tuning
strategy that encourages models to explicitly reason through problem-solving
before assessing solutions. Both contributions significantly enhance the
process evaluation capabilities of open-source models. All the resources will
be released to foster future research of reliable multi-modal process
evaluation.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 10:55:51 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Ai",
"Jiaxin",
""
],
[
"Zhou",
"Pengfei",
""
],
[
"Xu",
"Zhaopan",
""
],
[
"Li",
"Ming",
""
],
[
"Zhang",
"Fanrui",
""
],
[
"Li",
"Zizhen",
""
],
[
"Sun",
"Jianwen",
""
],
[
"Feng",
"Yukang",
""
],
[
"Huang",
"Baojin",
""
],
[
"Wang",
"Zhongyuan",
""
],
[
"Zhang",
"Kaipeng",
""
]
]
| TITLE: ProJudge: A Multi-Modal Multi-Discipline Benchmark and
Instruction-Tuning Dataset for MLLM-based Process Judges
ABSTRACT: As multi-modal large language models (MLLMs) frequently exhibit errors when
solving scientific problems, evaluating the validity of their reasoning
processes is critical for ensuring reliability and uncovering fine-grained
model weaknesses. Since human evaluation is laborious and costly, prompting
MLLMs as automated process judges has become a common practice. However, the
reliability of these model-based judges remains uncertain. To address this, we
introduce ProJudgeBench, the first comprehensive benchmark specifically
designed for evaluating abilities of MLLM-based process judges. ProJudgeBench
comprises 2,400 test cases and 50,118 step-level labels, spanning four
scientific disciplines with diverse difficulty levels and multi-modal content.
In ProJudgeBench, each step is meticulously annotated by human experts for
correctness, error type, and explanation, enabling a systematic evaluation of
judges' capabilities to detect, classify and diagnose errors. Evaluation on
ProJudgeBench reveals a significant performance gap between open-source and
proprietary models. To bridge this gap, we further propose ProJudge-173k, a
large-scale instruction-tuning dataset, and a Dynamic Dual-Phase fine-tuning
strategy that encourages models to explicitly reason through problem-solving
before assessing solutions. Both contributions significantly enhance the
process evaluation capabilities of open-source models. All the resources will
be released to foster future research of reliable multi-modal process
evaluation.
| new_dataset | 0.959573 |
2503.06554 | Chengcheng Zhu | Chengcheng Zhu, Jiale Zhang, Di Wu, Guodong Long | BDPFL: Backdoor Defense for Personalized Federated Learning via
Explainable Distillation | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning is a distributed learning paradigm that facilitates the
collaborative training of a global model across multiple clients while
preserving the privacy of local datasets. To address inherent challenges
related to data heterogeneity and satisfy personalized needs, a new direction
within FL, known as personalized Federated Learning (pFL), has gradually
evolved. Extensive attention has been directed toward developing novel
frameworks and methods to enhance the performance of pFL. Regrettably, the
aspect of security in pFL has been largely overlooked. Our objective is to fill
this gap. Similar to FL, pFL is susceptible to backdoor attacks. However,
existing backdoor defense strategies are primarily tailored to general FL
frameworks, and pFL lacks robustness against backdoor attacks. We propose a
novel, backdoor-robust pFL framework named BDPFL to address these challenges.
First, BDPFL introduces layer-wise mutual distillation that enables clients to
learn their personalized local models while mitigating potential backdoors.
Then, BDPFL employs explanation heatmap to learn high-quality intermediate
representations and enhance the effect of eliminating deeper and more
entrenched backdoors. Moreover, we perform empirical evaluations of BDPFL's
performance on three datasets and compare BDPFL with four backdoor defense
methods. The experiments demonstrate that BDPFL outperforms baseline methods
and is effective under various settings.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 10:59:18 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhu",
"Chengcheng",
""
],
[
"Zhang",
"Jiale",
""
],
[
"Wu",
"Di",
""
],
[
"Long",
"Guodong",
""
]
]
| TITLE: BDPFL: Backdoor Defense for Personalized Federated Learning via
Explainable Distillation
ABSTRACT: Federated learning is a distributed learning paradigm that facilitates the
collaborative training of a global model across multiple clients while
preserving the privacy of local datasets. To address inherent challenges
related to data heterogeneity and satisfy personalized needs, a new direction
within FL, known as personalized Federated Learning (pFL), has gradually
evolved. Extensive attention has been directed toward developing novel
frameworks and methods to enhance the performance of pFL. Regrettably, the
aspect of security in pFL has been largely overlooked. Our objective is to fill
this gap. Similar to FL, pFL is susceptible to backdoor attacks. However,
existing backdoor defense strategies are primarily tailored to general FL
frameworks, and pFL lacks robustness against backdoor attacks. We propose a
novel, backdoor-robust pFL framework named BDPFL to address these challenges.
First, BDPFL introduces layer-wise mutual distillation that enables clients to
learn their personalized local models while mitigating potential backdoors.
Then, BDPFL employs explanation heatmap to learn high-quality intermediate
representations and enhance the effect of eliminating deeper and more
entrenched backdoors. Moreover, we perform empirical evaluations of BDPFL's
performance on three datasets and compare BDPFL with four backdoor defense
methods. The experiments demonstrate that BDPFL outperforms baseline methods
and is effective under various settings.
| no_new_dataset | 0.938237 |
2503.06563 | Jiangdong Cai | Jiangdong Cai, Haotian Jiang, Zhenrong Shen, Yonghao Li, Honglin
Xiong, Lichi Zhang, Qian Wang | LSA: Latent Style Augmentation Towards Stain-Agnostic Cervical Cancer
Screening | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | The deployment of computer-aided diagnosis systems for cervical cancer
screening using whole slide images (WSIs) faces critical challenges due to
domain shifts caused by staining variations across different scanners and
imaging environments. While existing stain augmentation methods improve
patch-level robustness, they fail to scale to WSIs due to two key limitations:
(1) inconsistent stain patterns when extending patch operations to gigapixel
slides, and (2) prohibitive computational/storage costs from offline processing
of augmented WSIs.To address this, we propose Latent Style Augmentation (LSA),
a framework that performs efficient, online stain augmentation directly on
WSI-level latent features. We first introduce WSAug, a WSI-level stain
augmentation method ensuring consistent stain across patches within a WSI.
Using offline-augmented WSIs by WSAug, we design and train Stain Transformer,
which can simulate targeted style in the latent space, efficiently enhancing
the robustness of the WSI-level classifier. We validate our method on a
multi-scanner WSI dataset for cervical cancer diagnosis. Despite being trained
on data from a single scanner, our approach achieves significant performance
improvements on out-of-distribution data from other scanners. Code will be
available at https://github.com/caijd2000/LSA.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 11:33:59 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Cai",
"Jiangdong",
""
],
[
"Jiang",
"Haotian",
""
],
[
"Shen",
"Zhenrong",
""
],
[
"Li",
"Yonghao",
""
],
[
"Xiong",
"Honglin",
""
],
[
"Zhang",
"Lichi",
""
],
[
"Wang",
"Qian",
""
]
]
| TITLE: LSA: Latent Style Augmentation Towards Stain-Agnostic Cervical Cancer
Screening
ABSTRACT: The deployment of computer-aided diagnosis systems for cervical cancer
screening using whole slide images (WSIs) faces critical challenges due to
domain shifts caused by staining variations across different scanners and
imaging environments. While existing stain augmentation methods improve
patch-level robustness, they fail to scale to WSIs due to two key limitations:
(1) inconsistent stain patterns when extending patch operations to gigapixel
slides, and (2) prohibitive computational/storage costs from offline processing
of augmented WSIs.To address this, we propose Latent Style Augmentation (LSA),
a framework that performs efficient, online stain augmentation directly on
WSI-level latent features. We first introduce WSAug, a WSI-level stain
augmentation method ensuring consistent stain across patches within a WSI.
Using offline-augmented WSIs by WSAug, we design and train Stain Transformer,
which can simulate targeted style in the latent space, efficiently enhancing
the robustness of the WSI-level classifier. We validate our method on a
multi-scanner WSI dataset for cervical cancer diagnosis. Despite being trained
on data from a single scanner, our approach achieves significant performance
improvements on out-of-distribution data from other scanners. Code will be
available at https://github.com/caijd2000/LSA.
| no_new_dataset | 0.949295 |
2503.06565 | Shi-Jie Li | Shijie Li, Xun Xu, Si Yong Yeo, Xulei Yang | Future-Aware Interaction Network For Motion Forecasting | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Motion forecasting is a crucial component of autonomous driving systems,
enabling the generation of accurate and smooth future trajectories to ensure
safe navigation to the destination. In previous methods, potential future
trajectories are often absent in the scene encoding stage, which may lead to
suboptimal outcomes. Additionally, prior approaches typically employ
transformer architectures for spatiotemporal modeling of trajectories and map
information, which suffer from the quadratic scaling complexity of the
transformer architecture. In this work, we propose an interaction-based method,
named Future-Aware Interaction Network, that introduces potential future
trajectories into scene encoding for a comprehensive traffic representation.
Furthermore, a State Space Model (SSM), specifically Mamba, is introduced for
both spatial and temporal modeling. To adapt Mamba for spatial interaction
modeling, we propose an adaptive reordering strategy that transforms unordered
data into a structured sequence. Additionally, Mamba is employed to refine
generated future trajectories temporally, ensuring more consistent predictions.
These enhancements not only improve model efficiency but also enhance the
accuracy and diversity of predictions. We conduct comprehensive experiments on
the widely used Argoverse 1 and Argoverse 2 datasets, demonstrating that the
proposed method achieves superior performance compared to previous approaches
in a more efficient way. The code will be released according to the acceptance.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 11:38:34 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Shijie",
""
],
[
"Xu",
"Xun",
""
],
[
"Yeo",
"Si Yong",
""
],
[
"Yang",
"Xulei",
""
]
]
| TITLE: Future-Aware Interaction Network For Motion Forecasting
ABSTRACT: Motion forecasting is a crucial component of autonomous driving systems,
enabling the generation of accurate and smooth future trajectories to ensure
safe navigation to the destination. In previous methods, potential future
trajectories are often absent in the scene encoding stage, which may lead to
suboptimal outcomes. Additionally, prior approaches typically employ
transformer architectures for spatiotemporal modeling of trajectories and map
information, which suffer from the quadratic scaling complexity of the
transformer architecture. In this work, we propose an interaction-based method,
named Future-Aware Interaction Network, that introduces potential future
trajectories into scene encoding for a comprehensive traffic representation.
