Search is not available for this dataset
id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence | prompt
string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2411.10332 | Yongliang Wu | Yongliang Wu, Xinting Hu, Yuyang Sun, Yizhou Zhou, Wenbo Zhu, Fengyun
Rao, Bernt Schiele, Xu Yang | Number it: Temporal Grounding Videos like Flipping Manga | Accepted by CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video Large Language Models (Vid-LLMs) have made remarkable advancements in
comprehending video content for QA dialogue. However, they struggle to extend
this visual understanding to tasks requiring precise temporal localization,
known as Video Temporal Grounding (VTG). To address this gap, we introduce
Number-Prompt (NumPro), a novel method that empowers Vid-LLMs to bridge visual
comprehension with temporal grounding by adding unique numerical identifiers to
each video frame. Treating a video as a sequence of numbered frame images,
NumPro transforms VTG into an intuitive process: flipping through manga panels
in sequence. This allows Vid-LLMs to "read" event timelines, accurately linking
visual content with corresponding temporal information. Our experiments
demonstrate that NumPro significantly boosts VTG performance of top-tier
Vid-LLMs without additional computational cost. Furthermore, fine-tuning on a
NumPro-enhanced dataset defines a new state-of-the-art for VTG, surpassing
previous top-performing methods by up to 6.9\% in mIoU for moment retrieval and
8.5\% in mAP for highlight detection. The code will be available at
https://github.com/yongliang-wu/NumPro.
| [
{
"version": "v1",
"created": "Fri, 15 Nov 2024 16:32:34 GMT"
},
{
"version": "v2",
"created": "Thu, 28 Nov 2024 02:57:24 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 12:40:26 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Wu",
"Yongliang",
""
],
[
"Hu",
"Xinting",
""
],
[
"Sun",
"Yuyang",
""
],
[
"Zhou",
"Yizhou",
""
],
[
"Zhu",
"Wenbo",
""
],
[
"Rao",
"Fengyun",
""
],
[
"Schiele",
"Bernt",
""
],
[
"Yang",
"Xu",
""
]
] | TITLE: Number it: Temporal Grounding Videos like Flipping Manga
ABSTRACT: Video Large Language Models (Vid-LLMs) have made remarkable advancements in
comprehending video content for QA dialogue. However, they struggle to extend
this visual understanding to tasks requiring precise temporal localization,
known as Video Temporal Grounding (VTG). To address this gap, we introduce
Number-Prompt (NumPro), a novel method that empowers Vid-LLMs to bridge visual
comprehension with temporal grounding by adding unique numerical identifiers to
each video frame. Treating a video as a sequence of numbered frame images,
NumPro transforms VTG into an intuitive process: flipping through manga panels
in sequence. This allows Vid-LLMs to "read" event timelines, accurately linking
visual content with corresponding temporal information. Our experiments
demonstrate that NumPro significantly boosts VTG performance of top-tier
Vid-LLMs without additional computational cost. Furthermore, fine-tuning on a
NumPro-enhanced dataset defines a new state-of-the-art for VTG, surpassing
previous top-performing methods by up to 6.9\% in mIoU for moment retrieval and
8.5\% in mAP for highlight detection. The code will be available at
https://github.com/yongliang-wu/NumPro.
|
2411.12287 | Dongyoung Go | Dongyoung Go, Taesun Whang, Chanhee Lee, Hwa-Yeon Kim, Sunghoon Park,
Seunghwan Ji, Jinho Kim, Dongchan Kim, Young-Bum Kim | CUE-M: Contextual Understanding and Enhanced Search with Multimodal
Large Language Model | Preprint. Under review | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The integration of Retrieval-Augmented Generation (RAG) with Multimodal Large
Language Models (MLLMs) has revolutionized information retrieval and expanded
the practical applications of AI. However, current systems struggle in
accurately interpreting user intent, employing diverse retrieval strategies,
and effectively filtering unintended or inappropriate responses, limiting their
effectiveness. This paper introduces Contextual Understanding and Enhanced
Search with MLLM (CUE-M), a novel multimodal search framework that addresses
these challenges through a multi-stage pipeline comprising image context
enrichment, intent refinement, contextual query generation, external API
integration, and relevance-based filtering. CUE-M incorporates a robust
filtering pipeline combining image-based, text-based, and multimodal
classifiers, dynamically adapting to instance- and category-specific concern
defined by organizational policies. Extensive experiments on real-word datasets
and public benchmarks on knowledge-based VQA and safety demonstrated that CUE-M
outperforms baselines and establishes new state-of-the-art results, advancing
the capabilities of multimodal retrieval systems.
| [
{
"version": "v1",
"created": "Tue, 19 Nov 2024 07:16:48 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2024 05:43:58 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 00:37:43 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Go",
"Dongyoung",
""
],
[
"Whang",
"Taesun",
""
],
[
"Lee",
"Chanhee",
""
],
[
"Kim",
"Hwa-Yeon",
""
],
[
"Park",
"Sunghoon",
""
],
[
"Ji",
"Seunghwan",
""
],
[
"Kim",
"Jinho",
""
],
[
"Kim",
"Dongchan",
""
],
[
"Kim",
"Young-Bum",
""
]
] | TITLE: CUE-M: Contextual Understanding and Enhanced Search with Multimodal
Large Language Model
ABSTRACT: The integration of Retrieval-Augmented Generation (RAG) with Multimodal Large
Language Models (MLLMs) has revolutionized information retrieval and expanded
the practical applications of AI. However, current systems struggle in
accurately interpreting user intent, employing diverse retrieval strategies,
and effectively filtering unintended or inappropriate responses, limiting their
effectiveness. This paper introduces Contextual Understanding and Enhanced
Search with MLLM (CUE-M), a novel multimodal search framework that addresses
these challenges through a multi-stage pipeline comprising image context
enrichment, intent refinement, contextual query generation, external API
integration, and relevance-based filtering. CUE-M incorporates a robust
filtering pipeline combining image-based, text-based, and multimodal
classifiers, dynamically adapting to instance- and category-specific concern
defined by organizational policies. Extensive experiments on real-word datasets
and public benchmarks on knowledge-based VQA and safety demonstrated that CUE-M
outperforms baselines and establishes new state-of-the-art results, advancing
the capabilities of multimodal retrieval systems.
|
2411.13945 | Stein Stroobants | Stein Stroobants, Christophe de Wagter, Guido C.H.E. De Croon | Neuromorphic Attitude Estimation and Control | null | null | null | null | cs.RO cs.LG cs.NE | http://creativecommons.org/licenses/by/4.0/ | The real-world application of small drones is mostly hampered by energy
limitations. Neuromorphic computing promises extremely energy-efficient AI for
autonomous flight but is still challenging to train and deploy on real robots.
To reap the maximal benefits from neuromorphic computing, it is necessary to
perform all autonomy functions end-to-end on a single neuromorphic chip, from
low-level attitude control to high-level navigation. This research presents the
first neuromorphic control system using a spiking neural network (SNN) to
effectively map a drone's raw sensory input directly to motor commands. We
apply this method to low-level attitude estimation and control for a quadrotor,
deploying the SNN on a tiny Crazyflie. We propose a modular SNN, separately
training and then merging estimation and control sub-networks. The SNN is
trained with imitation learning, using a flight dataset of sensory-motor pairs.
Post-training, the network is deployed on the Crazyflie, issuing control
commands from sensor inputs at 500Hz. Furthermore, for the training procedure
we augmented training data by flying a controller with additional excitation
and time-shifting the target data to enhance the predictive capabilities of the
SNN. On the real drone, the perception-to-control SNN tracks attitude commands
with an average error of 3.0 degrees, compared to 2.7 degrees for the regular
flight stack. We also show the benefits of the proposed learning modifications
for reducing the average tracking error and reducing oscillations. Our work
shows the feasibility of performing neuromorphic end-to-end control, laying the
basis for highly energy-efficient and low-latency neuromorphic autopilots.
| [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 08:54:45 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 07:57:38 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Stroobants",
"Stein",
""
],
[
"de Wagter",
"Christophe",
""
],
[
"De Croon",
"Guido C. H. E.",
""
]
] | TITLE: Neuromorphic Attitude Estimation and Control
ABSTRACT: The real-world application of small drones is mostly hampered by energy
limitations. Neuromorphic computing promises extremely energy-efficient AI for
autonomous flight but is still challenging to train and deploy on real robots.
To reap the maximal benefits from neuromorphic computing, it is necessary to
perform all autonomy functions end-to-end on a single neuromorphic chip, from
low-level attitude control to high-level navigation. This research presents the
first neuromorphic control system using a spiking neural network (SNN) to
effectively map a drone's raw sensory input directly to motor commands. We
apply this method to low-level attitude estimation and control for a quadrotor,
deploying the SNN on a tiny Crazyflie. We propose a modular SNN, separately
training and then merging estimation and control sub-networks. The SNN is
trained with imitation learning, using a flight dataset of sensory-motor pairs.
Post-training, the network is deployed on the Crazyflie, issuing control
commands from sensor inputs at 500Hz. Furthermore, for the training procedure
we augmented training data by flying a controller with additional excitation
and time-shifting the target data to enhance the predictive capabilities of the
SNN. On the real drone, the perception-to-control SNN tracks attitude commands
with an average error of 3.0 degrees, compared to 2.7 degrees for the regular
flight stack. We also show the benefits of the proposed learning modifications
for reducing the average tracking error and reducing oscillations. Our work
shows the feasibility of performing neuromorphic end-to-end control, laying the
basis for highly energy-efficient and low-latency neuromorphic autopilots.
|
2411.16173 | Junho Kim | Junho Kim, Hyunjun Kim, Hosu Lee, Yong Man Ro | SALOVA: Segment-Augmented Long Video Assistant for Targeted Retrieval
and Routing in Long-Form Video Analysis | Project page: https://ivy-lvlm.github.io/SALOVA/ | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite advances in Large Multi-modal Models, applying them to long and
untrimmed video content remains challenging due to limitations in context
length and substantial memory overhead. These constraints often lead to
significant information loss and reduced relevance in the model responses. With
the exponential growth of video data across web platforms, understanding
long-form video is crucial for advancing generalized intelligence. In this
paper, we introduce SALOVA: Segment-Augmented LOng Video Assistant, a novel
video-LLM framework designed to enhance the comprehension of lengthy video
content through targeted retrieval process. We address two main challenges to
achieve it: (i) We present the SceneWalk dataset, a high-quality collection of
87.8K long videos, each densely captioned at the segment level to enable models
to capture scene continuity and maintain rich descriptive context. (ii) We
develop robust architectural designs integrating dynamic routing mechanism and
spatio-temporal projector to efficiently retrieve and process relevant video
segments based on user queries. Our framework mitigates the limitations of
current video-LMMs by allowing for precise identification and retrieval of
relevant video segments in response to queries, thereby improving the
contextual relevance of the generated responses. Through extensive experiments,
SALOVA demonstrates enhanced capability in processing complex long-form videos,
showing significant capability to maintain contextual integrity across extended
sequences.
| [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 08:04:47 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 10:44:15 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kim",
"Junho",
""
],
[
"Kim",
"Hyunjun",
""
],
[
"Lee",
"Hosu",
""
],
[
"Ro",
"Yong Man",
""
]
] | TITLE: SALOVA: Segment-Augmented Long Video Assistant for Targeted Retrieval
and Routing in Long-Form Video Analysis
ABSTRACT: Despite advances in Large Multi-modal Models, applying them to long and
untrimmed video content remains challenging due to limitations in context
length and substantial memory overhead. These constraints often lead to
significant information loss and reduced relevance in the model responses. With
the exponential growth of video data across web platforms, understanding
long-form video is crucial for advancing generalized intelligence. In this
paper, we introduce SALOVA: Segment-Augmented LOng Video Assistant, a novel
video-LLM framework designed to enhance the comprehension of lengthy video
content through targeted retrieval process. We address two main challenges to
achieve it: (i) We present the SceneWalk dataset, a high-quality collection of
87.8K long videos, each densely captioned at the segment level to enable models
to capture scene continuity and maintain rich descriptive context. (ii) We
develop robust architectural designs integrating dynamic routing mechanism and
spatio-temporal projector to efficiently retrieve and process relevant video
segments based on user queries. Our framework mitigates the limitations of
current video-LMMs by allowing for precise identification and retrieval of
relevant video segments in response to queries, thereby improving the
contextual relevance of the generated responses. Through extensive experiments,
SALOVA demonstrates enhanced capability in processing complex long-form videos,
showing significant capability to maintain contextual integrity across extended
sequences.
|
2411.18276 | Wenbo Cui | Wenbo Cui, Chengyang Zhao, Songlin Wei, Jiazhao Zhang, Haoran Geng,
Yaran Chen, Haoran Li, He Wang | GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic
Articulated Object Manipulation | Accepted by ICRA 2025. Project page:
https://pku-epic.github.io/GAPartManip/ | null | null | null | cs.RO cs.AI | http://creativecommons.org/licenses/by/4.0/ | Effectively manipulating articulated objects in household scenarios is a
crucial step toward achieving general embodied artificial intelligence.
Mainstream research in 3D vision has primarily focused on manipulation through
depth perception and pose detection. However, in real-world environments, these
methods often face challenges due to imperfect depth perception, such as with
transparent lids and reflective handles. Moreover, they generally lack the
diversity in part-based interactions required for flexible and adaptable
manipulation. To address these challenges, we introduced a large-scale
part-centric dataset for articulated object manipulation that features both
photo-realistic material randomization and detailed annotations of
part-oriented, scene-level actionable interaction poses. We evaluated the
effectiveness of our dataset by integrating it with several state-of-the-art
methods for depth estimation and interaction pose prediction. Additionally, we
proposed a novel modular framework that delivers superior and robust
performance for generalizable articulated object manipulation. Our extensive
experiments demonstrate that our dataset significantly improves the performance
of depth perception and actionable interaction pose prediction in both
simulation and real-world scenarios. More information and demos can be found
at: https://pku-epic.github.io/GAPartManip/.
| [
{
"version": "v1",
"created": "Wed, 27 Nov 2024 12:11:23 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 07:52:16 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Cui",
"Wenbo",
""
],
[
"Zhao",
"Chengyang",
""
],
[
"Wei",
"Songlin",
""
],
[
"Zhang",
"Jiazhao",
""
],
[
"Geng",
"Haoran",
""
],
[
"Chen",
"Yaran",
""
],
[
"Li",
"Haoran",
""
],
[
"Wang",
"He",
""
]
] | TITLE: GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic
Articulated Object Manipulation
ABSTRACT: Effectively manipulating articulated objects in household scenarios is a
crucial step toward achieving general embodied artificial intelligence.
Mainstream research in 3D vision has primarily focused on manipulation through
depth perception and pose detection. However, in real-world environments, these
methods often face challenges due to imperfect depth perception, such as with
transparent lids and reflective handles. Moreover, they generally lack the
diversity in part-based interactions required for flexible and adaptable
manipulation. To address these challenges, we introduced a large-scale
part-centric dataset for articulated object manipulation that features both
photo-realistic material randomization and detailed annotations of
part-oriented, scene-level actionable interaction poses. We evaluated the
effectiveness of our dataset by integrating it with several state-of-the-art
methods for depth estimation and interaction pose prediction. Additionally, we
proposed a novel modular framework that delivers superior and robust
performance for generalizable articulated object manipulation. Our extensive
experiments demonstrate that our dataset significantly improves the performance
of depth perception and actionable interaction pose prediction in both
simulation and real-world scenarios. More information and demos can be found
at: https://pku-epic.github.io/GAPartManip/.
|
2412.03473 | Ziwen Li | Ziwen Li, Jiaxin Huang, Runnan Chen, Yunlong Che, Yandong Guo,
Tongliang Liu, Fakhri Karray, Mingming Gong | UrbanGS: Semantic-Guided Gaussian Splatting for Urban Scene
Reconstruction | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Reconstructing urban scenes is challenging due to their complex geometries
and the presence of potentially dynamic objects. 3D Gaussian Splatting
(3DGS)-based methods have shown strong performance, but existing approaches
often incorporate manual 3D annotations to improve dynamic object modeling,
which is impractical due to high labeling costs. Some methods leverage 4D
Gaussian Splatting (4DGS) to represent the entire scene, but they treat static
and dynamic objects uniformly, leading to unnecessary updates for static
elements and ultimately degrading reconstruction quality. To address these
issues, we propose UrbanGS, which leverages 2D semantic maps and an existing
dynamic Gaussian approach to distinguish static objects from the scene,
enabling separate processing of definite static and potentially dynamic
elements. Specifically, for definite static regions, we enforce global
consistency to prevent unintended changes in dynamic Gaussian and introduce a
K-nearest neighbor (KNN)-based regularization to improve local coherence on
low-textured ground surfaces. Notably, for potentially dynamic objects, we
aggregate temporal information using learnable time embeddings, allowing each
Gaussian to model deformations over time. Extensive experiments on real-world
datasets demonstrate that our approach outperforms state-of-the-art methods in
reconstruction quality and efficiency, accurately preserving static content
while capturing dynamic elements.
| [
{
"version": "v1",
"created": "Wed, 4 Dec 2024 16:59:49 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 10:30:57 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Ziwen",
""
],
[
"Huang",
"Jiaxin",
""
],
[
"Chen",
"Runnan",
""
],
[
"Che",
"Yunlong",
""
],
[
"Guo",
"Yandong",
""
],
[
"Liu",
"Tongliang",
""
],
[
"Karray",
"Fakhri",
""
],
[
"Gong",
"Mingming",
""
]
] | TITLE: UrbanGS: Semantic-Guided Gaussian Splatting for Urban Scene
Reconstruction
ABSTRACT: Reconstructing urban scenes is challenging due to their complex geometries
and the presence of potentially dynamic objects. 3D Gaussian Splatting
(3DGS)-based methods have shown strong performance, but existing approaches
often incorporate manual 3D annotations to improve dynamic object modeling,
which is impractical due to high labeling costs. Some methods leverage 4D
Gaussian Splatting (4DGS) to represent the entire scene, but they treat static
and dynamic objects uniformly, leading to unnecessary updates for static
elements and ultimately degrading reconstruction quality. To address these
issues, we propose UrbanGS, which leverages 2D semantic maps and an existing
dynamic Gaussian approach to distinguish static objects from the scene,
enabling separate processing of definite static and potentially dynamic
elements. Specifically, for definite static regions, we enforce global
consistency to prevent unintended changes in dynamic Gaussian and introduce a
K-nearest neighbor (KNN)-based regularization to improve local coherence on
low-textured ground surfaces. Notably, for potentially dynamic objects, we
aggregate temporal information using learnable time embeddings, allowing each
Gaussian to model deformations over time. Extensive experiments on real-world
datasets demonstrate that our approach outperforms state-of-the-art methods in
reconstruction quality and efficiency, accurately preserving static content
while capturing dynamic elements.
|
2412.06699 | Huachen Gao | Baorui Ma, Huachen Gao, Haoge Deng, Zhengxiong Luo, Tiejun Huang, Lulu
Tang, Xinlong Wang | You See it, You Got it: Learning 3D Creation on Pose-Free Videos at
Scale | Accepted by CVPR 2025, Project Page: https://vision.baai.ac.cn/see3d | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent 3D generation models typically rely on limited-scale 3D `gold-labels'
or 2D diffusion priors for 3D content creation. However, their performance is
upper-bounded by constrained 3D priors due to the lack of scalable learning
paradigms. In this work, we present See3D, a visual-conditional multi-view
diffusion model trained on large-scale Internet videos for open-world 3D
creation. The model aims to Get 3D knowledge by solely Seeing the visual
contents from the vast and rapidly growing video data -- You See it, You Got
it. To achieve this, we first scale up the training data using a proposed data
curation pipeline that automatically filters out multi-view inconsistencies and
insufficient observations from source videos. This results in a high-quality,
richly diverse, large-scale dataset of multi-view images, termed WebVi3D,
containing 320M frames from 16M video clips. Nevertheless, learning generic 3D
priors from videos without explicit 3D geometry or camera pose annotations is
nontrivial, and annotating poses for web-scale videos is prohibitively
expensive. To eliminate the need for pose conditions, we introduce an
innovative visual-condition - a purely 2D-inductive visual signal generated by
adding time-dependent noise to the masked video data. Finally, we introduce a
novel visual-conditional 3D generation framework by integrating See3D into a
warping-based pipeline for high-fidelity 3D generation. Our numerical and
visual comparisons on single and sparse reconstruction benchmarks show that
See3D, trained on cost-effective and scalable video data, achieves notable
zero-shot and open-world generation capabilities, markedly outperforming models
trained on costly and constrained 3D datasets. Please refer to our project page
at: https://vision.baai.ac.cn/see3d
| [
{
"version": "v1",
"created": "Mon, 9 Dec 2024 17:44:56 GMT"
},
{
"version": "v2",
"created": "Sat, 14 Dec 2024 15:42:05 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 08:55:03 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ma",
"Baorui",
""
],
[
"Gao",
"Huachen",
""
],
[
"Deng",
"Haoge",
""
],
[
"Luo",
"Zhengxiong",
""
],
[
"Huang",
"Tiejun",
""
],
[
"Tang",
"Lulu",
""
],
[
"Wang",
"Xinlong",
""
]
] | TITLE: You See it, You Got it: Learning 3D Creation on Pose-Free Videos at
Scale
ABSTRACT: Recent 3D generation models typically rely on limited-scale 3D `gold-labels'
or 2D diffusion priors for 3D content creation. However, their performance is
upper-bounded by constrained 3D priors due to the lack of scalable learning
paradigms. In this work, we present See3D, a visual-conditional multi-view
diffusion model trained on large-scale Internet videos for open-world 3D
creation. The model aims to Get 3D knowledge by solely Seeing the visual
contents from the vast and rapidly growing video data -- You See it, You Got
it. To achieve this, we first scale up the training data using a proposed data
curation pipeline that automatically filters out multi-view inconsistencies and
insufficient observations from source videos. This results in a high-quality,
richly diverse, large-scale dataset of multi-view images, termed WebVi3D,
containing 320M frames from 16M video clips. Nevertheless, learning generic 3D
priors from videos without explicit 3D geometry or camera pose annotations is
nontrivial, and annotating poses for web-scale videos is prohibitively
expensive. To eliminate the need for pose conditions, we introduce an
innovative visual-condition - a purely 2D-inductive visual signal generated by
adding time-dependent noise to the masked video data. Finally, we introduce a
novel visual-conditional 3D generation framework by integrating See3D into a
warping-based pipeline for high-fidelity 3D generation. Our numerical and
visual comparisons on single and sparse reconstruction benchmarks show that
See3D, trained on cost-effective and scalable video data, achieves notable
zero-shot and open-world generation capabilities, markedly outperforming models
trained on costly and constrained 3D datasets. Please refer to our project page
at: https://vision.baai.ac.cn/see3d
|
2412.08376 | Siyan Dong | Siyan Dong, Shuzhe Wang, Shaohui Liu, Lulu Cai, Qingnan Fan, Juho
Kannala, Yanchao Yang | Reloc3r: Large-Scale Training of Relative Camera Pose Regression for
Generalizable, Fast, and Accurate Visual Localization | CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual localization aims to determine the camera pose of a query image
relative to a database of posed images. In recent years, deep neural networks
that directly regress camera poses have gained popularity due to their fast
inference capabilities. However, existing methods struggle to either generalize
well to new scenes or provide accurate camera pose estimates. To address these
issues, we present Reloc3r, a simple yet effective visual localization
framework. It consists of an elegantly designed relative pose regression
network, and a minimalist motion averaging module for absolute pose estimation.
Trained on approximately eight million posed image pairs, Reloc3r achieves
surprisingly good performance and generalization ability. We conduct extensive
experiments on six public datasets, consistently demonstrating the
effectiveness and efficiency of the proposed method. It provides high-quality
camera pose estimates in real time and generalizes to novel scenes. Code:
https://github.com/ffrivera0/reloc3r.
| [
{
"version": "v1",
"created": "Wed, 11 Dec 2024 13:36:18 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 10:18:18 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Dong",
"Siyan",
""
],
[
"Wang",
"Shuzhe",
""
],
[
"Liu",
"Shaohui",
""
],
[
"Cai",
"Lulu",
""
],
[
"Fan",
"Qingnan",
""
],
[
"Kannala",
"Juho",
""
],
[
"Yang",
"Yanchao",
""
]
] | TITLE: Reloc3r: Large-Scale Training of Relative Camera Pose Regression for
Generalizable, Fast, and Accurate Visual Localization
ABSTRACT: Visual localization aims to determine the camera pose of a query image
relative to a database of posed images. In recent years, deep neural networks
that directly regress camera poses have gained popularity due to their fast
inference capabilities. However, existing methods struggle to either generalize
well to new scenes or provide accurate camera pose estimates. To address these
issues, we present Reloc3r, a simple yet effective visual localization
framework. It consists of an elegantly designed relative pose regression
network, and a minimalist motion averaging module for absolute pose estimation.