Furthermore, a State Space Model (SSM), specifically Mamba, is introduced for
both spatial and temporal modeling. To adapt Mamba for spatial interaction
modeling, we propose an adaptive reordering strategy that transforms unordered
data into a structured sequence. Additionally, Mamba is employed to refine
generated future trajectories temporally, ensuring more consistent predictions.
These enhancements not only improve model efficiency but also enhance the
accuracy and diversity of predictions. We conduct comprehensive experiments on
the widely used Argoverse 1 and Argoverse 2 datasets, demonstrating that the
proposed method achieves superior performance compared to previous approaches
in a more efficient way. The code will be released according to the acceptance.
| no_new_dataset | 0.950411 |
2503.06567 | Yao Cheng | Yao Cheng, Yibo Zhao, Jiapeng Zhu, Yao Liu, Xing Sun, Xiang Li | Human Cognition Inspired RAG with Knowledge Graph for Complex Problem
Solving | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have demonstrated transformative potential
across various domains, yet they face significant challenges in knowledge
integration and complex problem reasoning, often leading to hallucinations and
unreliable outputs. Retrieval-Augmented Generation (RAG) has emerged as a
promising solution to enhance LLMs accuracy by incorporating external
knowledge. However, traditional RAG systems struggle with processing complex
relational information and multi-step reasoning, limiting their effectiveness
in advanced problem-solving tasks. To address these limitations, we propose
CogGRAG, a cognition inspired graph-based RAG framework, designed to improve
LLMs performance in Knowledge Graph Question Answering (KGQA). Inspired by the
human cognitive process of decomposing complex problems and performing
self-verification, our framework introduces a three-stage methodology:
decomposition, retrieval, and reasoning with self-verification. By integrating
these components, CogGRAG enhances the accuracy of LLMs in complex problem
solving. We conduct systematic experiments with three LLM backbones on four
benchmark datasets, where CogGRAG outperforms the baselines.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 11:50:39 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Cheng",
"Yao",
""
],
[
"Zhao",
"Yibo",
""
],
[
"Zhu",
"Jiapeng",
""
],
[
"Liu",
"Yao",
""
],
[
"Sun",
"Xing",
""
],
[
"Li",
"Xiang",
""
]
]
| TITLE: Human Cognition Inspired RAG with Knowledge Graph for Complex Problem
Solving
ABSTRACT: Large language models (LLMs) have demonstrated transformative potential
across various domains, yet they face significant challenges in knowledge
integration and complex problem reasoning, often leading to hallucinations and
unreliable outputs. Retrieval-Augmented Generation (RAG) has emerged as a
promising solution to enhance LLMs accuracy by incorporating external
knowledge. However, traditional RAG systems struggle with processing complex
relational information and multi-step reasoning, limiting their effectiveness
in advanced problem-solving tasks. To address these limitations, we propose
CogGRAG, a cognition inspired graph-based RAG framework, designed to improve
LLMs performance in Knowledge Graph Question Answering (KGQA). Inspired by the
human cognitive process of decomposing complex problems and performing
self-verification, our framework introduces a three-stage methodology:
decomposition, retrieval, and reasoning with self-verification. By integrating
these components, CogGRAG enhances the accuracy of LLMs in complex problem
solving. We conduct systematic experiments with three LLM backbones on four
benchmark datasets, where CogGRAG outperforms the baselines.
| no_new_dataset | 0.941115 |
2503.06569 | Shi-Jie Li | Shijie Li, Zhongyao Cheng, Rong Li, Shuai Li, Juergen Gall, Xun Xu,
Xulei Yang | Global-Aware Monocular Semantic Scene Completion with State Space Models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Monocular Semantic Scene Completion (MonoSSC) reconstructs and interprets 3D
environments from a single image, enabling diverse real-world applications.
However, existing methods are often constrained by the local receptive field of
Convolutional Neural Networks (CNNs), making it challenging to handle the
non-uniform distribution of projected points (Fig. \ref{fig:perspective}) and
effectively reconstruct missing information caused by the 3D-to-2D projection.
In this work, we introduce GA-MonoSSC, a hybrid architecture for MonoSSC that
effectively captures global context in both the 2D image domain and 3D space.
Specifically, we propose a Dual-Head Multi-Modality Encoder, which leverages a
Transformer architecture to capture spatial relationships across all features
in the 2D image domain, enabling more comprehensive 2D feature extraction.
Additionally, we introduce the Frustum Mamba Decoder, built on the State Space
Model (SSM), to efficiently capture long-range dependencies in 3D space.
Furthermore, we propose a frustum reordering strategy within the Frustum Mamba
Decoder to mitigate feature discontinuities in the reordered voxel sequence,
ensuring better alignment with the scan mechanism of the State Space Model
(SSM) for improved 3D representation learning. We conduct extensive experiments
on the widely used Occ-ScanNet and NYUv2 datasets, demonstrating that our
proposed method achieves state-of-the-art performance, validating its
effectiveness. The code will be released upon acceptance.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 11:55:40 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Shijie",
""
],
[
"Cheng",
"Zhongyao",
""
],
[
"Li",
"Rong",
""
],
[
"Li",
"Shuai",
""
],
[
"Gall",
"Juergen",
""
],
[
"Xu",
"Xun",
""
],
[
"Yang",
"Xulei",
""
]
]
| TITLE: Global-Aware Monocular Semantic Scene Completion with State Space Models
ABSTRACT: Monocular Semantic Scene Completion (MonoSSC) reconstructs and interprets 3D
environments from a single image, enabling diverse real-world applications.
However, existing methods are often constrained by the local receptive field of
Convolutional Neural Networks (CNNs), making it challenging to handle the
non-uniform distribution of projected points (Fig. \ref{fig:perspective}) and
effectively reconstruct missing information caused by the 3D-to-2D projection.
In this work, we introduce GA-MonoSSC, a hybrid architecture for MonoSSC that
effectively captures global context in both the 2D image domain and 3D space.
Specifically, we propose a Dual-Head Multi-Modality Encoder, which leverages a
Transformer architecture to capture spatial relationships across all features
in the 2D image domain, enabling more comprehensive 2D feature extraction.
Additionally, we introduce the Frustum Mamba Decoder, built on the State Space
Model (SSM), to efficiently capture long-range dependencies in 3D space.
Furthermore, we propose a frustum reordering strategy within the Frustum Mamba
Decoder to mitigate feature discontinuities in the reordered voxel sequence,
ensuring better alignment with the scan mechanism of the State Space Model
(SSM) for improved 3D representation learning. We conduct extensive experiments
on the widely used Occ-ScanNet and NYUv2 datasets, demonstrating that our
proposed method achieves state-of-the-art performance, validating its
effectiveness. The code will be released upon acceptance.
| no_new_dataset | 0.952442 |
2503.06573 | Gili Lior | Gili Lior, Asaf Yehudai, Ariel Gera, Liat Ein-Dor | WildIFEval: Instruction Following in the Wild | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Recent LLMs have shown remarkable success in following user instructions, yet
handling instructions with multiple constraints remains a significant
challenge. In this work, we introduce WildIFEval - a large-scale dataset of 12K
real user instructions with diverse, multi-constraint conditions. Unlike prior
datasets, our collection spans a broad lexical and topical spectrum of
constraints, in natural user prompts. We categorize these constraints into
eight high-level classes to capture their distribution and dynamics in
real-world scenarios. Leveraging WildIFEval, we conduct extensive experiments
to benchmark the instruction-following capabilities of leading LLMs. Our
findings reveal that all evaluated models experience performance degradation
with an increasing number of constraints. Thus, we show that all models have a
large room for improvement on such tasks. Moreover, we observe that the
specific type of constraint plays a critical role in model performance. We
release our dataset to promote further research on instruction-following under
complex, realistic conditions.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 12:06:29 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Lior",
"Gili",
""
],
[
"Yehudai",
"Asaf",
""
],
[
"Gera",
"Ariel",
""
],
[
"Ein-Dor",
"Liat",
""
]
]
| TITLE: WildIFEval: Instruction Following in the Wild
ABSTRACT: Recent LLMs have shown remarkable success in following user instructions, yet
handling instructions with multiple constraints remains a significant
challenge. In this work, we introduce WildIFEval - a large-scale dataset of 12K
real user instructions with diverse, multi-constraint conditions. Unlike prior
datasets, our collection spans a broad lexical and topical spectrum of
constraints, in natural user prompts. We categorize these constraints into
eight high-level classes to capture their distribution and dynamics in
real-world scenarios. Leveraging WildIFEval, we conduct extensive experiments
to benchmark the instruction-following capabilities of leading LLMs. Our
findings reveal that all evaluated models experience performance degradation
with an increasing number of constraints. Thus, we show that all models have a
large room for improvement on such tasks. Moreover, we observe that the
specific type of constraint plays a critical role in model performance. We
release our dataset to promote further research on instruction-following under
complex, realistic conditions.
| new_dataset | 0.958109 |
2503.06580 | Yuxiang Zhang | Yuxiang Zhang, Yuqi Yang, Jiangming Shu, Xinyan Wen, Jitao Sang | Agent models: Internalizing Chain-of-Action Generation into Reasoning
models | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Traditional agentic workflows rely on external prompts to manage interactions
with tools and the environment, which limits the autonomy of reasoning models.