Trained on approximately eight million posed image pairs, Reloc3r achieves
surprisingly good performance and generalization ability. We conduct extensive
experiments on six public datasets, consistently demonstrating the
effectiveness and efficiency of the proposed method. It provides high-quality
camera pose estimates in real time and generalizes to novel scenes. Code:
https://github.com/ffrivera0/reloc3r.
|
2412.09010 | Yusuke Sakemi Ph.D. | Yusuke Sakemi, Yuji Okamoto, Takashi Morie, Sou Nobukawa, Takeo
Hosomi, Kazuyuki Aihara | Harnessing Nonidealities in Analog In-Memory Computing Circuits: A
Physical Modeling Approach for Neuromorphic Systems | Title changed | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large-scale deep learning models are increasingly constrained by their
immense energy consumption, limiting their scalability and applicability for
edge intelligence. In-memory computing (IMC) offers a promising solution by
addressing the von Neumann bottleneck inherent in traditional deep learning
accelerators, significantly reducing energy consumption. However, the analog
nature of IMC introduces hardware nonidealities that degrade model performance
and reliability. This paper presents a novel approach to directly train
physical models of IMC, formulated as ordinary-differential-equation
(ODE)-based physical neural networks (PNNs). To enable the training of
large-scale networks, we propose a technique called differentiable spike-time
discretization (DSTD), which reduces the computational cost of ODE-based PNNs
by up to 20 times in speed and 100 times in memory. We demonstrate that such
large-scale networks enhance the learning performance by exploiting hardware
nonidealities on the CIFAR-10 dataset. The proposed bottom-up methodology is
validated through the post-layout SPICE simulations on the IMC circuit with
nonideal characteristics using the sky130 process. The proposed PNN approach
reduces the discrepancy between the model behavior and circuit dynamics by at
least an order of magnitude. This work paves the way for leveraging nonideal
physical devices, such as non-volatile resistive memories, for energy-efficient
deep learning applications.
| [
{
"version": "v1",
"created": "Thu, 12 Dec 2024 07:22:23 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 03:08:11 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Sakemi",
"Yusuke",
""
],
[
"Okamoto",
"Yuji",
""
],
[
"Morie",
"Takashi",
""
],
[
"Nobukawa",
"Sou",
""
],
[
"Hosomi",
"Takeo",
""
],
[
"Aihara",
"Kazuyuki",
""
]
] | TITLE: Harnessing Nonidealities in Analog In-Memory Computing Circuits: A
Physical Modeling Approach for Neuromorphic Systems
ABSTRACT: Large-scale deep learning models are increasingly constrained by their
immense energy consumption, limiting their scalability and applicability for
edge intelligence. In-memory computing (IMC) offers a promising solution by
addressing the von Neumann bottleneck inherent in traditional deep learning
accelerators, significantly reducing energy consumption. However, the analog
nature of IMC introduces hardware nonidealities that degrade model performance
and reliability. This paper presents a novel approach to directly train
physical models of IMC, formulated as ordinary-differential-equation
(ODE)-based physical neural networks (PNNs). To enable the training of
large-scale networks, we propose a technique called differentiable spike-time
discretization (DSTD), which reduces the computational cost of ODE-based PNNs
by up to 20 times in speed and 100 times in memory. We demonstrate that such
large-scale networks enhance the learning performance by exploiting hardware
nonidealities on the CIFAR-10 dataset. The proposed bottom-up methodology is
validated through the post-layout SPICE simulations on the IMC circuit with
nonideal characteristics using the sky130 process. The proposed PNN approach
reduces the discrepancy between the model behavior and circuit dynamics by at
least an order of magnitude. This work paves the way for leveraging nonideal
physical devices, such as non-volatile resistive memories, for energy-efficient
deep learning applications.
|
2412.12406 | Meisam Kabiri | Meisam Kabiri, Holger Voos | Global SLAM Using 5G ToA Integration: Performance Analysis with Unknown
Base Stations and Loop Closure Alternatives | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel approach that integrates 5G Time of Arrival (ToA)
measurements into ORB-SLAM3 to enable global localization and enhance mapping
capabilities for indoor drone navigation. We extend ORB-SLAM3's optimization
pipeline to jointly process ToA data from 5G base stations alongside visual and
inertial measurements while estimating system biases. This integration
transforms the inherently local SLAM estimates into globally referenced
trajectories and effectively resolves scale ambiguity in monocular
configurations. Our method is evaluated using both Aerolab indoor datasets with
RGB-D cameras and the EuRoC MAV benchmark, complemented by simulated 5G ToA
measurements at 28 GHz and 78 GHz frequencies using MATLAB and QuaDRiGa.
Extensive experiments across multiple SLAM configurations demonstrate that ToA
integration enables consistent global positioning across all modes while
maintaining local accuracy. For monocular configurations, ToA integration
successfully resolves scale ambiguity and improves consistency. We further
investigate scenarios with unknown base station positions and demonstrate that
ToA measurements can effectively serve as an alternative to loop closure for
drift correction. We also analyze how different geometric arrangements of base
stations impact SLAM performance. Comparative analysis with state-of-the-art
methods, including UWB-VO, confirms our approach's robustness even with lower
measurement frequencies and sequential base station operation. The results
validate that 5G ToA integration provides substantial benefits for global SLAM
applications, particularly in challenging indoor environments where accurate
positioning is critical.
| [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 23:17:40 GMT"
},
{
"version": "v2",
"created": "Sat, 28 Dec 2024 13:57:49 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Jan 2025 16:55:40 GMT"
},
{
"version": "v4",
"created": "Mon, 17 Mar 2025 21:48:33 GMT"
},
{
"version": "v5",
"created": "Fri, 21 Mar 2025 13:48:46 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kabiri",
"Meisam",
""
],
[
"Voos",
"Holger",
""
]
] | TITLE: Global SLAM Using 5G ToA Integration: Performance Analysis with Unknown
Base Stations and Loop Closure Alternatives
ABSTRACT: This paper presents a novel approach that integrates 5G Time of Arrival (ToA)
measurements into ORB-SLAM3 to enable global localization and enhance mapping
capabilities for indoor drone navigation. We extend ORB-SLAM3's optimization
pipeline to jointly process ToA data from 5G base stations alongside visual and
inertial measurements while estimating system biases. This integration
transforms the inherently local SLAM estimates into globally referenced
trajectories and effectively resolves scale ambiguity in monocular
configurations. Our method is evaluated using both Aerolab indoor datasets with
RGB-D cameras and the EuRoC MAV benchmark, complemented by simulated 5G ToA
measurements at 28 GHz and 78 GHz frequencies using MATLAB and QuaDRiGa.
Extensive experiments across multiple SLAM configurations demonstrate that ToA
integration enables consistent global positioning across all modes while
maintaining local accuracy. For monocular configurations, ToA integration
successfully resolves scale ambiguity and improves consistency. We further
investigate scenarios with unknown base station positions and demonstrate that
ToA measurements can effectively serve as an alternative to loop closure for
drift correction. We also analyze how different geometric arrangements of base
stations impact SLAM performance. Comparative analysis with state-of-the-art
methods, including UWB-VO, confirms our approach's robustness even with lower
measurement frequencies and sequential base station operation. The results
validate that 5G ToA integration provides substantial benefits for global SLAM
applications, particularly in challenging indoor environments where accurate
positioning is critical.
|
2412.15289 | Xiaoning Dong | Xiaoning Dong, Wenbo Hu, Wei Xu, Tianxing He | SATA: A Paradigm for LLM Jailbreak via Simple Assistive Task Linkage | null | null | null | null | cs.CR cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have made significant advancements across
various tasks, but their safety alignment remain a major concern. Exploring
jailbreak prompts can expose LLMs' vulnerabilities and guide efforts to secure
them. Existing methods primarily design sophisticated instructions for the LLM
to follow, or rely on multiple iterations, which could hinder the performance
and efficiency of jailbreaks. In this work, we propose a novel jailbreak
paradigm, Simple Assistive Task Linkage (SATA), which can effectively
circumvent LLM safeguards and elicit harmful responses. Specifically, SATA
first masks harmful keywords within a malicious query to generate a relatively
benign query containing one or multiple [MASK] special tokens. It then employs
a simple assistive task such as a masked language model task or an element
lookup by position task to encode the semantics of the masked keywords.
Finally, SATA links the assistive task with the masked query to jointly perform
the jailbreak. Extensive experiments show that SATA achieves state-of-the-art
performance and outperforms baselines by a large margin. Specifically, on
AdvBench dataset, with mask language model (MLM) assistive task, SATA achieves
an overall attack success rate (ASR) of 85% and harmful score (HS) of 4.57, and
with element lookup by position (ELP) assistive task, SATA attains an overall
ASR of 76% and HS of 4.43.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 05:57:37 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 13:00:44 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Dong",
"Xiaoning",
""
],
[
"Hu",
"Wenbo",
""
],
[
"Xu",
"Wei",
""
],
[
"He",
"Tianxing",
""
]
] | TITLE: SATA: A Paradigm for LLM Jailbreak via Simple Assistive Task Linkage
ABSTRACT: Large language models (LLMs) have made significant advancements across
various tasks, but their safety alignment remain a major concern. Exploring
jailbreak prompts can expose LLMs' vulnerabilities and guide efforts to secure
them. Existing methods primarily design sophisticated instructions for the LLM
to follow, or rely on multiple iterations, which could hinder the performance
and efficiency of jailbreaks. In this work, we propose a novel jailbreak
paradigm, Simple Assistive Task Linkage (SATA), which can effectively
circumvent LLM safeguards and elicit harmful responses. Specifically, SATA
first masks harmful keywords within a malicious query to generate a relatively
benign query containing one or multiple [MASK] special tokens. It then employs
a simple assistive task such as a masked language model task or an element
lookup by position task to encode the semantics of the masked keywords.
Finally, SATA links the assistive task with the masked query to jointly perform
the jailbreak. Extensive experiments show that SATA achieves state-of-the-art
performance and outperforms baselines by a large margin. Specifically, on
AdvBench dataset, with mask language model (MLM) assistive task, SATA achieves
an overall attack success rate (ASR) of 85% and harmful score (HS) of 4.57, and
with element lookup by position (ELP) assistive task, SATA attains an overall
ASR of 76% and HS of 4.43.
|
2412.18600 | Hongjie Li | Hongjie Li, Hong-Xing Yu, Jiaman Li, Jiajun Wu | ZeroHSI: Zero-Shot 4D Human-Scene Interaction by Video Generation | Project website: https://awfuact.github.io/zerohsi/ The first two
authors contribute equally | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human-scene interaction (HSI) generation is crucial for applications in
embodied AI, virtual reality, and robotics. Yet, existing methods cannot
synthesize interactions in unseen environments such as in-the-wild scenes or
reconstructed scenes, as they rely on paired 3D scenes and captured human
motion data for training, which are unavailable for unseen environments. We
present ZeroHSI, a novel approach that enables zero-shot 4D human-scene
interaction synthesis, eliminating the need for training on any MoCap data. Our
key insight is to distill human-scene interactions from state-of-the-art video
generation models, which have been trained on vast amounts of natural human
movements and interactions, and use differentiable rendering to reconstruct
human-scene interactions. ZeroHSI can synthesize realistic human motions in
both static scenes and environments with dynamic objects, without requiring any
ground-truth motion data. We evaluate ZeroHSI on a curated dataset of different
types of various indoor and outdoor scenes with different interaction prompts,
demonstrating its ability to generate diverse and contextually appropriate
human-scene interactions.
| [
{
"version": "v1",
"created": "Tue, 24 Dec 2024 18:55:38 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 16:17:28 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Hongjie",
""
],
[
"Yu",
"Hong-Xing",
""
],
[
"Li",
"Jiaman",
""
],
[
"Wu",
"Jiajun",
""
]
] | TITLE: ZeroHSI: Zero-Shot 4D Human-Scene Interaction by Video Generation
ABSTRACT: Human-scene interaction (HSI) generation is crucial for applications in
embodied AI, virtual reality, and robotics. Yet, existing methods cannot
synthesize interactions in unseen environments such as in-the-wild scenes or
reconstructed scenes, as they rely on paired 3D scenes and captured human
motion data for training, which are unavailable for unseen environments. We
present ZeroHSI, a novel approach that enables zero-shot 4D human-scene
interaction synthesis, eliminating the need for training on any MoCap data. Our
key insight is to distill human-scene interactions from state-of-the-art video
generation models, which have been trained on vast amounts of natural human
movements and interactions, and use differentiable rendering to reconstruct
human-scene interactions. ZeroHSI can synthesize realistic human motions in
both static scenes and environments with dynamic objects, without requiring any
ground-truth motion data. We evaluate ZeroHSI on a curated dataset of different
types of various indoor and outdoor scenes with different interaction prompts,
demonstrating its ability to generate diverse and contextually appropriate
human-scene interactions.
|
2412.20735 | Li Yang | Yang Li, Dong Du, Linfeng Song, Chen Li, Weikang Wang, Tao Yang,
Haitao Mi | HunyuanProver: A Scalable Data Synthesis Framework and Guided Tree
Search for Automated Theorem Proving | null | null | null | null | cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | We introduce HunyuanProver, an language model finetuned from the Hunyuan 7B
for interactive automatic theorem proving with LEAN4. To alleviate the data
sparsity issue, we design a scalable framework to iterative synthesize data
with low cost. Besides, guided tree search algorithms are designed to enable
effective ``system 2 thinking`` of the prover. HunyuanProver achieves
state-of-the-art (SOTA) performances on major benchmarks. Specifically, it
achieves a pass of 68.4% on the miniF2F-test compared to 65.9%, the current
SOTA results. It proves 4 IMO statements (imo_1960_p2, imo_1962_p2},
imo_1964_p2 and imo_1983_p6) in miniF2F-test. To benefit the community, we will
open-source a dataset of 30k synthesized instances, where each instance
contains the original question in natural language, the converted statement by
autoformalization, and the proof by HunyuanProver.
| [
{
"version": "v1",
"created": "Mon, 30 Dec 2024 06:18:33 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Dec 2024 10:48:14 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 02:00:37 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Yang",
""
],
[
"Du",
"Dong",
""
],
[
"Song",
"Linfeng",
""
],
[
"Li",
"Chen",
""
],
[
"Wang",
"Weikang",
""
],
[
"Yang",
"Tao",
""
],
[
"Mi",
"Haitao",
""
]
] | TITLE: HunyuanProver: A Scalable Data Synthesis Framework and Guided Tree
Search for Automated Theorem Proving
ABSTRACT: We introduce HunyuanProver, an language model finetuned from the Hunyuan 7B
for interactive automatic theorem proving with LEAN4. To alleviate the data
sparsity issue, we design a scalable framework to iterative synthesize data
with low cost. Besides, guided tree search algorithms are designed to enable
effective ``system 2 thinking`` of the prover. HunyuanProver achieves
state-of-the-art (SOTA) performances on major benchmarks. Specifically, it
achieves a pass of 68.4% on the miniF2F-test compared to 65.9%, the current
SOTA results. It proves 4 IMO statements (imo_1960_p2, imo_1962_p2},
imo_1964_p2 and imo_1983_p6) in miniF2F-test. To benefit the community, we will
open-source a dataset of 30k synthesized instances, where each instance
contains the original question in natural language, the converted statement by
autoformalization, and the proof by HunyuanProver.
|
2501.00174 | Marco Siino | Marco Siino, Ilenia Tinnirello, Marco La Cascia | The Text Classification Pipeline: Starting Shallow going Deeper | null | null | null | null | cs.CL cs.AI cs.IR | http://creativecommons.org/licenses/by/4.0/ | Text classification stands as a cornerstone within the realm of Natural
Language Processing (NLP), particularly when viewed through computer science
and engineering. The past decade has seen deep learning revolutionize text
classification, propelling advancements in text retrieval, categorization,
information extraction, and summarization. The scholarly literature includes
datasets, models, and evaluation criteria, with English being the predominant
language of focus, despite studies involving Arabic, Chinese, Hindi, and
others. The efficacy of text classification models relies heavily on their
ability to capture intricate textual relationships and non-linear correlations,
necessitating a comprehensive examination of the entire text classification
pipeline.
In the NLP domain, a plethora of text representation techniques and model
architectures have emerged, with Large Language Models (LLMs) and Generative
Pre-trained Transformers (GPTs) at the forefront. These models are adept at
transforming extensive textual data into meaningful vector representations
encapsulating semantic information. The multidisciplinary nature of text
classification, encompassing data mining, linguistics, and information
retrieval, highlights the importance of collaborative research to advance the
field. This work integrates traditional and contemporary text mining
methodologies, fostering a holistic understanding of text classification.
| [
{
"version": "v1",
"created": "Mon, 30 Dec 2024 23:01:19 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 19:18:07 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Siino",
"Marco",
""
],
[
"Tinnirello",
"Ilenia",
""
],
[
"La Cascia",
"Marco",
""
]
] | TITLE: The Text Classification Pipeline: Starting Shallow going Deeper
ABSTRACT: Text classification stands as a cornerstone within the realm of Natural
Language Processing (NLP), particularly when viewed through computer science
and engineering. The past decade has seen deep learning revolutionize text
classification, propelling advancements in text retrieval, categorization,
information extraction, and summarization. The scholarly literature includes
datasets, models, and evaluation criteria, with English being the predominant
language of focus, despite studies involving Arabic, Chinese, Hindi, and
others. The efficacy of text classification models relies heavily on their
ability to capture intricate textual relationships and non-linear correlations,
necessitating a comprehensive examination of the entire text classification
pipeline.
In the NLP domain, a plethora of text representation techniques and model
architectures have emerged, with Large Language Models (LLMs) and Generative
Pre-trained Transformers (GPTs) at the forefront. These models are adept at
transforming extensive textual data into meaningful vector representations
encapsulating semantic information. The multidisciplinary nature of text
classification, encompassing data mining, linguistics, and information
retrieval, highlights the importance of collaborative research to advance the
field. This work integrates traditional and contemporary text mining
methodologies, fostering a holistic understanding of text classification.
|
2501.08837 | Olga Zatsarynna | Olga Zatsarynna, Emad Bahrami, Yazan Abu Farha, Gianpiero Francesca,
Juergen Gall | MANTA: Diffusion Mamba for Efficient and Effective Stochastic Long-Term
Dense Anticipation | Accepted to CVPR2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Long-term dense action anticipation is very challenging since it requires
predicting actions and their durations several minutes into the future based on
provided video observations. To model the uncertainty of future outcomes,
stochastic models predict several potential future action sequences for the
same observation. Recent work has further proposed to incorporate uncertainty
modelling for observed frames by simultaneously predicting per-frame past and
future actions in a unified manner. While such joint modelling of actions is
beneficial, it requires long-range temporal capabilities to connect events
across distant past and future time points. However, the previous work
struggles to achieve such a long-range understanding due to its limited and/or
sparse receptive field. To alleviate this issue, we propose a novel MANTA
(MAmba for ANTicipation) network. Our model enables effective long-term
temporal modelling even for very long sequences while maintaining linear
complexity in sequence length. We demonstrate that our approach achieves
state-of-the-art results on three datasets - Breakfast, 50Salads, and
Assembly101 - while also significantly improving computational and memory
efficiency. Our code is available at https://github.com/olga-zats/DIFF_MANTA .
| [
{
"version": "v1",
"created": "Wed, 15 Jan 2025 14:46:44 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 17:04:07 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zatsarynna",
"Olga",
""
],
[
"Bahrami",
"Emad",
""
],
[
"Farha",
"Yazan Abu",
""
],
[
"Francesca",
"Gianpiero",
""
],
[
"Gall",
"Juergen",
""
]
] | TITLE: MANTA: Diffusion Mamba for Efficient and Effective Stochastic Long-Term
Dense Anticipation
ABSTRACT: Long-term dense action anticipation is very challenging since it requires
predicting actions and their durations several minutes into the future based on
provided video observations. To model the uncertainty of future outcomes,
stochastic models predict several potential future action sequences for the
same observation. Recent work has further proposed to incorporate uncertainty
modelling for observed frames by simultaneously predicting per-frame past and
future actions in a unified manner. While such joint modelling of actions is
beneficial, it requires long-range temporal capabilities to connect events
across distant past and future time points. However, the previous work
struggles to achieve such a long-range understanding due to its limited and/or
sparse receptive field. To alleviate this issue, we propose a novel MANTA
(MAmba for ANTicipation) network. Our model enables effective long-term
temporal modelling even for very long sequences while maintaining linear
complexity in sequence length. We demonstrate that our approach achieves
state-of-the-art results on three datasets - Breakfast, 50Salads, and
Assembly101 - while also significantly improving computational and memory
efficiency. Our code is available at https://github.com/olga-zats/DIFF_MANTA .
|
2501.08909 | Ruizhen Gu | Ruizhen Gu and Jos\'e Miguel Rojas and Donghwan Shin | Software Testing for Extended Reality Applications: A Systematic Mapping
Study | 51 pages, 10 figures | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Extended Reality (XR) is an emerging technology spanning diverse application
domains and offering immersive user experiences. However, its unique
characteristics, such as six degrees of freedom interactions, present
significant testing challenges distinct from traditional 2D GUI applications,
demanding novel testing techniques to build high-quality XR applications. This
paper presents the first systematic mapping study on software testing for XR
applications. We selected 34 studies focusing on techniques and empirical
approaches in XR software testing for detailed examination. The studies are
classified and reviewed to address the current research landscape, test facets,
and evaluation methodologies in the XR testing domain. Additionally, we provide
a repository summarising the mapping study, including datasets and tools
referenced in the selected studies, to support future research and practical
applications. Our study highlights open challenges in XR testing and proposes
actionable future research directions to address the gaps and advance the field
of XR software testing.
| [
{
"version": "v1",
"created": "Wed, 15 Jan 2025 16:19:12 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 18:11:30 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Gu",
"Ruizhen",
""
],
[
"Rojas",
"José Miguel",
""
],
[
"Shin",
"Donghwan",
""
]
] | TITLE: Software Testing for Extended Reality Applications: A Systematic Mapping
Study
ABSTRACT: Extended Reality (XR) is an emerging technology spanning diverse application
domains and offering immersive user experiences. However, its unique
characteristics, such as six degrees of freedom interactions, present
significant testing challenges distinct from traditional 2D GUI applications,
demanding novel testing techniques to build high-quality XR applications. This
paper presents the first systematic mapping study on software testing for XR
applications. We selected 34 studies focusing on techniques and empirical
approaches in XR software testing for detailed examination. The studies are
classified and reviewed to address the current research landscape, test facets,
and evaluation methodologies in the XR testing domain. Additionally, we provide
a repository summarising the mapping study, including datasets and tools
referenced in the selected studies, to support future research and practical
applications. Our study highlights open challenges in XR testing and proposes
actionable future research directions to address the gaps and advance the field
of XR software testing.
|
2501.11006 | Shashikant Ilager Mr | Shashikant Ilager, Lukas Florian Briem, Ivona Brandic | GREEN-CODE: Learning to Optimize Energy Efficiency in LLM-based Code
Generation | Under submission in ACM/IEEE conference, 11 pages | null | null | null | cs.DC cs.AI cs.PF cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) are becoming integral to daily life, showcasing
their vast potential across various Natural Language Processing (NLP) tasks.
Beyond NLP, LLMs are increasingly used in software development tasks, such as
code completion, modification, bug fixing, and code translation. Software
engineers widely use tools like GitHub Copilot and Amazon Q, streamlining
workflows and automating tasks with high accuracy. While the resource and
energy intensity of LLM training is often highlighted, inference can be even
more resource-intensive over time, as it's a continuous process with a high
number of invocations. Therefore, developing resource-efficient alternatives
for LLM inference is crucial for sustainability. This work proposes GREEN-CODE,
a framework for energy-aware code generation in LLMs. GREEN-CODE performs
dynamic early exit during LLM inference. We train a Reinforcement Learning (RL)
agent that learns to balance the trade-offs between accuracy, latency, and
energy consumption. Our approach is evaluated on two open-source LLMs, Llama
3.2 3B and OPT 2.7B, using the JavaCorpus and PY150 datasets. Results show that
our method reduces the energy consumption between 23-50 % on average for code
generation tasks without significantly affecting accuracy.
| [
{
"version": "v1",
"created": "Sun, 19 Jan 2025 10:44:03 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 15:07:55 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ilager",
"Shashikant",
""
],
[
"Briem",
"Lukas Florian",
""
],
[
"Brandic",
"Ivona",
""
]
] | TITLE: GREEN-CODE: Learning to Optimize Energy Efficiency in LLM-based Code
Generation
ABSTRACT: Large Language Models (LLMs) are becoming integral to daily life, showcasing
their vast potential across various Natural Language Processing (NLP) tasks.