We position \emph{Large Agent Models (LAMs)} that internalize the generation of
\emph{Chain-of-Action (CoA)}, enabling the model to autonomously decide when
and how to use external tools. Our proposed AutoCoA framework combines
supervised fine-tuning (SFT) and reinforcement learning (RL), allowing the
model to seamlessly switch between reasoning and action while efficiently
managing environment interactions. Main components include step-level action
triggering, trajectory-level CoA optimization, and an internal world model to
reduce real-environment interaction costs. Evaluations on open-domain QA tasks
demonstrate that AutoCoA-trained agent models significantly outperform
ReAct-based workflows in task completion, especially in tasks that require
long-term reasoning and multi-step actions. Code and dataset are available at
https://github.com/ADaM-BJTU/AutoCoA
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 12:19:47 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhang",
"Yuxiang",
""
],
[
"Yang",
"Yuqi",
""
],
[
"Shu",
"Jiangming",
""
],
[
"Wen",
"Xinyan",
""
],
[
"Sang",
"Jitao",
""
]
]
| TITLE: Agent models: Internalizing Chain-of-Action Generation into Reasoning
models
ABSTRACT: Traditional agentic workflows rely on external prompts to manage interactions
with tools and the environment, which limits the autonomy of reasoning models.
We position \emph{Large Agent Models (LAMs)} that internalize the generation of
\emph{Chain-of-Action (CoA)}, enabling the model to autonomously decide when
and how to use external tools. Our proposed AutoCoA framework combines
supervised fine-tuning (SFT) and reinforcement learning (RL), allowing the
model to seamlessly switch between reasoning and action while efficiently
managing environment interactions. Main components include step-level action
triggering, trajectory-level CoA optimization, and an internal world model to
reduce real-environment interaction costs. Evaluations on open-domain QA tasks
demonstrate that AutoCoA-trained agent models significantly outperform
ReAct-based workflows in task completion, especially in tasks that require
long-term reasoning and multi-step actions. Code and dataset are available at
https://github.com/ADaM-BJTU/AutoCoA
| no_new_dataset | 0.945399 |
2503.06587 | Yang Zhou | Xiaoming Peng, Yixin Yang, Yang Zhou, Hui Huang | Introducing Unbiased Depth into 2D Gaussian Splatting for High-accuracy
Surface Reconstruction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, 2D Gaussian Splatting (2DGS) has demonstrated superior geometry
reconstruction quality than the popular 3DGS by using 2D surfels to approximate
thin surfaces. However, it falls short when dealing with glossy surfaces,
resulting in visible holes in these areas. We found the reflection
discontinuity causes the issue. To fit the jump from diffuse to specular
reflection at different viewing angles, depth bias is introduced in the
optimized Gaussian primitives. To address that, we first replace the depth
distortion loss in 2DGS with a novel depth convergence loss, which imposes a
strong constraint on depth continuity. Then, we rectified the depth criterion
in determining the actual surface, which fully accounts for all the
intersecting Gaussians along the ray. Qualitative and quantitative evaluations
across various datasets reveal that our method significantly improves
reconstruction quality, with more complete and accurate surfaces than 2DGS.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 12:38:01 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Peng",
"Xiaoming",
""
],
[
"Yang",
"Yixin",
""
],
[
"Zhou",
"Yang",
""
],
[
"Huang",
"Hui",
""
]
]
| TITLE: Introducing Unbiased Depth into 2D Gaussian Splatting for High-accuracy
Surface Reconstruction
ABSTRACT: Recently, 2D Gaussian Splatting (2DGS) has demonstrated superior geometry
reconstruction quality than the popular 3DGS by using 2D surfels to approximate
thin surfaces. However, it falls short when dealing with glossy surfaces,
resulting in visible holes in these areas. We found the reflection
discontinuity causes the issue. To fit the jump from diffuse to specular
reflection at different viewing angles, depth bias is introduced in the
optimized Gaussian primitives. To address that, we first replace the depth
distortion loss in 2DGS with a novel depth convergence loss, which imposes a
strong constraint on depth continuity. Then, we rectified the depth criterion
in determining the actual surface, which fully accounts for all the
intersecting Gaussians along the ray. Qualitative and quantitative evaluations
across various datasets reveal that our method significantly improves
reconstruction quality, with more complete and accurate surfaces than 2DGS.
| no_new_dataset | 0.947186 |
2503.06588 | Yaxuan Li | Yaxuan Li, Han Jiang, Yifei Ma, Shihua Qin, Fangxu Xing | Speech Audio Generation from dynamic MRI via a Knowledge Enhanced
Conditional Variational Autoencoder | null | null | null | null | cs.SD cs.CV | http://creativecommons.org/licenses/by/4.0/ | Dynamic Magnetic Resonance Imaging (MRI) of the vocal tract has become an
increasingly adopted imaging modality for speech motor studies. Beyond image
signals, systematic data loss, noise pollution, and audio file corruption can
occur due to the unpredictability of the MRI acquisition environment. In such
cases, generating audio from images is critical for data recovery in both
clinical and research applications. However, this remains challenging due to
hardware constraints, acoustic interference, and data corruption. Existing
solutions, such as denoising and multi-stage synthesis methods, face
limitations in audio fidelity and generalizability. To address these
challenges, we propose a Knowledge Enhanced Conditional Variational Autoencoder
(KE-CVAE), a novel two-step "knowledge enhancement + variational inference"
framework for generating speech audio signals from cine dynamic MRI sequences.
This approach introduces two key innovations: (1) integration of unlabeled MRI
data for knowledge enhancement, and (2) a variational inference architecture to
improve generative modeling capacity. To the best of our knowledge, this is one
of the first attempts at synthesizing speech audio directly from dynamic MRI
video sequences. The proposed method was trained and evaluated on an
open-source dynamic vocal tract MRI dataset recorded during speech.
Experimental results demonstrate its effectiveness in generating natural speech
waveforms while addressing MRI-specific acoustic challenges, outperforming
conventional deep learning-based synthesis approaches.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 12:40:16 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Li",
"Yaxuan",
""
],
[
"Jiang",
"Han",
""
],
[
"Ma",
"Yifei",
""
],
[
"Qin",
"Shihua",
""
],
[
"Xing",
"Fangxu",
""
]
]
| TITLE: Speech Audio Generation from dynamic MRI via a Knowledge Enhanced
Conditional Variational Autoencoder
ABSTRACT: Dynamic Magnetic Resonance Imaging (MRI) of the vocal tract has become an
increasingly adopted imaging modality for speech motor studies. Beyond image
signals, systematic data loss, noise pollution, and audio file corruption can
occur due to the unpredictability of the MRI acquisition environment. In such
cases, generating audio from images is critical for data recovery in both
clinical and research applications. However, this remains challenging due to
hardware constraints, acoustic interference, and data corruption. Existing
solutions, such as denoising and multi-stage synthesis methods, face
limitations in audio fidelity and generalizability. To address these
challenges, we propose a Knowledge Enhanced Conditional Variational Autoencoder
(KE-CVAE), a novel two-step "knowledge enhancement + variational inference"
framework for generating speech audio signals from cine dynamic MRI sequences.
This approach introduces two key innovations: (1) integration of unlabeled MRI
data for knowledge enhancement, and (2) a variational inference architecture to
improve generative modeling capacity. To the best of our knowledge, this is one
of the first attempts at synthesizing speech audio directly from dynamic MRI
video sequences. The proposed method was trained and evaluated on an
open-source dynamic vocal tract MRI dataset recorded during speech.
Experimental results demonstrate its effectiveness in generating natural speech
waveforms while addressing MRI-specific acoustic challenges, outperforming
conventional deep learning-based synthesis approaches.
| no_new_dataset | 0.949153 |
2503.06594 | Yingfeng Luo | Yingfeng Luo, Tong Zheng, Yongyu Mu, Bei Li, Qinghong Zhang, Yongqi
Gao, Ziqiang Xu, Peinan Feng, Xiaoqian Liu, Tong Xiao, Jingbo Zhu | Beyond Decoder-only: Large Language Models Can be Good Encoders for
Machine Translation | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The field of neural machine translation (NMT) has changed with the advent of
large language models (LLMs). Much of the recent emphasis in natural language
processing (NLP) has been on modeling machine translation and many other
problems using a single pre-trained Transformer decoder, while encoder-decoder
architectures, which were the standard in earlier NMT models, have received
relatively less attention. In this paper, we explore translation models that
are universal, efficient, and easy to optimize, by marrying the world of LLMs
with the world of NMT. We apply LLMs to NMT encoding and leave the NMT decoder
unchanged. We also develop methods for adapting LLMs to work better with the
NMT decoder. Furthermore, we construct a new dataset involving multiple tasks
to assess how well the machine translation system generalizes across various
tasks. Evaluations on the WMT and our datasets show that results using our
method match or surpass a range of baselines in terms of translation quality,
but achieve $2.4 \sim 6.5 \times$ inference speedups and a $75\%$ reduction in
the memory footprint of the KV cache. It also demonstrates strong
generalization across a variety of translation-related tasks.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 12:54:05 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Luo",
"Yingfeng",
""
],
[
"Zheng",
"Tong",
""
],
[
"Mu",
"Yongyu",
""
],
[
"Li",
"Bei",
""
],
[
"Zhang",
"Qinghong",
""
],
[
"Gao",
"Yongqi",
""
],
[
"Xu",
"Ziqiang",
""
],
[
"Feng",
"Peinan",
""
],
[
"Liu",
"Xiaoqian",
""
],
[
"Xiao",
"Tong",
""
],
[
"Zhu",
"Jingbo",
""
]
]
| TITLE: Beyond Decoder-only: Large Language Models Can be Good Encoders for
Machine Translation
ABSTRACT: The field of neural machine translation (NMT) has changed with the advent of
large language models (LLMs). Much of the recent emphasis in natural language
processing (NLP) has been on modeling machine translation and many other
problems using a single pre-trained Transformer decoder, while encoder-decoder
architectures, which were the standard in earlier NMT models, have received
relatively less attention. In this paper, we explore translation models that
are universal, efficient, and easy to optimize, by marrying the world of LLMs
with the world of NMT. We apply LLMs to NMT encoding and leave the NMT decoder
unchanged. We also develop methods for adapting LLMs to work better with the
NMT decoder. Furthermore, we construct a new dataset involving multiple tasks
to assess how well the machine translation system generalizes across various
tasks. Evaluations on the WMT and our datasets show that results using our
method match or surpass a range of baselines in terms of translation quality,
but achieve $2.4 \sim 6.5 \times$ inference speedups and a $75\%$ reduction in
the memory footprint of the KV cache. It also demonstrates strong
generalization across a variety of translation-related tasks.
| new_dataset | 0.962497 |
2503.06598 | Hao Xu | Hao Xu, Tengfei Xue, Dongnan Liu, Yuqian Chen, Fan Zhang, Carl-Fredrik
Westin, Ron Kikinis, Lauren J. O'Donnell, Weidong Cai | MultiCo3D: Multi-Label Voxel Contrast for One-Shot Incremental
Segmentation of 3D Neuroimages | 13 pages, 6 figures, 6 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 3D neuroimages provide a comprehensive view of brain structure and function,
aiding in precise localization and functional connectivity analysis.
Segmentation of white matter (WM) tracts using 3D neuroimages is vital for
understanding the brain's structural connectivity in both healthy and diseased
states. One-shot Class Incremental Semantic Segmentation (OCIS) refers to
effectively segmenting new (novel) classes using only a single sample while
retaining knowledge of old (base) classes without forgetting. Voxel-contrastive
OCIS methods adjust the feature space to alleviate the feature overlap problem
between the base and novel classes. However, since WM tract segmentation is a
multi-label segmentation task, existing single-label voxel contrastive-based
methods may cause inherent contradictions. To address this, we propose a new
multi-label voxel contrast framework called MultiCo3D for one-shot class
incremental tract segmentation. Our method utilizes uncertainty distillation to
preserve base tract segmentation knowledge while adjusting the feature space
with multi-label voxel contrast to alleviate feature overlap when learning
novel tracts and dynamically weighting multi losses to balance overall loss. We
compare our method against several state-of-the-art (SOTA) approaches. The
experimental results show that our method significantly enhances one-shot class
incremental tract segmentation accuracy across five different experimental
setups on HCP and Preto datasets.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:06:20 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Xu",
"Hao",
""
],
[
"Xue",
"Tengfei",
""
],
[
"Liu",
"Dongnan",
""
],
[
"Chen",
"Yuqian",
""
],
[
"Zhang",
"Fan",
""
],
[
"Westin",
"Carl-Fredrik",
""
],
[
"Kikinis",
"Ron",
""
],
[
"O'Donnell",
"Lauren J.",
""
],
[
"Cai",
"Weidong",
""
]
]
| TITLE: MultiCo3D: Multi-Label Voxel Contrast for One-Shot Incremental
Segmentation of 3D Neuroimages
ABSTRACT: 3D neuroimages provide a comprehensive view of brain structure and function,
aiding in precise localization and functional connectivity analysis.
Segmentation of white matter (WM) tracts using 3D neuroimages is vital for
understanding the brain's structural connectivity in both healthy and diseased
states. One-shot Class Incremental Semantic Segmentation (OCIS) refers to
effectively segmenting new (novel) classes using only a single sample while
retaining knowledge of old (base) classes without forgetting. Voxel-contrastive
OCIS methods adjust the feature space to alleviate the feature overlap problem
between the base and novel classes. However, since WM tract segmentation is a
multi-label segmentation task, existing single-label voxel contrastive-based
methods may cause inherent contradictions. To address this, we propose a new
multi-label voxel contrast framework called MultiCo3D for one-shot class
incremental tract segmentation. Our method utilizes uncertainty distillation to
preserve base tract segmentation knowledge while adjusting the feature space
with multi-label voxel contrast to alleviate feature overlap when learning
novel tracts and dynamically weighting multi losses to balance overall loss. We
compare our method against several state-of-the-art (SOTA) approaches. The
experimental results show that our method significantly enhances one-shot class
incremental tract segmentation accuracy across five different experimental
setups on HCP and Preto datasets.
| no_new_dataset | 0.951774 |
2503.06602 | Pranoy Panda | Pranoy Panda, Siddharth Tandon, Vineeth N Balasubramanian | FW-Shapley: Real-time Estimation of Weighted Shapley Values | Accepted at ICASSP 2024 | null | 10.1109/ICASSP48485.2024.10446778 | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Fair credit assignment is essential in various machine learning (ML)
applications, and Shapley values have emerged as a valuable tool for this
purpose. However, in critical ML applications such as data valuation and
feature attribution, the uniform weighting of Shapley values across subset
cardinalities leads to unintuitive credit assignments. To address this,
weighted Shapley values were proposed as a generalization, allowing different
weights for subsets with different cardinalities. Despite their advantages,
similar to Shapley values, Weighted Shapley values suffer from exponential
compute costs, making them impractical for high-dimensional datasets. To tackle
this issue, we present two key contributions. Firstly, we provide a weighted
least squares characterization of weighted Shapley values. Next, using this
characterization, we propose Fast Weighted Shapley (FW-Shapley), an amortized
framework for efficiently computing weighted Shapley values using a learned
estimator. We further show that our estimator's training procedure is
theoretically valid even though we do not use ground truth Weighted Shapley
values during training. On the feature attribution task, we outperform the
learned estimator FastSHAP by $27\%$ (on average) in terms of Inclusion AUC.
For data valuation, we are much faster (14 times) while being comparable to the
state-of-the-art KNN Shapley.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:13:14 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Panda",
"Pranoy",
""
],
[
"Tandon",
"Siddharth",
""
],
[
"Balasubramanian",
"Vineeth N",
""
]
]
| TITLE: FW-Shapley: Real-time Estimation of Weighted Shapley Values
ABSTRACT: Fair credit assignment is essential in various machine learning (ML)
applications, and Shapley values have emerged as a valuable tool for this
purpose. However, in critical ML applications such as data valuation and
feature attribution, the uniform weighting of Shapley values across subset
cardinalities leads to unintuitive credit assignments. To address this,
weighted Shapley values were proposed as a generalization, allowing different
weights for subsets with different cardinalities. Despite their advantages,
similar to Shapley values, Weighted Shapley values suffer from exponential
compute costs, making them impractical for high-dimensional datasets. To tackle
this issue, we present two key contributions. Firstly, we provide a weighted
least squares characterization of weighted Shapley values. Next, using this
characterization, we propose Fast Weighted Shapley (FW-Shapley), an amortized
framework for efficiently computing weighted Shapley values using a learned
estimator. We further show that our estimator's training procedure is
theoretically valid even though we do not use ground truth Weighted Shapley
values during training. On the feature attribution task, we outperform the
learned estimator FastSHAP by $27\%$ (on average) in terms of Inclusion AUC.
For data valuation, we are much faster (14 times) while being comparable to the
state-of-the-art KNN Shapley.
| no_new_dataset | 0.949435 |
2503.06604 | Renhao Lu | Renhao Lu | Steerable Pyramid Weighted Loss: Multi-Scale Adaptive Weighting for
Semantic Segmentation | 9 pages, 4 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Semantic segmentation is a core task in computer vision with applications in
biomedical imaging, remote sensing, and autonomous driving. While standard loss
functions such as cross-entropy and Dice loss perform well in general cases,
they often struggle with fine structures, particularly in tasks involving thin
structures or closely packed objects. Various weight map-based loss functions
have been proposed to address this issue by assigning higher loss weights to
pixels prone to misclassification. However, these methods typically rely on
precomputed or runtime-generated weight maps based on distance transforms,
which impose significant computational costs and fail to adapt to evolving
network predictions. In this paper, we propose a novel steerable pyramid-based
weighted (SPW) loss function that efficiently generates adaptive weight maps.
Unlike traditional boundary-aware losses that depend on static or iteratively
updated distance maps, our method leverages steerable pyramids to dynamically
emphasize regions across multiple frequency bands (capturing features at
different scales) while maintaining computational efficiency. Additionally, by
incorporating network predictions into the weight computation, our approach
enables adaptive refinement during training. We evaluate our method on the
SNEMI3D, GlaS, and DRIVE datasets, benchmarking it against 11 state-of-the-art
loss functions. Our results demonstrate that the proposed SPW loss function
achieves superior pixel precision and segmentation accuracy with minimal
computational overhead. This work provides an effective and efficient solution
for improving semantic segmentation, particularly for applications requiring
multiscale feature representation. The code is avaiable at
https://anonymous.4open.science/r/SPW-0884
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:15:01 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Lu",
"Renhao",
""
]
]
| TITLE: Steerable Pyramid Weighted Loss: Multi-Scale Adaptive Weighting for
Semantic Segmentation
ABSTRACT: Semantic segmentation is a core task in computer vision with applications in
biomedical imaging, remote sensing, and autonomous driving. While standard loss
functions such as cross-entropy and Dice loss perform well in general cases,
they often struggle with fine structures, particularly in tasks involving thin
structures or closely packed objects. Various weight map-based loss functions
have been proposed to address this issue by assigning higher loss weights to
pixels prone to misclassification. However, these methods typically rely on
precomputed or runtime-generated weight maps based on distance transforms,
which impose significant computational costs and fail to adapt to evolving
network predictions. In this paper, we propose a novel steerable pyramid-based
weighted (SPW) loss function that efficiently generates adaptive weight maps.