Beyond NLP, LLMs are increasingly used in software development tasks, such as
code completion, modification, bug fixing, and code translation. Software
engineers widely use tools like GitHub Copilot and Amazon Q, streamlining
workflows and automating tasks with high accuracy. While the resource and
energy intensity of LLM training is often highlighted, inference can be even
more resource-intensive over time, as it's a continuous process with a high
number of invocations. Therefore, developing resource-efficient alternatives
for LLM inference is crucial for sustainability. This work proposes GREEN-CODE,
a framework for energy-aware code generation in LLMs. GREEN-CODE performs
dynamic early exit during LLM inference. We train a Reinforcement Learning (RL)
agent that learns to balance the trade-offs between accuracy, latency, and
energy consumption. Our approach is evaluated on two open-source LLMs, Llama
3.2 3B and OPT 2.7B, using the JavaCorpus and PY150 datasets. Results show that
our method reduces the energy consumption between 23-50 % on average for code
generation tasks without significantly affecting accuracy.
|
2502.00212 | Kefan Dong | Kefan Dong, Tengyu Ma | STP: Self-play LLM Theorem Provers with Iterative Conjecturing and
Proving | 25 pages, 5 figures | null | null | null | cs.LG cs.AI cs.LO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A fundamental challenge in formal theorem proving by LLMs is the lack of
high-quality training data. Although reinforcement learning or expert iteration
partially mitigates this issue by alternating between LLM generating proofs and
finetuning them on correctly generated ones, performance quickly plateaus due
to the scarcity of correct proofs (sparse rewards). To keep improving the
models with limited data, we draw inspiration from mathematicians, who
continuously develop new results, partly by proposing novel conjectures or
exercises (which are often variants of known results) and attempting to solve
them. We design the Self-play Theorem Prover (STP) that simultaneously takes on
two roles, conjecturer and prover, each providing training signals to the
other. The conjecturer is trained iteratively on previously generated
conjectures that are barely provable by the current prover, which incentivizes
it to generate increasingly challenging conjectures over time. The prover
attempts to prove the conjectures with standard expert iteration. We evaluate
STP with both Lean and Isabelle formal versifiers. With 51.3 billion tokens
generated during the training in Lean, STP proves 28.5% of the statements in
the LeanWorkbook dataset, doubling the previous best result of 13.2% achieved
through expert iteration. The final model achieves state-of-the-art performance
among whole-proof generation methods on miniF2F-test (65.0%, pass@3200),
Proofnet-test (23.9%, pass@3200) and PutnamBench (8/644, pass@3200). We release
our code, model, and dataset in this URL: https://github.com/kfdong/STP.
| [
{
"version": "v1",
"created": "Fri, 31 Jan 2025 23:01:48 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Feb 2025 07:20:28 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Feb 2025 03:52:52 GMT"
},
{
"version": "v4",
"created": "Fri, 21 Mar 2025 03:27:55 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Dong",
"Kefan",
""
],
[
"Ma",
"Tengyu",
""
]
] | TITLE: STP: Self-play LLM Theorem Provers with Iterative Conjecturing and
Proving
ABSTRACT: A fundamental challenge in formal theorem proving by LLMs is the lack of
high-quality training data. Although reinforcement learning or expert iteration
partially mitigates this issue by alternating between LLM generating proofs and
finetuning them on correctly generated ones, performance quickly plateaus due
to the scarcity of correct proofs (sparse rewards). To keep improving the
models with limited data, we draw inspiration from mathematicians, who
continuously develop new results, partly by proposing novel conjectures or
exercises (which are often variants of known results) and attempting to solve
them. We design the Self-play Theorem Prover (STP) that simultaneously takes on
two roles, conjecturer and prover, each providing training signals to the
other. The conjecturer is trained iteratively on previously generated
conjectures that are barely provable by the current prover, which incentivizes
it to generate increasingly challenging conjectures over time. The prover
attempts to prove the conjectures with standard expert iteration. We evaluate
STP with both Lean and Isabelle formal versifiers. With 51.3 billion tokens
generated during the training in Lean, STP proves 28.5% of the statements in
the LeanWorkbook dataset, doubling the previous best result of 13.2% achieved
through expert iteration. The final model achieves state-of-the-art performance
among whole-proof generation methods on miniF2F-test (65.0%, pass@3200),
Proofnet-test (23.9%, pass@3200) and PutnamBench (8/644, pass@3200). We release
our code, model, and dataset in this URL: https://github.com/kfdong/STP.
|
2502.02454 | Yuxin Qi | Quan Zhang, Yuxin Qi, Xi Tang, Jinwei Fang, Xi Lin, Ke Zhang, Chun
Yuan | IMDPrompter: Adapting SAM to Image Manipulation Detection by Cross-View
Automated Prompt Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using extensive training data from SA-1B, the Segment Anything Model (SAM)
has demonstrated exceptional generalization and zero-shot capabilities,
attracting widespread attention in areas such as medical image segmentation and
remote sensing image segmentation. However, its performance in the field of
image manipulation detection remains largely unexplored and unconfirmed. There
are two main challenges in applying SAM to image manipulation detection: a)
reliance on manual prompts, and b) the difficulty of single-view information in
supporting cross-dataset generalization. To address these challenges, we
develops a cross-view prompt learning paradigm called IMDPrompter based on SAM.
Benefiting from the design of automated prompts, IMDPrompter no longer relies
on manual guidance, enabling automated detection and localization.
Additionally, we propose components such as Cross-view Feature Perception,
Optimal Prompt Selection, and Cross-View Prompt Consistency, which facilitate
cross-view perceptual learning and guide SAM to generate accurate masks.
Extensive experimental results from five datasets (CASIA, Columbia, Coverage,
IMD2020, and NIST16) validate the effectiveness of our proposed method.
| [
{
"version": "v1",
"created": "Tue, 4 Feb 2025 16:20:41 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 09:53:55 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 08:02:45 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zhang",
"Quan",
""
],
[
"Qi",
"Yuxin",
""
],
[
"Tang",
"Xi",
""
],
[
"Fang",
"Jinwei",
""
],
[
"Lin",
"Xi",
""
],
[
"Zhang",
"Ke",
""
],
[
"Yuan",
"Chun",
""
]
] | TITLE: IMDPrompter: Adapting SAM to Image Manipulation Detection by Cross-View
Automated Prompt Learning
ABSTRACT: Using extensive training data from SA-1B, the Segment Anything Model (SAM)
has demonstrated exceptional generalization and zero-shot capabilities,
attracting widespread attention in areas such as medical image segmentation and
remote sensing image segmentation. However, its performance in the field of
image manipulation detection remains largely unexplored and unconfirmed. There
are two main challenges in applying SAM to image manipulation detection: a)
reliance on manual prompts, and b) the difficulty of single-view information in
supporting cross-dataset generalization. To address these challenges, we
develops a cross-view prompt learning paradigm called IMDPrompter based on SAM.
Benefiting from the design of automated prompts, IMDPrompter no longer relies
on manual guidance, enabling automated detection and localization.
Additionally, we propose components such as Cross-view Feature Perception,
Optimal Prompt Selection, and Cross-View Prompt Consistency, which facilitate
cross-view perceptual learning and guide SAM to generate accurate masks.
Extensive experimental results from five datasets (CASIA, Columbia, Coverage,
IMD2020, and NIST16) validate the effectiveness of our proposed method.
|
2502.03490 | Nora Belrose | David Johnston, Nora Belrose | Examining Two Hop Reasoning Through Information Content Scaling | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Prior work has found that transformers have an inconsistent ability to learn
to answer latent two-hop questions -- questions of the form "Who is Bob's
mother's boss?" We study why this is the case by examining how transformers'
capacity to learn datasets of two-hop questions and answers (two-hop QA) scales
with their size, motivated by prior work on transformer knowledge capacity for
simple factual memorization. We find that capacity scaling and generalization
both support the hypothesis that latent two-hop QA requires transformers to
learn each fact twice, while two-hop QA with chain of thought does not. We also
show that with appropriate dataset parameters, it is possible to "trap" very
small models in a regime where they memorize answers to two-hop questions
independently, even though they would perform better if they could learn to
answer them with function composition. Our findings show that measurement of
capacity scaling can complement existing interpretability methods, though there
are challenges in using it for this purpose.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 02:13:04 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 03:49:12 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Johnston",
"David",
""
],
[
"Belrose",
"Nora",
""
]
] | TITLE: Examining Two Hop Reasoning Through Information Content Scaling
ABSTRACT: Prior work has found that transformers have an inconsistent ability to learn
to answer latent two-hop questions -- questions of the form "Who is Bob's
mother's boss?" We study why this is the case by examining how transformers'
capacity to learn datasets of two-hop questions and answers (two-hop QA) scales
with their size, motivated by prior work on transformer knowledge capacity for
simple factual memorization. We find that capacity scaling and generalization
both support the hypothesis that latent two-hop QA requires transformers to
learn each fact twice, while two-hop QA with chain of thought does not. We also
show that with appropriate dataset parameters, it is possible to "trap" very
small models in a regime where they memorize answers to two-hop questions
independently, even though they would perform better if they could learn to
answer them with function composition. Our findings show that measurement of
capacity scaling can complement existing interpretability methods, though there
are challenges in using it for this purpose.
|
2502.07411 | Sheng Zhou | Sheng Zhou, Junbin Xiao, Qingyun Li, Yicong Li, Xun Yang, Dan Guo,
Meng Wang, Tat-Seng Chua, Angela Yao | EgoTextVQA: Towards Egocentric Scene-Text Aware Video Question Answering | Accepted by CVPR 2025 | null | null | null | cs.CV cs.MM | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We introduce EgoTextVQA, a novel and rigorously constructed benchmark for
egocentric QA assistance involving scene text. EgoTextVQA contains 1.5K
ego-view videos and 7K scene-text aware questions that reflect real user needs
in outdoor driving and indoor house-keeping activities. The questions are
designed to elicit identification and reasoning on scene text in an egocentric
and dynamic environment. With EgoTextVQA, we comprehensively evaluate 10
prominent multimodal large language models. Currently, all models struggle, and
the best results (Gemini 1.5 Pro) are around 33\% accuracy, highlighting the
severe deficiency of these techniques in egocentric QA assistance. Our further
investigations suggest that precise temporal grounding and multi-frame
reasoning, along with high resolution and auxiliary scene-text inputs, are key
for better performance. With thorough analyses and heuristic suggestions, we
hope EgoTextVQA can serve as a solid testbed for research in egocentric
scene-text QA assistance. Our dataset is released at:
https://github.com/zhousheng97/EgoTextVQA.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 09:45:06 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 14:21:30 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zhou",
"Sheng",
""
],
[
"Xiao",
"Junbin",
""
],
[
"Li",
"Qingyun",
""
],
[
"Li",
"Yicong",
""
],
[
"Yang",
"Xun",
""
],
[
"Guo",
"Dan",
""
],
[
"Wang",
"Meng",
""
],
[
"Chua",
"Tat-Seng",
""
],
[
"Yao",
"Angela",
""
]
] | TITLE: EgoTextVQA: Towards Egocentric Scene-Text Aware Video Question Answering
ABSTRACT: We introduce EgoTextVQA, a novel and rigorously constructed benchmark for
egocentric QA assistance involving scene text. EgoTextVQA contains 1.5K
ego-view videos and 7K scene-text aware questions that reflect real user needs
in outdoor driving and indoor house-keeping activities. The questions are
designed to elicit identification and reasoning on scene text in an egocentric
and dynamic environment. With EgoTextVQA, we comprehensively evaluate 10
prominent multimodal large language models. Currently, all models struggle, and
the best results (Gemini 1.5 Pro) are around 33\% accuracy, highlighting the
severe deficiency of these techniques in egocentric QA assistance. Our further
investigations suggest that precise temporal grounding and multi-frame
reasoning, along with high resolution and auxiliary scene-text inputs, are key
for better performance. With thorough analyses and heuristic suggestions, we
hope EgoTextVQA can serve as a solid testbed for research in egocentric
scene-text QA assistance. Our dataset is released at:
https://github.com/zhousheng97/EgoTextVQA.
|
2502.08317 | Jiarui Wu | Jiarui Wu, Zhuo Liu, Hangfeng He | Mitigating Hallucinations in Multimodal Spatial Relations through
Constraint-Aware Prompting | 19 pages | null | null | null | cs.CL cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatial relation hallucinations pose a persistent challenge in large
vision-language models (LVLMs), leading to generate incorrect predictions about
object positions and spatial configurations within an image. To address this
issue, we propose a constraint-aware prompting framework designed to reduce
spatial relation hallucinations. Specifically, we introduce two types of
constraints: (1) bidirectional constraint, which ensures consistency in
pairwise object relations, and (2) transitivity constraint, which enforces
relational dependence across multiple objects. By incorporating these
constraints, LVLMs can produce more spatially coherent and consistent outputs.