Unlike traditional boundary-aware losses that depend on static or iteratively
updated distance maps, our method leverages steerable pyramids to dynamically
emphasize regions across multiple frequency bands (capturing features at
different scales) while maintaining computational efficiency. Additionally, by
incorporating network predictions into the weight computation, our approach
enables adaptive refinement during training. We evaluate our method on the
SNEMI3D, GlaS, and DRIVE datasets, benchmarking it against 11 state-of-the-art
loss functions. Our results demonstrate that the proposed SPW loss function
achieves superior pixel precision and segmentation accuracy with minimal
computational overhead. This work provides an effective and efficient solution
for improving semantic segmentation, particularly for applications requiring
multiscale feature representation. The code is avaiable at
https://anonymous.4open.science/r/SPW-0884
| no_new_dataset | 0.949902 |
2503.06606 | Pranoy Panda | Pranoy Panda, Kancheti Sai Srinivas, Vineeth N Balasubramanian, Gaurav
Sinha | Interpretable Model Drift Detection | Accepted at CODS-COMAD 2024 | null | 10.1145/3632410.3632434 | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Data in the real world often has an evolving distribution. Thus, machine
learning models trained on such data get outdated over time. This phenomenon is
called model drift. Knowledge of this drift serves two purposes: (i) Retain an
accurate model and (ii) Discovery of knowledge or insights about change in the
relationship between input features and output variable w.r.t. the model. Most
existing works focus only on detecting model drift but offer no
interpretability. In this work, we take a principled approach to study the
problem of interpretable model drift detection from a risk perspective using a
feature-interaction aware hypothesis testing framework, which enjoys guarantees
on test power. The proposed framework is generic, i.e., it can be adapted to
both classification and regression tasks. Experiments on several standard drift
detection datasets show that our method is superior to existing interpretable
methods (especially on real-world datasets) and on par with state-of-the-art
black-box drift detection methods. We also quantitatively and qualitatively
study the interpretability aspect including a case study on USENET2 dataset. We
find our method focuses on model and drift sensitive features compared to
baseline interpretable drift detectors.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:19:06 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Panda",
"Pranoy",
""
],
[
"Srinivas",
"Kancheti Sai",
""
],
[
"Balasubramanian",
"Vineeth N",
""
],
[
"Sinha",
"Gaurav",
""
]
]
| TITLE: Interpretable Model Drift Detection
ABSTRACT: Data in the real world often has an evolving distribution. Thus, machine
learning models trained on such data get outdated over time. This phenomenon is
called model drift. Knowledge of this drift serves two purposes: (i) Retain an
accurate model and (ii) Discovery of knowledge or insights about change in the
relationship between input features and output variable w.r.t. the model. Most
existing works focus only on detecting model drift but offer no
interpretability. In this work, we take a principled approach to study the
problem of interpretable model drift detection from a risk perspective using a
feature-interaction aware hypothesis testing framework, which enjoys guarantees
on test power. The proposed framework is generic, i.e., it can be adapted to
both classification and regression tasks. Experiments on several standard drift
detection datasets show that our method is superior to existing interpretable
methods (especially on real-world datasets) and on par with state-of-the-art
black-box drift detection methods. We also quantitatively and qualitatively
study the interpretability aspect including a case study on USENET2 dataset. We
find our method focuses on model and drift sensitive features compared to
baseline interpretable drift detectors.
| no_new_dataset | 0.944228 |
2503.06608 | Ruchi Bhatt | Ruchi Bhatt, Shreya Bansal, Amanpreet Chander, Rupinder Kaur, Malya
Singh, Mohan Kankanhalli, Abdulmotaleb El Saddik, Mukesh Kumar Saini | GroMo: Plant Growth Modeling with Multiview Images | 7 pages, 5 Figures, 3 Tables | null | null | null | cs.CV cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding plant growth dynamics is essential for applications in
agriculture and plant phenotyping. We present the Growth Modelling (GroMo)
challenge, which is designed for two primary tasks: (1) plant age prediction
and (2) leaf count estimation, both essential for crop monitoring and precision
agriculture. For this challenge, we introduce GroMo25, a dataset with images of
four crops: radish, okra, wheat, and mustard. Each crop consists of multiple
plants (p1, p2, ..., pn) captured over different days (d1, d2, ..., dm) and
categorized into five levels (L1, L2, L3, L4, L5). Each plant is captured from
24 different angles with a 15-degree gap between images. Participants are
required to perform both tasks for all four crops with these multiview images.
We proposed a Multiview Vision Transformer (MVVT) model for the GroMo challenge
and evaluated the crop-wise performance on GroMo25. MVVT reports an average MAE
of 7.74 for age prediction and an MAE of 5.52 for leaf count. The GroMo
Challenge aims to advance plant phenotyping research by encouraging innovative
solutions for tracking and predicting plant growth. The GitHub repository is
publicly available at
https://github.com/mriglab/GroMo-Plant-Growth-Modeling-with-Multiview-Images.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:23:16 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Bhatt",
"Ruchi",
""
],
[
"Bansal",
"Shreya",
""
],
[
"Chander",
"Amanpreet",
""
],
[
"Kaur",
"Rupinder",
""
],
[
"Singh",
"Malya",
""
],
[
"Kankanhalli",
"Mohan",
""
],
[
"Saddik",
"Abdulmotaleb El",
""
],
[
"Saini",
"Mukesh Kumar",
""
]
]
| TITLE: GroMo: Plant Growth Modeling with Multiview Images
ABSTRACT: Understanding plant growth dynamics is essential for applications in
agriculture and plant phenotyping. We present the Growth Modelling (GroMo)
challenge, which is designed for two primary tasks: (1) plant age prediction
and (2) leaf count estimation, both essential for crop monitoring and precision
agriculture. For this challenge, we introduce GroMo25, a dataset with images of
four crops: radish, okra, wheat, and mustard. Each crop consists of multiple
plants (p1, p2, ..., pn) captured over different days (d1, d2, ..., dm) and
categorized into five levels (L1, L2, L3, L4, L5). Each plant is captured from
24 different angles with a 15-degree gap between images. Participants are
required to perform both tasks for all four crops with these multiview images.
We proposed a Multiview Vision Transformer (MVVT) model for the GroMo challenge
and evaluated the crop-wise performance on GroMo25. MVVT reports an average MAE
of 7.74 for age prediction and an MAE of 5.52 for leaf count. The GroMo
Challenge aims to advance plant phenotyping research by encouraging innovative
solutions for tracking and predicting plant growth. The GitHub repository is
publicly available at
https://github.com/mriglab/GroMo-Plant-Growth-Modeling-with-Multiview-Images.
| new_dataset | 0.959724 |
2503.06611 | Raghvendra Cowlagi | Alexandra E. Ballentine and Raghvendra V. Cowlagi | Inverse Reinforcement Learning for Minimum-Exposure Paths in
Spatiotemporally Varying Scalar Fields | Joint submission to MECC-JAVS 2025 | null | null | null | cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Performance and reliability analyses of autonomous vehicles (AVs) can benefit
from tools that ``amplify'' small datasets to synthesize larger volumes of
plausible samples of the AV's behavior. We consider a specific instance of this
data synthesis problem that addresses minimizing the AV's exposure to adverse
environmental conditions during travel to a fixed goal location. The
environment is characterized by a threat field, which is a strictly positive
scalar field with higher intensities corresponding to hazardous and unfavorable
conditions for the AV. We address the problem of synthesizing datasets of
minimum exposure paths that resemble a training dataset of such paths. The main
contribution of this paper is an inverse reinforcement learning (IRL) model to
solve this problem. We consider time-invariant (static) as well as time-varying
(dynamic) threat fields. We find that the proposed IRL model provides excellent
performance in synthesizing paths from initial conditions not seen in the
training dataset, when the threat field is the same as that used for training.
Furthermore, we evaluate model performance on unseen threat fields and find low
error in that case as well. Finally, we demonstrate the model's ability to
synthesize distinct datasets when trained on different datasets with distinct
characteristics.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:30:11 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Ballentine",
"Alexandra E.",
""
],
[
"Cowlagi",
"Raghvendra V.",
""
]
]
| TITLE: Inverse Reinforcement Learning for Minimum-Exposure Paths in
Spatiotemporally Varying Scalar Fields
ABSTRACT: Performance and reliability analyses of autonomous vehicles (AVs) can benefit
from tools that ``amplify'' small datasets to synthesize larger volumes of
plausible samples of the AV's behavior. We consider a specific instance of this
data synthesis problem that addresses minimizing the AV's exposure to adverse
environmental conditions during travel to a fixed goal location. The
environment is characterized by a threat field, which is a strictly positive
scalar field with higher intensities corresponding to hazardous and unfavorable
conditions for the AV. We address the problem of synthesizing datasets of
minimum exposure paths that resemble a training dataset of such paths. The main
contribution of this paper is an inverse reinforcement learning (IRL) model to
solve this problem. We consider time-invariant (static) as well as time-varying
(dynamic) threat fields. We find that the proposed IRL model provides excellent
performance in synthesizing paths from initial conditions not seen in the
training dataset, when the threat field is the same as that used for training.
Furthermore, we evaluate model performance on unseen threat fields and find low
error in that case as well. Finally, we demonstrate the model's ability to
synthesize distinct datasets when trained on different datasets with distinct
characteristics.
| no_new_dataset | 0.946349 |
2503.06614 | Qian Zeng | Qian Zeng, Xin Lin, Jingyi Gao and Yang Yu | Using Subgraph GNNs for Node Classification:an Overlooked Potential
Approach | 16 pages | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Previous studies have demonstrated the strong performance of Graph Neural
Networks (GNNs) in node classification. However, most existing GNNs adopt a
node-centric perspective and rely on global message passing, leading to high
computational and memory costs that hinder scalability. To mitigate these
challenges, subgraph-based methods have been introduced, leveraging local
subgraphs as approximations of full computational trees. While this approach
improves efficiency, it often suffers from performance degradation due to the
loss of global contextual information, limiting its effectiveness compared to
global GNNs. To address this trade-off between scalability and classification
accuracy, we reformulate the node classification task as a subgraph
classification problem and propose SubGND (Subgraph GNN for NoDe). This
framework introduces a differentiated zero-padding strategy and an Ego-Alter
subgraph representation method to resolve label conflicts while incorporating
an Adaptive Feature Scaling Mechanism to dynamically adjust feature
contributions based on dataset-specific dependencies. Experimental results on
six benchmark datasets demonstrate that SubGND achieves performance comparable
to or surpassing global message-passing GNNs, particularly in heterophilic
settings, highlighting its effectiveness and scalability as a promising
solution for node classification.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:37:38 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zeng",
"Qian",
""
],
[
"Lin",
"Xin",
""
],
[
"Gao",
"Jingyi",
""
],
[
"Yu",
"Yang",
""
]
]
| TITLE: Using Subgraph GNNs for Node Classification:an Overlooked Potential
Approach
ABSTRACT: Previous studies have demonstrated the strong performance of Graph Neural
Networks (GNNs) in node classification. However, most existing GNNs adopt a
node-centric perspective and rely on global message passing, leading to high
computational and memory costs that hinder scalability. To mitigate these
challenges, subgraph-based methods have been introduced, leveraging local
subgraphs as approximations of full computational trees. While this approach
improves efficiency, it often suffers from performance degradation due to the
loss of global contextual information, limiting its effectiveness compared to
global GNNs. To address this trade-off between scalability and classification
accuracy, we reformulate the node classification task as a subgraph
classification problem and propose SubGND (Subgraph GNN for NoDe). This
framework introduces a differentiated zero-padding strategy and an Ego-Alter
subgraph representation method to resolve label conflicts while incorporating
an Adaptive Feature Scaling Mechanism to dynamically adjust feature
contributions based on dataset-specific dependencies. Experimental results on
six benchmark datasets demonstrate that SubGND achieves performance comparable
to or surpassing global message-passing GNNs, particularly in heterophilic
settings, highlighting its effectiveness and scalability as a promising
solution for node classification.