We evaluate our method on three widely-used spatial relation datasets,
demonstrating performance improvements over existing approaches. Additionally,
a systematic analysis of various bidirectional relation analysis choices and
transitivity reference selections highlights greater possibilities of our
methods in incorporating constraints to mitigate spatial relation
hallucinations.
| [
{
"version": "v1",
"created": "Wed, 12 Feb 2025 11:32:19 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 03:39:57 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Wu",
"Jiarui",
""
],
[
"Liu",
"Zhuo",
""
],
[
"He",
"Hangfeng",
""
]
] | TITLE: Mitigating Hallucinations in Multimodal Spatial Relations through
Constraint-Aware Prompting
ABSTRACT: Spatial relation hallucinations pose a persistent challenge in large
vision-language models (LVLMs), leading to generate incorrect predictions about
object positions and spatial configurations within an image. To address this
issue, we propose a constraint-aware prompting framework designed to reduce
spatial relation hallucinations. Specifically, we introduce two types of
constraints: (1) bidirectional constraint, which ensures consistency in
pairwise object relations, and (2) transitivity constraint, which enforces
relational dependence across multiple objects. By incorporating these
constraints, LVLMs can produce more spatially coherent and consistent outputs.
We evaluate our method on three widely-used spatial relation datasets,
demonstrating performance improvements over existing approaches. Additionally,
a systematic analysis of various bidirectional relation analysis choices and
transitivity reference selections highlights greater possibilities of our
methods in incorporating constraints to mitigate spatial relation
hallucinations.
|
2502.10127 | Gamal Elghazaly Dr. | Gamal Elghazaly and Raphael Frank | Leveraging V2X for Collaborative HD Maps Construction Using Scene Graph
Generation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | High-Definition (HD) maps play a crucial role in autonomous vehicle
navigation, complementing onboard perception sensors for improved accuracy and
safety. Traditional HD map generation relies on dedicated mapping vehicles,
which are costly and fail to capture real-time infrastructure changes. This
paper presents HDMapLaneNet, a novel framework leveraging V2X communication and
Scene Graph Generation to collaboratively construct a localized geometric layer
of HD maps. The approach extracts lane centerlines from front-facing camera
images, represents them as graphs, and transmits the data for global
aggregation to the cloud via V2X. Preliminary results on the nuScenes dataset
demonstrate superior association prediction performance compared to a
state-of-the-art method.
| [
{
"version": "v1",
"created": "Fri, 14 Feb 2025 12:56:10 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 16:34:23 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Elghazaly",
"Gamal",
""
],
[
"Frank",
"Raphael",
""
]
] | TITLE: Leveraging V2X for Collaborative HD Maps Construction Using Scene Graph
Generation
ABSTRACT: High-Definition (HD) maps play a crucial role in autonomous vehicle
navigation, complementing onboard perception sensors for improved accuracy and
safety. Traditional HD map generation relies on dedicated mapping vehicles,
which are costly and fail to capture real-time infrastructure changes. This
paper presents HDMapLaneNet, a novel framework leveraging V2X communication and
Scene Graph Generation to collaboratively construct a localized geometric layer
of HD maps. The approach extracts lane centerlines from front-facing camera
images, represents them as graphs, and transmits the data for global
aggregation to the cloud via V2X. Preliminary results on the nuScenes dataset
demonstrate superior association prediction performance compared to a
state-of-the-art method.
|
2502.12537 | Sina Montazeri | Sina Montazeri, Haseebullah Jumakhan, Amir Mirzaeinia | Finding Optimal Trading History in Reinforcement Learning for Stock
Market Trading | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper investigates the optimization of temporal windows in Financial
Deep Reinforcement Learning (DRL) models using 2D Convolutional Neural Networks
(CNNs). We introduce a novel approach to treating the temporal field as a
hyperparameter and examine its impact on model performance across various
datasets and feature arrangements. We introduce a new hyperparameter for the
CNN policy, proposing that this temporal field can and should be treated as a
hyperparameter for these models. We examine the significance of this temporal
field by iteratively expanding the window of observations presented to the CNN
policy during the deep reinforcement learning process. Our iterative process
involves progressively increasing the observation period from two weeks to
twelve weeks, allowing us to examine the effects of different temporal windows
on the model's performance. This window expansion is implemented in two
settings. In one setting, we rearrange the features in the dataset to group
them by company, allowing the model to have a full view of company data in its
observation window and CNN kernel. In the second setting, we do not group the
features by company, and features are arranged by category. Our study reveals
that shorter temporal windows are most effective when no feature rearrangement
to group per company is in effect. However, the model will utilize longer
temporal windows and yield better performance once we introduce the feature
rearrangement. To examine the consistency of our findings, we repeated our
experiment on two datasets containing the same thirty companies from the Dow
Jones Index but with different features in each dataset and consistently
observed the above-mentioned patterns. The result is a trading model
significantly outperforming global financial services firms such as the Global
X Guru by the established Mirae Asset.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2025 04:50:00 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Feb 2025 10:24:59 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Montazeri",
"Sina",
""
],
[
"Jumakhan",
"Haseebullah",
""
],
[
"Mirzaeinia",
"Amir",
""
]
] | TITLE: Finding Optimal Trading History in Reinforcement Learning for Stock
Market Trading
ABSTRACT: This paper investigates the optimization of temporal windows in Financial
Deep Reinforcement Learning (DRL) models using 2D Convolutional Neural Networks
(CNNs). We introduce a novel approach to treating the temporal field as a
hyperparameter and examine its impact on model performance across various
datasets and feature arrangements. We introduce a new hyperparameter for the
CNN policy, proposing that this temporal field can and should be treated as a
hyperparameter for these models. We examine the significance of this temporal
field by iteratively expanding the window of observations presented to the CNN
policy during the deep reinforcement learning process. Our iterative process
involves progressively increasing the observation period from two weeks to
twelve weeks, allowing us to examine the effects of different temporal windows
on the model's performance. This window expansion is implemented in two
settings. In one setting, we rearrange the features in the dataset to group
them by company, allowing the model to have a full view of company data in its
observation window and CNN kernel. In the second setting, we do not group the
features by company, and features are arranged by category. Our study reveals
that shorter temporal windows are most effective when no feature rearrangement
to group per company is in effect. However, the model will utilize longer
temporal windows and yield better performance once we introduce the feature
rearrangement. To examine the consistency of our findings, we repeated our
experiment on two datasets containing the same thirty companies from the Dow
Jones Index but with different features in each dataset and consistently
observed the above-mentioned patterns. The result is a trading model
significantly outperforming global financial services firms such as the Global
X Guru by the established Mirae Asset.
|
2502.15749 | Joonghyuk Hahn | Joonghyuk Hahn, Hyeseon Ahn, Jungin Kim, Soohan Lim, Yo-Sub Han | TCProF: Time-Complexity Prediction SSL Framework | 26 pages, 13 figures, This paper has been accepted to NAACL 2025 | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Time complexity is a theoretic measure to determine the amount of time the
algorithm needs for its execution. In reality, developers write algorithms into
code snippets within limited resources, making the calculation of a code's time
complexity a fundamental task. However, determining the precise time complexity
of a code is theoretically undecidable. In response, recent advancements have
leaned toward deploying datasets for code time complexity prediction and
initiating preliminary experiments for this challenge. We investigate the
challenge in low-resource scenarios where only a few labeled instances are
given for training. Remarkably, we are the first to introduce TCProF: a
Time-Complexity Prediction SSL Framework as an effective solution for code time
complexity prediction in low-resource settings. TCProF significantly boosts
performance by integrating our augmentation, symbolic modules, and a
co-training mechanism, achieving a more than 60% improvement over self-training
approaches. We further provide an extensive comparative analysis between
TCProF, ChatGPT, and Gemini-Pro, offering a detailed evaluation of our
approach. Our code is at https://github.com/peer0/few-shot-tc.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 12:39:33 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 01:48:59 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Hahn",
"Joonghyuk",
""
],
[
"Ahn",
"Hyeseon",
""
],
[
"Kim",
"Jungin",
""
],
[
"Lim",
"Soohan",
""
],
[
"Han",
"Yo-Sub",
""
]
] | TITLE: TCProF: Time-Complexity Prediction SSL Framework
ABSTRACT: Time complexity is a theoretic measure to determine the amount of time the
algorithm needs for its execution. In reality, developers write algorithms into
code snippets within limited resources, making the calculation of a code's time
complexity a fundamental task. However, determining the precise time complexity
of a code is theoretically undecidable. In response, recent advancements have
leaned toward deploying datasets for code time complexity prediction and
initiating preliminary experiments for this challenge. We investigate the
challenge in low-resource scenarios where only a few labeled instances are
given for training. Remarkably, we are the first to introduce TCProF: a
Time-Complexity Prediction SSL Framework as an effective solution for code time
complexity prediction in low-resource settings. TCProF significantly boosts
performance by integrating our augmentation, symbolic modules, and a
co-training mechanism, achieving a more than 60% improvement over self-training
approaches. We further provide an extensive comparative analysis between
TCProF, ChatGPT, and Gemini-Pro, offering a detailed evaluation of our
approach. Our code is at https://github.com/peer0/few-shot-tc.
|
2502.18321 | Shuyi Chen | Shuyi Chen, Ferdinando Fioretto, Feng Qiu, Shixiang Zhu | Global-Decision-Focused Neural ODEs for Proactive Grid Resilience
Management | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extreme hazard events such as wildfires and hurricanes increasingly threaten
power systems, causing widespread outages and disrupting critical services.
Recently, predict-then-optimize approaches have gained traction in grid
operations, where system functionality forecasts are first generated and then
used as inputs for downstream decision-making. However, this two-stage method
often results in a misalignment between prediction and optimization objectives,
leading to suboptimal resource allocation. To address this, we propose
predict-all-then-optimize-globally (PATOG), a framework that integrates outage
prediction with globally optimized interventions. At its core, our
global-decision-focused (GDF) neural ODE model captures outage dynamics while
optimizing resilience strategies in a decision-aware manner. Unlike
conventional methods, our approach ensures spatially and temporally coherent
decision-making, improving both predictive accuracy and operational efficiency.
Experiments on synthetic and real-world datasets demonstrate significant
improvements in outage prediction consistency and grid resilience.
| [
{
"version": "v1",
"created": "Tue, 25 Feb 2025 16:15:35 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 15:16:16 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Chen",
"Shuyi",
""
],
[
"Fioretto",
"Ferdinando",
""
],
[
"Qiu",
"Feng",
""
],
[
"Zhu",
"Shixiang",
""
]
] | TITLE: Global-Decision-Focused Neural ODEs for Proactive Grid Resilience
Management
ABSTRACT: Extreme hazard events such as wildfires and hurricanes increasingly threaten
power systems, causing widespread outages and disrupting critical services.
Recently, predict-then-optimize approaches have gained traction in grid
operations, where system functionality forecasts are first generated and then
used as inputs for downstream decision-making. However, this two-stage method
often results in a misalignment between prediction and optimization objectives,
leading to suboptimal resource allocation. To address this, we propose
predict-all-then-optimize-globally (PATOG), a framework that integrates outage
prediction with globally optimized interventions. At its core, our
global-decision-focused (GDF) neural ODE model captures outage dynamics while
optimizing resilience strategies in a decision-aware manner. Unlike
conventional methods, our approach ensures spatially and temporally coherent
decision-making, improving both predictive accuracy and operational efficiency.