| no_new_dataset | 0.943712 |
2503.06623 | Feng Liu | Sijie Zhao, Feng Liu, Xueliang Zhang, Hao Chen, Tao Han, Junchao Gong,
Ran Tao, Pengfeng Xiao, Lei Bai, Wanli Ouyang | Transforming Weather Data from Pixel to Latent Space | 8 pages, 6 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing impact of climate change and extreme weather events has
spurred growing interest in deep learning for weather research. However,
existing studies often rely on weather data in pixel space, which presents
several challenges such as smooth outputs in model outputs, limited
applicability to a single pressure-variable subset (PVS), and high data storage
and computational costs. To address these challenges, we propose a novel
Weather Latent Autoencoder (WLA) that transforms weather data from pixel space
to latent space, enabling efficient weather task modeling. By decoupling
weather reconstruction from downstream tasks, WLA improves the accuracy and
sharpness of weather task model results. The incorporated Pressure-Variable
Unified Module transforms multiple PVS into a unified representation, enhancing
the adaptability of the model in multiple weather scenarios. Furthermore,
weather tasks can be performed in a low-storage latent space of WLA rather than
a high-storage pixel space, thus significantly reducing data storage and
computational costs. Through extensive experimentation, we demonstrate its
superior compression and reconstruction performance, enabling the creation of
the ERA5-latent dataset with unified representations of multiple PVS from ERA5
data. The compressed full PVS in the ERA5-latent dataset reduces the original
244.34 TB of data to 0.43 TB. The downstream task further demonstrates that
task models can apply to multiple PVS with low data costs in latent space and
achieve superior performance compared to models in pixel space. Code,
ERA5-latent data, and pre-trained models are available at
https://anonymous.4open.science/r/Weather-Latent-Autoencoder-8467.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:55:33 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhao",
"Sijie",
""
],
[
"Liu",
"Feng",
""
],
[
"Zhang",
"Xueliang",
""
],
[
"Chen",
"Hao",
""
],
[
"Han",
"Tao",
""
],
[
"Gong",
"Junchao",
""
],
[
"Tao",
"Ran",
""
],
[
"Xiao",
"Pengfeng",
""
],
[
"Bai",
"Lei",
""
],
[
"Ouyang",
"Wanli",
""
]
]
| TITLE: Transforming Weather Data from Pixel to Latent Space
ABSTRACT: The increasing impact of climate change and extreme weather events has
spurred growing interest in deep learning for weather research. However,
existing studies often rely on weather data in pixel space, which presents
several challenges such as smooth outputs in model outputs, limited
applicability to a single pressure-variable subset (PVS), and high data storage
and computational costs. To address these challenges, we propose a novel
Weather Latent Autoencoder (WLA) that transforms weather data from pixel space
to latent space, enabling efficient weather task modeling. By decoupling
weather reconstruction from downstream tasks, WLA improves the accuracy and
sharpness of weather task model results. The incorporated Pressure-Variable
Unified Module transforms multiple PVS into a unified representation, enhancing
the adaptability of the model in multiple weather scenarios. Furthermore,
weather tasks can be performed in a low-storage latent space of WLA rather than
a high-storage pixel space, thus significantly reducing data storage and
computational costs. Through extensive experimentation, we demonstrate its
superior compression and reconstruction performance, enabling the creation of
the ERA5-latent dataset with unified representations of multiple PVS from ERA5
data. The compressed full PVS in the ERA5-latent dataset reduces the original
244.34 TB of data to 0.43 TB. The downstream task further demonstrates that
task models can apply to multiple PVS with low data costs in latent space and
achieve superior performance compared to models in pixel space. Code,
ERA5-latent data, and pre-trained models are available at
https://anonymous.4open.science/r/Weather-Latent-Autoencoder-8467.
| no_new_dataset | 0.901704 |
2503.06624 | Nankai Lin | Meiyu Zeng, Xingming Liao, Canyu Chen, Nankai Lin, Zhuowei Wang, Chong
Chen, Aimin Yang | Chameleon: On the Scene Diversity and Domain Variety of AI-Generated
Videos Detection | 17 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artificial intelligence generated content (AIGC), known as DeepFakes, has
emerged as a growing concern because it is being utilized as a tool for
spreading disinformation. While much research exists on identifying
AI-generated text and images, research on detecting AI-generated videos is
limited. Existing datasets for AI-generated videos detection exhibit
limitations in terms of diversity, complexity, and realism. To address these
issues, this paper focuses on AI-generated videos detection and constructs a
diverse dataset named Chameleon. We generate videos through multiple generation
tools and various real video sources. At the same time, we preserve the videos'
real-world complexity, including scene switches and dynamic perspective
changes, and expand beyond face-centered detection to include human actions and
environment generation. Our work bridges the gap between AI-generated dataset
construction and real-world forensic needs, offering a valuable benchmark to
counteract the evolving threats of AI-generated content.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 13:58:43 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zeng",
"Meiyu",
""
],
[
"Liao",
"Xingming",
""
],
[
"Chen",
"Canyu",
""
],
[
"Lin",
"Nankai",
""
],
[
"Wang",
"Zhuowei",
""
],
[
"Chen",
"Chong",
""
],
[
"Yang",
"Aimin",
""
]
]
| TITLE: Chameleon: On the Scene Diversity and Domain Variety of AI-Generated
Videos Detection
ABSTRACT: Artificial intelligence generated content (AIGC), known as DeepFakes, has
emerged as a growing concern because it is being utilized as a tool for
spreading disinformation. While much research exists on identifying
AI-generated text and images, research on detecting AI-generated videos is
limited. Existing datasets for AI-generated videos detection exhibit
limitations in terms of diversity, complexity, and realism. To address these
issues, this paper focuses on AI-generated videos detection and constructs a
diverse dataset named Chameleon. We generate videos through multiple generation
tools and various real video sources. At the same time, we preserve the videos'
real-world complexity, including scene switches and dynamic perspective
changes, and expand beyond face-centered detection to include human actions and
environment generation. Our work bridges the gap between AI-generated dataset
construction and real-world forensic needs, offering a valuable benchmark to
counteract the evolving threats of AI-generated content.
| new_dataset | 0.949059 |
2503.06629 | Tomasz Kryjak | Hiroshi Nakano and Krzysztof Blachut and Kamil Jeziorek and Piotr
Wzorek and Manon Dampfhoffer and Thomas Mesquida and Hiroaki Nishi and Tomasz
Kryjak and Thomas Dalgaty | Hardware-Accelerated Event-Graph Neural Networks for Low-Latency
Time-Series Classification on SoC FPGA | Paper accepted for the 21st International Symposium on Applied
Reconfigurable Computing ARC 2025, Sevilla, Spain, April 9-11, 2025 | null | null | null | cs.LG cs.AI eess.SP | http://creativecommons.org/licenses/by/4.0/ | As the quantities of data recorded by embedded edge sensors grow, so too does
the need for intelligent local processing. Such data often comes in the form of
time-series signals, based on which real-time predictions can be made locally
using an AI model. However, a hardware-software approach capable of making
low-latency predictions with low power consumption is required. In this paper,
we present a hardware implementation of an event-graph neural network for
time-series classification. We leverage an artificial cochlea model to convert
the input time-series signals into a sparse event-data format that allows the
event-graph to drastically reduce the number of calculations relative to other
AI methods. We implemented the design on a SoC FPGA and applied it to the
real-time processing of the Spiking Heidelberg Digits (SHD) dataset to
benchmark our approach against competitive solutions. Our method achieves a
floating-point accuracy of 92.7% on the SHD dataset for the base model, which
is only 2.4% and 2% less than the state-of-the-art models with over 10% and 67%
fewer model parameters, respectively. It also outperforms FPGA-based spiking
neural network implementations by 19.3% and 4.5%, achieving 92.3% accuracy for
the quantised model while using fewer computational resources and reducing
latency.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 14:08:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Nakano",
"Hiroshi",
""
],
[
"Blachut",
"Krzysztof",
""
],
[
"Jeziorek",
"Kamil",
""
],
[
"Wzorek",
"Piotr",
""
],
[
"Dampfhoffer",
"Manon",
""
],
[
"Mesquida",
"Thomas",
""
],
[
"Nishi",
"Hiroaki",
""
],
[
"Kryjak",
"Tomasz",
""
],
[
"Dalgaty",
"Thomas",
""
]
]
| TITLE: Hardware-Accelerated Event-Graph Neural Networks for Low-Latency
Time-Series Classification on SoC FPGA
ABSTRACT: As the quantities of data recorded by embedded edge sensors grow, so too does
the need for intelligent local processing. Such data often comes in the form of
time-series signals, based on which real-time predictions can be made locally
using an AI model. However, a hardware-software approach capable of making
low-latency predictions with low power consumption is required. In this paper,
we present a hardware implementation of an event-graph neural network for
time-series classification. We leverage an artificial cochlea model to convert
the input time-series signals into a sparse event-data format that allows the
event-graph to drastically reduce the number of calculations relative to other
AI methods. We implemented the design on a SoC FPGA and applied it to the
real-time processing of the Spiking Heidelberg Digits (SHD) dataset to
benchmark our approach against competitive solutions. Our method achieves a
floating-point accuracy of 92.7% on the SHD dataset for the base model, which
is only 2.4% and 2% less than the state-of-the-art models with over 10% and 67%
fewer model parameters, respectively. It also outperforms FPGA-based spiking
neural network implementations by 19.3% and 4.5%, achieving 92.3% accuracy for
the quantised model while using fewer computational resources and reducing
latency.
| no_new_dataset | 0.947575 |
2503.06633 | Yu Zhou | Yu Zhou and Bingyan Liu | BTFL: A Bayesian-based Test-Time Generalization Method for Internal and
External Data Distributions in Federated learning | accepted as KDD 2025 research track paper | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated Learning (FL) enables multiple clients to collaboratively develop a
global model while maintaining data privacy. However, online FL deployment
faces challenges due to distribution shifts and evolving test samples.
Personalized Federated Learning (PFL) tailors the global model to individual
client distributions, but struggles with Out-Of-Distribution (OOD) samples
during testing, leading to performance degradation. In real-world scenarios,
balancing personalization and generalization during online testing is crucial
and existing methods primarily focus on training-phase generalization. To
address the test-time trade-off, we introduce a new scenario: Test-time
Generalization for Internal and External Distributions in Federated Learning
(TGFL), which evaluates adaptability under Internal Distribution (IND) and
External Distribution (EXD). We propose BTFL, a Bayesian-based test-time
generalization method for TGFL, which balances generalization and
personalization at the sample level during testing. BTFL employs a two-head
architecture to store local and global knowledge, interpolating predictions via
a dual-Bayesian framework that considers both historical test data and current
sample characteristics with theoretical guarantee and faster speed. Our
experiments demonstrate that BTFL achieves improved performance across various
datasets and models with less time cost. The source codes are made publicly
available at https://github.com/ZhouYuCS/BTFL .
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 14:16:34 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhou",
"Yu",
""
],
[
"Liu",
"Bingyan",
""
]
]
| TITLE: BTFL: A Bayesian-based Test-Time Generalization Method for Internal and
External Data Distributions in Federated learning
ABSTRACT: Federated Learning (FL) enables multiple clients to collaboratively develop a
global model while maintaining data privacy. However, online FL deployment
faces challenges due to distribution shifts and evolving test samples.
Personalized Federated Learning (PFL) tailors the global model to individual
client distributions, but struggles with Out-Of-Distribution (OOD) samples
during testing, leading to performance degradation. In real-world scenarios,
balancing personalization and generalization during online testing is crucial
and existing methods primarily focus on training-phase generalization. To
address the test-time trade-off, we introduce a new scenario: Test-time
Generalization for Internal and External Distributions in Federated Learning
(TGFL), which evaluates adaptability under Internal Distribution (IND) and
External Distribution (EXD). We propose BTFL, a Bayesian-based test-time
generalization method for TGFL, which balances generalization and
personalization at the sample level during testing. BTFL employs a two-head
architecture to store local and global knowledge, interpolating predictions via
a dual-Bayesian framework that considers both historical test data and current
sample characteristics with theoretical guarantee and faster speed. Our
experiments demonstrate that BTFL achieves improved performance across various
datasets and models with less time cost. The source codes are made publicly
available at https://github.com/ZhouYuCS/BTFL .
| no_new_dataset | 0.947624 |
2503.06637 | Lei Shi | Lei Shi, Andreas Bulling | CLAD: Constrained Latent Action Diffusion for Vision-Language Procedure
Planning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We propose CLAD -- a Constrained Latent Action Diffusion model for
vision-language procedure planning in instructional videos. Procedure planning
is the challenging task of predicting intermediate actions given a visual
observation of a start and a goal state. However, future interactive AI systems
must also be able to plan procedures using multi-modal input, e.g., where
visual observations are augmented with language descriptions. To tackle this
vision-language procedure planning task, our method uses a Variational
Autoencoder (VAE) to learn the latent representation of actions and
observations as constraints and integrate them into the diffusion process. This
approach exploits that the latent space of diffusion models already has
semantics that can be used. We use the latent constraints to steer the
diffusion model to better generate actions. We report extensive experiments on
the popular CrossTask, Coin, and NIV datasets and show that our method
outperforms state-of-the-art methods by a large margin. By evaluating ablated
versions of our method, we further show that the proposed integration of the
action and observation representations learnt in the VAE latent space is key to
these performance improvements.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 14:31:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Shi",
"Lei",
""
],
[
"Bulling",
"Andreas",
""
]
]
| TITLE: CLAD: Constrained Latent Action Diffusion for Vision-Language Procedure
Planning
ABSTRACT: We propose CLAD -- a Constrained Latent Action Diffusion model for
vision-language procedure planning in instructional videos. Procedure planning
is the challenging task of predicting intermediate actions given a visual
observation of a start and a goal state. However, future interactive AI systems
must also be able to plan procedures using multi-modal input, e.g., where
visual observations are augmented with language descriptions. To tackle this
vision-language procedure planning task, our method uses a Variational
Autoencoder (VAE) to learn the latent representation of actions and
observations as constraints and integrate them into the diffusion process. This
approach exploits that the latent space of diffusion models already has
semantics that can be used. We use the latent constraints to steer the
diffusion model to better generate actions. We report extensive experiments on
the popular CrossTask, Coin, and NIV datasets and show that our method
outperforms state-of-the-art methods by a large margin. By evaluating ablated
versions of our method, we further show that the proposed integration of the
action and observation representations learnt in the VAE latent space is key to
these performance improvements.
| no_new_dataset | 0.94366 |
2503.06647 | Hassan Kazemi Tehrani | Hassan Kazemi Tehrani, Jun Cai, Abbas Yekanlou, and Sylvia Santosa | Personalized Class Incremental Context-Aware Food Classification for
Food Intake Monitoring Systems | null | null | null | null | cs.CV cs.CE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate food intake monitoring is crucial for maintaining a healthy diet and
preventing nutrition-related diseases. With the diverse range of foods consumed
across various cultures, classic food classification models have limitations
due to their reliance on fixed-sized food datasets. Studies show that people
consume only a small range of foods across the existing ones, each consuming a
unique set of foods. Existing class-incremental models have low accuracy for
the new classes and lack personalization. This paper introduces a personalized,
class-incremental food classification model designed to overcome these
challenges and improve the performance of food intake monitoring systems. Our
approach adapts itself to the new array of food classes, maintaining
applicability and accuracy, both for new and existing classes by using
personalization. Our model's primary focus is personalization, which improves
classification accuracy by prioritizing a subset of foods based on an
individual's eating habits, including meal frequency, times, and locations. A
modified version of DSN is utilized to expand on the appearance of new food
classes. Additionally, we propose a comprehensive framework that integrates
this model into a food intake monitoring system. This system analyzes meal
images provided by users, makes use of a smart scale to estimate food weight,
utilizes a nutrient content database to calculate the amount of each
macro-nutrient, and creates a dietary user profile through a mobile
application. Finally, experimental evaluations on two new benchmark datasets
FOOD101-Personal and VFN-Personal, personalized versions of well-known datasets
for food classification, are conducted to demonstrate the effectiveness of our
model in improving the classification accuracy of both new and existing
classes, addressing the limitations of both conventional and class-incremental
food classification models.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 14:50:56 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Tehrani",
"Hassan Kazemi",
""
],
[
"Cai",
"Jun",
""
],
[
"Yekanlou",
"Abbas",
""
],
[
"Santosa",
"Sylvia",
""
]
]
| TITLE: Personalized Class Incremental Context-Aware Food Classification for
Food Intake Monitoring Systems
ABSTRACT: Accurate food intake monitoring is crucial for maintaining a healthy diet and
preventing nutrition-related diseases. With the diverse range of foods consumed
across various cultures, classic food classification models have limitations
due to their reliance on fixed-sized food datasets. Studies show that people
consume only a small range of foods across the existing ones, each consuming a
unique set of foods. Existing class-incremental models have low accuracy for
the new classes and lack personalization. This paper introduces a personalized,
class-incremental food classification model designed to overcome these
challenges and improve the performance of food intake monitoring systems. Our
approach adapts itself to the new array of food classes, maintaining
applicability and accuracy, both for new and existing classes by using
personalization. Our model's primary focus is personalization, which improves
classification accuracy by prioritizing a subset of foods based on an
individual's eating habits, including meal frequency, times, and locations. A
modified version of DSN is utilized to expand on the appearance of new food
classes. Additionally, we propose a comprehensive framework that integrates
this model into a food intake monitoring system. This system analyzes meal
images provided by users, makes use of a smart scale to estimate food weight,
utilizes a nutrient content database to calculate the amount of each
macro-nutrient, and creates a dietary user profile through a mobile
application. Finally, experimental evaluations on two new benchmark datasets
FOOD101-Personal and VFN-Personal, personalized versions of well-known datasets
for food classification, are conducted to demonstrate the effectiveness of our
model in improving the classification accuracy of both new and existing
classes, addressing the limitations of both conventional and class-incremental
food classification models.