Experiments on synthetic and real-world datasets demonstrate significant
improvements in outage prediction consistency and grid resilience.
|
2503.01924 | Yuhang Wang | Wang YuHang, Junkang Guo, Aolei Liu, Kaihao Wang, Zaitong Wu, Zhenyu
Liu, Wenfei Yin, Jian Liu | TAET: Two-Stage Adversarial Equalization Training on Long-Tailed
Distributions | Text: 8 pages of main content, 5 pages of appendices have been
accepted by CVPR2025 | Computer Vision and Pattern Recognition 2025 | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Adversarial robustness is a critical challenge in deploying deep neural
networks for real-world applications. While adversarial training is a widely
recognized defense strategy, most existing studies focus on balanced datasets,
overlooking the prevalence of long-tailed distributions in real-world data,
which significantly complicates robustness. This paper provides a comprehensive
analysis of adversarial training under long-tailed distributions and identifies
limitations in the current state-of-the-art method, AT-BSL, in achieving robust
performance under such conditions. To address these challenges, we propose a
novel training framework, TAET, which integrates an initial stabilization phase
followed by a stratified equalization adversarial training phase. Additionally,
prior work on long-tailed robustness has largely ignored the crucial evaluation
metric of balanced accuracy. To bridge this gap, we introduce the concept of
balanced robustness, a comprehensive metric tailored for assessing robustness
under long-tailed distributions. Extensive experiments demonstrate that our
method surpasses existing advanced defenses, achieving significant improvements
in both memory and computational efficiency. This work represents a substantial
advancement in addressing robustness challenges in real-world applications. Our
code is available at:
https://github.com/BuhuiOK/TAET-Two-Stage-Adversarial-Equalization-Training-on-Long-Tailed-Distributions.
| [
{
"version": "v1",
"created": "Sun, 2 Mar 2025 12:07:00 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 08:49:42 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 09:56:29 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"YuHang",
"Wang",
""
],
[
"Guo",
"Junkang",
""
],
[
"Liu",
"Aolei",
""
],
[
"Wang",
"Kaihao",
""
],
[
"Wu",
"Zaitong",
""
],
[
"Liu",
"Zhenyu",
""
],
[
"Yin",
"Wenfei",
""
],
[
"Liu",
"Jian",
""
]
] | TITLE: TAET: Two-Stage Adversarial Equalization Training on Long-Tailed
Distributions
ABSTRACT: Adversarial robustness is a critical challenge in deploying deep neural
networks for real-world applications. While adversarial training is a widely
recognized defense strategy, most existing studies focus on balanced datasets,
overlooking the prevalence of long-tailed distributions in real-world data,
which significantly complicates robustness. This paper provides a comprehensive
analysis of adversarial training under long-tailed distributions and identifies
limitations in the current state-of-the-art method, AT-BSL, in achieving robust
performance under such conditions. To address these challenges, we propose a
novel training framework, TAET, which integrates an initial stabilization phase
followed by a stratified equalization adversarial training phase. Additionally,
prior work on long-tailed robustness has largely ignored the crucial evaluation
metric of balanced accuracy. To bridge this gap, we introduce the concept of
balanced robustness, a comprehensive metric tailored for assessing robustness
under long-tailed distributions. Extensive experiments demonstrate that our
method surpasses existing advanced defenses, achieving significant improvements
in both memory and computational efficiency. This work represents a substantial
advancement in addressing robustness challenges in real-world applications. Our
code is available at:
https://github.com/BuhuiOK/TAET-Two-Stage-Adversarial-Equalization-Training-on-Long-Tailed-Distributions.
|
2503.03234 | Dakarai Crowder | Dakarai Crowder, Kojo Vandyck, Xiping Sun, James McCann, Wenzhen Yuan | Social Gesture Recognition in spHRI: Leveraging Fabric-Based Tactile
Sensing on Humanoid Robots | Accepted to ICRA 25. 8 pages, 8 figures | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Humans are able to convey different messages using only touch. Equipping
robots with the ability to understand social touch adds another modality in
which humans and robots can communicate. In this paper, we present a social
gesture recognition system using a fabric-based, large-scale tactile sensor
placed onto the arms of a humanoid robot. We built a social gesture dataset
using multiple participants and extracted temporal features for classification.
By collecting tactile data on a humanoid robot, our system provides insights
into human-robot social touch, and displays that the use of fabric based
sensors could be a potential way of advancing the development of spHRI systems
for more natural and effective communication.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2025 07:24:00 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Mar 2025 22:36:51 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 16:50:42 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Crowder",
"Dakarai",
""
],
[
"Vandyck",
"Kojo",
""
],
[
"Sun",
"Xiping",
""
],
[
"McCann",
"James",
""
],
[
"Yuan",
"Wenzhen",
""
]
] | TITLE: Social Gesture Recognition in spHRI: Leveraging Fabric-Based Tactile
Sensing on Humanoid Robots
ABSTRACT: Humans are able to convey different messages using only touch. Equipping
robots with the ability to understand social touch adds another modality in
which humans and robots can communicate. In this paper, we present a social
gesture recognition system using a fabric-based, large-scale tactile sensor
placed onto the arms of a humanoid robot. We built a social gesture dataset
using multiple participants and extracted temporal features for classification.
By collecting tactile data on a humanoid robot, our system provides insights
into human-robot social touch, and displays that the use of fabric based
sensors could be a potential way of advancing the development of spHRI systems
for more natural and effective communication.
|
2503.03750 | Mantas Mazeika | Richard Ren, Arunim Agarwal, Mantas Mazeika, Cristina Menghini, Robert
Vacareanu, Brad Kenstler, Mick Yang, Isabelle Barrass, Alice Gatti, Xuwang
Yin, Eduardo Trevino, Matias Geralnik, Adam Khoja, Dean Lee, Summer Yue, Dan
Hendrycks | The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems | Website: https://www.mask-benchmark.ai | null | null | null | cs.LG cs.AI cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | As large language models (LLMs) become more capable and agentic, the
requirement for trust in their outputs grows significantly, yet at the same
time concerns have been mounting that models may learn to lie in pursuit of
their goals. To address these concerns, a body of work has emerged around the
notion of "honesty" in LLMs, along with interventions aimed at mitigating
deceptive behaviors. However, evaluations of honesty are currently highly
limited, with no benchmark combining large scale and applicability to all
models. Moreover, many benchmarks claiming to measure honesty in fact simply
measure accuracy--the correctness of a model's beliefs--in disguise. In this
work, we introduce a large-scale human-collected dataset for measuring honesty
directly, allowing us to disentangle accuracy from honesty for the first time.
Across a diverse set of LLMs, we find that while larger models obtain higher
accuracy on our benchmark, they do not become more honest. Surprisingly, while
most frontier LLMs obtain high scores on truthfulness benchmarks, we find a
substantial propensity in frontier LLMs to lie when pressured to do so,
resulting in low honesty scores on our benchmark. We find that simple methods,
such as representation engineering interventions, can improve honesty. These
results underscore the growing need for robust evaluations and effective
interventions to ensure LLMs remain trustworthy.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2025 18:59:23 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 23:06:17 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ren",
"Richard",
""
],
[
"Agarwal",
"Arunim",
""
],
[
"Mazeika",
"Mantas",
""
],
[
"Menghini",
"Cristina",
""
],
[
"Vacareanu",
"Robert",
""
],
[
"Kenstler",
"Brad",
""
],
[
"Yang",
"Mick",
""
],
[
"Barrass",
"Isabelle",
""
],
[
"Gatti",
"Alice",
""
],
[
"Yin",
"Xuwang",
""
],
[
"Trevino",
"Eduardo",
""
],
[
"Geralnik",
"Matias",
""
],
[
"Khoja",
"Adam",
""
],
[
"Lee",
"Dean",
""
],
[
"Yue",
"Summer",
""
],
[
"Hendrycks",
"Dan",
""
]
] | TITLE: The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems
ABSTRACT: As large language models (LLMs) become more capable and agentic, the
requirement for trust in their outputs grows significantly, yet at the same
time concerns have been mounting that models may learn to lie in pursuit of
their goals. To address these concerns, a body of work has emerged around the
notion of "honesty" in LLMs, along with interventions aimed at mitigating
deceptive behaviors. However, evaluations of honesty are currently highly
limited, with no benchmark combining large scale and applicability to all
models. Moreover, many benchmarks claiming to measure honesty in fact simply
measure accuracy--the correctness of a model's beliefs--in disguise. In this
work, we introduce a large-scale human-collected dataset for measuring honesty
directly, allowing us to disentangle accuracy from honesty for the first time.
Across a diverse set of LLMs, we find that while larger models obtain higher
accuracy on our benchmark, they do not become more honest. Surprisingly, while
most frontier LLMs obtain high scores on truthfulness benchmarks, we find a
substantial propensity in frontier LLMs to lie when pressured to do so,
resulting in low honesty scores on our benchmark. We find that simple methods,
such as representation engineering interventions, can improve honesty. These
results underscore the growing need for robust evaluations and effective
interventions to ensure LLMs remain trustworthy.
|
2503.06146 | Ziyue Huang | Ziyue Huang, Yongchao Feng, Shuai Yang, Ziqi Liu, Qingjie Liu, and
Yunhong Wang | OpenRSD: Towards Open-prompts for Object Detection in Remote Sensing
Images | 11 pages, 4 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Remote sensing object detection has made significant progress, but most
studies still focus on closed-set detection, limiting generalization across
diverse datasets. Open-vocabulary object detection (OVD) provides a solution by
leveraging multimodal associations between text prompts and visual features.
However, existing OVD methods for remote sensing (RS) images are constrained by
small-scale datasets and fail to address the unique challenges of remote
sensing interpretation, include oriented object detection and the need for both
high precision and real-time performance in diverse scenarios. To tackle these
challenges, we propose OpenRSD, a universal open-prompt RS object detection
framework. OpenRSD supports multimodal prompts and integrates multi-task
detection heads to balance accuracy and real-time requirements. Additionally,
we design a multi-stage training pipeline to enhance the generalization of
model. Evaluated on seven public datasets, OpenRSD demonstrates superior
performance in oriented and horizontal bounding box detection, with real-time
inference capabilities suitable for large-scale RS image analysis. Compared to
YOLO-World, OpenRSD exhibits an 8.7\% higher average precision and achieves an
inference speed of 20.8 FPS. Codes and models will be released.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 10:08:46 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 06:47:18 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Huang",
"Ziyue",
""
],
[
"Feng",
"Yongchao",
""
],
[
"Yang",
"Shuai",
""
],
[
"Liu",
"Ziqi",
""
],
[
"Liu",
"Qingjie",
""
],
[
"Wang",
"Yunhong",
""
]
] | TITLE: OpenRSD: Towards Open-prompts for Object Detection in Remote Sensing
Images
ABSTRACT: Remote sensing object detection has made significant progress, but most
studies still focus on closed-set detection, limiting generalization across
diverse datasets. Open-vocabulary object detection (OVD) provides a solution by
leveraging multimodal associations between text prompts and visual features.