| no_new_dataset | 0.956063 |
2503.06648 | Hender Lin | Hender Lin | Enhancing NLP Robustness and Generalization through LLM-Generated
Contrast Sets: A Scalable Framework for Systematic Evaluation and Adversarial
Training | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Standard NLP benchmarks often fail to capture vulnerabilities stemming from
dataset artifacts and spurious correlations. Contrast sets address this gap by
challenging models near decision boundaries but are traditionally
labor-intensive to create and limited in diversity. This study leverages large
language models to automate the generation of diverse contrast sets. Using the
SNLI dataset, we created a 3,000-example contrast set to evaluate and improve
model robustness. Fine-tuning on these contrast sets enhanced performance on
systematically perturbed examples, maintained standard test accuracy, and
modestly improved generalization to novel perturbations. This automated
approach offers a scalable solution for evaluating and improving NLP models,
addressing systematic generalization challenges, and advancing robustness in
real-world applications.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 14:52:53 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Lin",
"Hender",
""
]
]
| TITLE: Enhancing NLP Robustness and Generalization through LLM-Generated
Contrast Sets: A Scalable Framework for Systematic Evaluation and Adversarial
Training
ABSTRACT: Standard NLP benchmarks often fail to capture vulnerabilities stemming from
dataset artifacts and spurious correlations. Contrast sets address this gap by
challenging models near decision boundaries but are traditionally
labor-intensive to create and limited in diversity. This study leverages large
language models to automate the generation of diverse contrast sets. Using the
SNLI dataset, we created a 3,000-example contrast set to evaluate and improve
model robustness. Fine-tuning on these contrast sets enhanced performance on
systematically perturbed examples, maintained standard test accuracy, and
modestly improved generalization to novel perturbations. This automated
approach offers a scalable solution for evaluating and improving NLP models,
addressing systematic generalization challenges, and advancing robustness in
real-world applications.
| no_new_dataset | 0.926503 |
2503.06651 | Tengjiao Wang | Tengjiao Wang, Zhenyu Kang, Ting Li, Zhihui Chen, Shaobo Wang, Yingpei
Lin, Yan Wang, and Yichuan Yu | Electromagnetic Information Theory: Fundamentals, Paradigm Shifts, and
Applications | null | null | null | null | cs.IT eess.SP math.IT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper explores the emerging research direction of electromagnetic
information theory (EIT), which aims to integrate traditional Shannon-based
methodologies with physical consistency, particularly the electromagnetic
properties of communication channels. We propose an EIT-based multiple-input
multiple-output (MIMO) paradigm that enhances conventional spatially-discrete
MIMO models by incorporating the concepts of electromagnetic (EM) precoding and
EM combining. This approach aims to improve the modeling of next-generation
systems while remaining consistent with Shannon's theoretical foundations. We
explore typical EIT applications, such as densely spaced MIMO, near-field
communications, and tri-polarized antennas, and analyze their channel
characteristics through theoretical simulations and measured datasets. The
paper also discusses critical research challenges and opportunities for EIT
applications from an industrial perspective, emphasizing the field's potential
for practical applications.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 15:05:40 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Wang",
"Tengjiao",
""
],
[
"Kang",
"Zhenyu",
""
],
[
"Li",
"Ting",
""
],
[
"Chen",
"Zhihui",
""
],
[
"Wang",
"Shaobo",
""
],
[
"Lin",
"Yingpei",
""
],
[
"Wang",
"Yan",
""
],
[
"Yu",
"Yichuan",
""
]
]
| TITLE: Electromagnetic Information Theory: Fundamentals, Paradigm Shifts, and
Applications
ABSTRACT: This paper explores the emerging research direction of electromagnetic
information theory (EIT), which aims to integrate traditional Shannon-based
methodologies with physical consistency, particularly the electromagnetic
properties of communication channels. We propose an EIT-based multiple-input
multiple-output (MIMO) paradigm that enhances conventional spatially-discrete
MIMO models by incorporating the concepts of electromagnetic (EM) precoding and
EM combining. This approach aims to improve the modeling of next-generation
systems while remaining consistent with Shannon's theoretical foundations. We
explore typical EIT applications, such as densely spaced MIMO, near-field
communications, and tri-polarized antennas, and analyze their channel
characteristics through theoretical simulations and measured datasets. The
paper also discusses critical research challenges and opportunities for EIT
applications from an industrial perspective, emphasizing the field's potential
for practical applications.
| no_new_dataset | 0.945248 |
2503.06664 | Tommaso Bendinelli | Tommaso Bendinelli, Artur Dox, Christian Holz | Exploring LLM Agents for Cleaning Tabular Machine Learning Datasets | 14 pages, 1 main figure, 3 plots, Published at ICLR 2025 Workshop on
Foundation Models in the Wild | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | High-quality, error-free datasets are a key ingredient in building reliable,
accurate, and unbiased machine learning (ML) models. However, real world
datasets often suffer from errors due to sensor malfunctions, data entry
mistakes, or improper data integration across multiple sources that can
severely degrade model performance. Detecting and correcting these issues
typically require tailor-made solutions and demand extensive domain expertise.
Consequently, automation is challenging, rendering the process labor-intensive
and tedious. In this study, we investigate whether Large Language Models (LLMs)
can help alleviate the burden of manual data cleaning. We set up an experiment
in which an LLM, paired with Python, is tasked with cleaning the training
dataset to improve the performance of a learning algorithm without having the
ability to modify the training pipeline or perform any feature engineering. We
run this experiment on multiple Kaggle datasets that have been intentionally
corrupted with errors. Our results show that LLMs can identify and correct
erroneous entries, such as illogical values or outlier, by leveraging
contextual information from other features within the same row, as well as
feedback from previous iterations. However, they struggle to detect more
complex errors that require understanding data distribution across multiple
rows, such as trends and biases.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 15:29:46 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Bendinelli",
"Tommaso",
""
],
[
"Dox",
"Artur",
""
],
[
"Holz",
"Christian",
""
]
]
| TITLE: Exploring LLM Agents for Cleaning Tabular Machine Learning Datasets
ABSTRACT: High-quality, error-free datasets are a key ingredient in building reliable,
accurate, and unbiased machine learning (ML) models. However, real world
datasets often suffer from errors due to sensor malfunctions, data entry
mistakes, or improper data integration across multiple sources that can
severely degrade model performance. Detecting and correcting these issues
typically require tailor-made solutions and demand extensive domain expertise.
Consequently, automation is challenging, rendering the process labor-intensive
and tedious. In this study, we investigate whether Large Language Models (LLMs)
can help alleviate the burden of manual data cleaning. We set up an experiment
in which an LLM, paired with Python, is tasked with cleaning the training
dataset to improve the performance of a learning algorithm without having the
ability to modify the training pipeline or perform any feature engineering. We
run this experiment on multiple Kaggle datasets that have been intentionally
corrupted with errors. Our results show that LLMs can identify and correct
erroneous entries, such as illogical values or outlier, by leveraging
contextual information from other features within the same row, as well as
feedback from previous iterations. However, they struggle to detect more
complex errors that require understanding data distribution across multiple
rows, such as trends and biases.
| no_new_dataset | 0.941654 |
2503.06678 | Hantao Zhou | Hantao Zhou, Rui Yang, Longxiang Tang, Guanyi Qin, Yan Zhang, Runze
Hu, Xiu Li | Gamma: Toward Generic Image Assessment with Mixture of Assessment
Experts | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image assessment aims to evaluate the quality and aesthetics of images and
has been applied across various scenarios, such as natural and AIGC scenes.
Existing methods mostly address these sub-tasks or scenes individually. While
some works attempt to develop unified image assessment models, they have
struggled to achieve satisfactory performance or cover a broad spectrum of
assessment scenarios. In this paper, we present \textbf{Gamma}, a
\textbf{G}eneric im\textbf{A}ge assess\textbf{M}ent model using
\textbf{M}ixture of \textbf{A}ssessment Experts, which can effectively assess
images from diverse scenes through mixed-dataset training. Achieving unified
training in image assessment presents significant challenges due to annotation
biases across different datasets. To address this issue, we first propose a
Mixture of Assessment Experts (MoAE) module, which employs shared and adaptive
experts to dynamically learn common and specific knowledge for different
datasets, respectively. In addition, we introduce a Scene-based Differential
Prompt (SDP) strategy, which uses scene-specific prompts to provide prior
knowledge and guidance during the learning process, further boosting adaptation
for various scenes. Our Gamma model is trained and evaluated on 12 datasets
spanning 6 image assessment scenarios. Extensive experiments show that our
unified Gamma outperforms other state-of-the-art mixed-training methods by
significant margins while covering more scenes. Code:
https://github.com/zht8506/Gamma.
| [
{
"version": "v1",
"created": "Sun, 9 Mar 2025 16:07:58 GMT"
}
]
| 2025-03-11T00:00:00 | [
[
"Zhou",
"Hantao",
""
],
[
"Yang",
"Rui",
""
],
[
"Tang",
"Longxiang",
""
],
[
"Qin",
"Guanyi",
""
],
[
"Zhang",
"Yan",
""
],
[
"Hu",
"Runze",
""
],
[
"Li",
"Xiu",
""
]
]
| TITLE: Gamma: Toward Generic Image Assessment with Mixture of Assessment
Experts
ABSTRACT: Image assessment aims to evaluate the quality and aesthetics of images and
has been applied across various scenarios, such as natural and AIGC scenes.
Existing methods mostly address these sub-tasks or scenes individually. While
some works attempt to develop unified image assessment models, they have
struggled to achieve satisfactory performance or cover a broad spectrum of
assessment scenarios. In this paper, we present \textbf{Gamma}, a
\textbf{G}eneric im\textbf{A}ge assess\textbf{M}ent model using
\textbf{M}ixture of \textbf{A}ssessment Experts, which can effectively assess
images from diverse scenes through mixed-dataset training. Achieving unified
training in image assessment presents significant challenges due to annotation
biases across different datasets. To address this issue, we first propose a
Mixture of Assessment Experts (MoAE) module, which employs shared and adaptive
experts to dynamically learn common and specific knowledge for different
datasets, respectively. In addition, we introduce a Scene-based Differential
Prompt (SDP) strategy, which uses scene-specific prompts to provide prior
knowledge and guidance during the learning process, further boosting adaptation
for various scenes. Our Gamma model is trained and evaluated on 12 datasets
spanning 6 image assessment scenarios. Extensive experiments show that our
unified Gamma outperforms other state-of-the-art mixed-training methods by
significant margins while covering more scenes. Code:
https://github.com/zht8506/Gamma.
| no_new_dataset | 0.941975 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.