However, existing OVD methods for remote sensing (RS) images are constrained by
small-scale datasets and fail to address the unique challenges of remote
sensing interpretation, include oriented object detection and the need for both
high precision and real-time performance in diverse scenarios. To tackle these
challenges, we propose OpenRSD, a universal open-prompt RS object detection
framework. OpenRSD supports multimodal prompts and integrates multi-task
detection heads to balance accuracy and real-time requirements. Additionally,
we design a multi-stage training pipeline to enhance the generalization of
model. Evaluated on seven public datasets, OpenRSD demonstrates superior
performance in oriented and horizontal bounding box detection, with real-time
inference capabilities suitable for large-scale RS image analysis. Compared to
YOLO-World, OpenRSD exhibits an 8.7\% higher average precision and achieves an
inference speed of 20.8 FPS. Codes and models will be released.
|
2503.06158 | Ziruo Hao | Ziruo Hao, Zhenhua Cui, Tao Yang, Bo Hu, Xiaofeng Wu, Hui Feng | Invariant Federated Learning for Edge Intelligence: Mitigating
Heterogeneity and Asynchrony via Exit Strategy and Invariant Penalty | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper provides an invariant federated learning system for
resource-constrained edge intelligence. This framework can avoid the impact of
heterogeneity and asynchrony by exit strategy and invariant penalty. We
decompose local information into two orthogonal components to measure the
contribution or impact of heterogeneous and asynchronous clients. We propose
that the exit of abnormal clients can guarantee the effect of the model on most
clients. Meanwhile, to ensure the models' performance on exited abnormal
clients and those who lack training resources, we propose Federated Learning
with Invariant Penalty for Generalization (FedIPG) based on the invariant
orthogonal decomposition of parameters. Theoretical proof shows that FedIPG
reduces the Out-Of-Distribution prediction loss without increasing the
communication burden. The performance of FedIPG combined with an exit strategy
is tested empirically in multiple scales using four datasets. It shows our
system can enhance In-Distribution performance and outperform the
state-of-the-art algorithm in Out-Of-Distribution generalization while
maintaining model convergence. Additionally, the results of the visual
experiment prove that FedIPG contains preliminary causality in terms of
ignoring confounding features.
| [
{
"version": "v1",
"created": "Sat, 8 Mar 2025 10:47:27 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 12:03:44 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Hao",
"Ziruo",
""
],
[
"Cui",
"Zhenhua",
""
],
[
"Yang",
"Tao",
""
],
[
"Hu",
"Bo",
""
],
[
"Wu",
"Xiaofeng",
""
],
[
"Feng",
"Hui",
""
]
] | TITLE: Invariant Federated Learning for Edge Intelligence: Mitigating
Heterogeneity and Asynchrony via Exit Strategy and Invariant Penalty
ABSTRACT: This paper provides an invariant federated learning system for
resource-constrained edge intelligence. This framework can avoid the impact of
heterogeneity and asynchrony by exit strategy and invariant penalty. We
decompose local information into two orthogonal components to measure the
contribution or impact of heterogeneous and asynchronous clients. We propose
that the exit of abnormal clients can guarantee the effect of the model on most
clients. Meanwhile, to ensure the models' performance on exited abnormal
clients and those who lack training resources, we propose Federated Learning
with Invariant Penalty for Generalization (FedIPG) based on the invariant
orthogonal decomposition of parameters. Theoretical proof shows that FedIPG
reduces the Out-Of-Distribution prediction loss without increasing the
communication burden. The performance of FedIPG combined with an exit strategy
is tested empirically in multiple scales using four datasets. It shows our
system can enhance In-Distribution performance and outperform the
state-of-the-art algorithm in Out-Of-Distribution generalization while
maintaining model convergence. Additionally, the results of the visual
experiment prove that FedIPG contains preliminary causality in terms of
ignoring confounding features.
|
2503.07390 | Boeun Kim | Boeun Kim, Hea In Jeong, JungHoon Sung, Yihua Cheng, Jeongmin Lee, Ju
Yong Chang, Sang-Il Choi, Younggeun Choi, Saim Shin, Jungho Kim, Hyung Jin
Chang | PersonaBooth: Personalized Text-to-Motion Generation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper introduces Motion Personalization, a new task that generates
personalized motions aligned with text descriptions using several basic motions
containing Persona. To support this novel task, we introduce a new large-scale
motion dataset called PerMo (PersonaMotion), which captures the unique personas
of multiple actors. We also propose a multi-modal finetuning method of a
pretrained motion diffusion model called PersonaBooth. PersonaBooth addresses
two main challenges: i) A significant distribution gap between the
persona-focused PerMo dataset and the pretraining datasets, which lack
persona-specific data, and ii) the difficulty of capturing a consistent persona
from the motions vary in content (action type). To tackle the dataset
distribution gap, we introduce a persona token to accept new persona features
and perform multi-modal adaptation for both text and visuals during finetuning.
To capture a consistent persona, we incorporate a contrastive learning
technique to enhance intra-cohesion among samples with the same persona.
Furthermore, we introduce a context-aware fusion mechanism to maximize the
integration of persona cues from multiple input motions. PersonaBooth
outperforms state-of-the-art motion style transfer methods, establishing a new
benchmark for motion personalization.
| [
{
"version": "v1",
"created": "Mon, 10 Mar 2025 14:38:00 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 09:23:01 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 04:22:55 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kim",
"Boeun",
""
],
[
"Jeong",
"Hea In",
""
],
[
"Sung",
"JungHoon",
""
],
[
"Cheng",
"Yihua",
""
],
[
"Lee",
"Jeongmin",
""
],
[
"Chang",
"Ju Yong",
""
],
[
"Choi",
"Sang-Il",
""
],
[
"Choi",
"Younggeun",
""
],
[
"Shin",
"Saim",
""
],
[
"Kim",
"Jungho",
""
],
[
"Chang",
"Hyung Jin",
""
]
] | TITLE: PersonaBooth: Personalized Text-to-Motion Generation
ABSTRACT: This paper introduces Motion Personalization, a new task that generates
personalized motions aligned with text descriptions using several basic motions
containing Persona. To support this novel task, we introduce a new large-scale
motion dataset called PerMo (PersonaMotion), which captures the unique personas
of multiple actors. We also propose a multi-modal finetuning method of a
pretrained motion diffusion model called PersonaBooth. PersonaBooth addresses
two main challenges: i) A significant distribution gap between the
persona-focused PerMo dataset and the pretraining datasets, which lack
persona-specific data, and ii) the difficulty of capturing a consistent persona
from the motions vary in content (action type). To tackle the dataset
distribution gap, we introduce a persona token to accept new persona features
and perform multi-modal adaptation for both text and visuals during finetuning.
To capture a consistent persona, we incorporate a contrastive learning
technique to enhance intra-cohesion among samples with the same persona.
Furthermore, we introduce a context-aware fusion mechanism to maximize the
integration of persona cues from multiple input motions. PersonaBooth
outperforms state-of-the-art motion style transfer methods, establishing a new
benchmark for motion personalization.
|
2503.08140 | Ethan Griffiths | Ethan Griffiths, Maryam Haghighat, Simon Denman, Clinton Fookes, Milad
Ramezani | HOTFormerLoc: Hierarchical Octree Transformer for Versatile Lidar Place
Recognition Across Ground and Aerial Views | 16 pages, 13 figures, 10 tables, Accepted to CVPR 2025 | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present HOTFormerLoc, a novel and versatile Hierarchical Octree-based
TransFormer, for large-scale 3D place recognition in both ground-to-ground and
ground-to-aerial scenarios across urban and forest environments. We propose an
octree-based multi-scale attention mechanism that captures spatial and semantic
features across granularities. To address the variable density of point
distributions from spinning lidar, we present cylindrical octree attention
windows to reflect the underlying distribution during attention. We introduce
relay tokens to enable efficient global-local interactions and multi-scale
representation learning at reduced computational cost. Our pyramid attentional
pooling then synthesises a robust global descriptor for end-to-end place
recognition in challenging environments. In addition, we introduce
CS-Wild-Places, a novel 3D cross-source dataset featuring point cloud data from
aerial and ground lidar scans captured in dense forests. Point clouds in
CS-Wild-Places contain representational gaps and distinctive attributes such as
varying point densities and noise patterns, making it a challenging benchmark
for cross-view localisation in the wild. HOTFormerLoc achieves a top-1 average
recall improvement of 5.5% - 11.5% on the CS-Wild-Places benchmark.
Furthermore, it consistently outperforms SOTA 3D place recognition methods,
with an average performance gain of 4.9% on well-established urban and forest
datasets. The code and CS-Wild-Places benchmark is available at
https://csiro-robotics.github.io/HOTFormerLoc.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 07:59:45 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 07:00:11 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Griffiths",
"Ethan",
""
],
[
"Haghighat",
"Maryam",
""
],
[
"Denman",
"Simon",
""
],
[
"Fookes",
"Clinton",
""
],
[
"Ramezani",
"Milad",
""
]
] | TITLE: HOTFormerLoc: Hierarchical Octree Transformer for Versatile Lidar Place
Recognition Across Ground and Aerial Views
ABSTRACT: We present HOTFormerLoc, a novel and versatile Hierarchical Octree-based
TransFormer, for large-scale 3D place recognition in both ground-to-ground and
ground-to-aerial scenarios across urban and forest environments. We propose an
octree-based multi-scale attention mechanism that captures spatial and semantic
features across granularities. To address the variable density of point
distributions from spinning lidar, we present cylindrical octree attention
windows to reflect the underlying distribution during attention. We introduce
relay tokens to enable efficient global-local interactions and multi-scale
representation learning at reduced computational cost. Our pyramid attentional
pooling then synthesises a robust global descriptor for end-to-end place
recognition in challenging environments. In addition, we introduce
CS-Wild-Places, a novel 3D cross-source dataset featuring point cloud data from
aerial and ground lidar scans captured in dense forests. Point clouds in
CS-Wild-Places contain representational gaps and distinctive attributes such as
varying point densities and noise patterns, making it a challenging benchmark
for cross-view localisation in the wild. HOTFormerLoc achieves a top-1 average
recall improvement of 5.5% - 11.5% on the CS-Wild-Places benchmark.
Furthermore, it consistently outperforms SOTA 3D place recognition methods,
with an average performance gain of 4.9% on well-established urban and forest
datasets. The code and CS-Wild-Places benchmark is available at
https://csiro-robotics.github.io/HOTFormerLoc.
|
2503.09050 | Alvin Kimbowa | Alvin Kimbowa and Arjun Parmar and Maziar Badii and David Liu and
Matthew Harkey and Ilker Hacihaliloglu | Mono2D: A Trainable Monogenic Layer for Robust Knee Cartilage
Segmentation on Out-of-Distribution 2D Ultrasound Data | 11 pages, removed unrelated LaTeX template figure from last page | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Automated knee cartilage segmentation using point-of-care ultrasound devices
and deep-learning networks has the potential to enhance the management of knee
osteoarthritis. However, segmentation algorithms often struggle with domain
shifts caused by variations in ultrasound devices and acquisition parameters,
limiting their generalizability. In this paper, we propose Mono2D, a monogenic
layer that extracts multi-scale, contrast- and intensity-invariant local phase
features using trainable bandpass quadrature filters. This layer mitigates
domain shifts, improving generalization to out-of-distribution domains. Mono2D
is integrated before the first layer of a segmentation network, and its
parameters jointly trained alongside the network's parameters. We evaluated
Mono2D on a multi-domain 2D ultrasound knee cartilage dataset for single-source
domain generalization (SSDG). Our results demonstrate that Mono2D outperforms
other SSDG methods in terms of Dice score and mean average surface distance. To
further assess its generalizability, we evaluate Mono2D on a multi-site
prostate MRI dataset, where it continues to outperform other SSDG methods,
highlighting its potential to improve domain generalization in medical imaging.
Nevertheless, further evaluation on diverse datasets is still necessary to
assess its clinical utility.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2025 04:27:45 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 15:07:07 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kimbowa",
"Alvin",
""
],
[
"Parmar",
"Arjun",
""
],
[
"Badii",
"Maziar",
""
],
[
"Liu",
"David",
""
],
[
"Harkey",
"Matthew",
""
],
[
"Hacihaliloglu",
"Ilker",
""
]
] | TITLE: Mono2D: A Trainable Monogenic Layer for Robust Knee Cartilage
Segmentation on Out-of-Distribution 2D Ultrasound Data
ABSTRACT: Automated knee cartilage segmentation using point-of-care ultrasound devices
and deep-learning networks has the potential to enhance the management of knee
osteoarthritis. However, segmentation algorithms often struggle with domain
shifts caused by variations in ultrasound devices and acquisition parameters,
limiting their generalizability. In this paper, we propose Mono2D, a monogenic
layer that extracts multi-scale, contrast- and intensity-invariant local phase
features using trainable bandpass quadrature filters. This layer mitigates
domain shifts, improving generalization to out-of-distribution domains. Mono2D
is integrated before the first layer of a segmentation network, and its
parameters jointly trained alongside the network's parameters. We evaluated
Mono2D on a multi-domain 2D ultrasound knee cartilage dataset for single-source
domain generalization (SSDG). Our results demonstrate that Mono2D outperforms
other SSDG methods in terms of Dice score and mean average surface distance. To
further assess its generalizability, we evaluate Mono2D on a multi-site
prostate MRI dataset, where it continues to outperform other SSDG methods,
highlighting its potential to improve domain generalization in medical imaging.
Nevertheless, further evaluation on diverse datasets is still necessary to
assess its clinical utility.
|
2503.09417 | Dima Taji | Dima Taji and Daniel Zeman | Towards Generating Automatic Anaphora Annotations | 7 pages, 0 figures, 2 tables | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Training models that can perform well on various NLP tasks require large
amounts of data, and this becomes more apparent with nuanced tasks such as
anaphora and conference resolution. To combat the prohibitive costs of creating
manual gold annotated data, this paper explores two methods to automatically
create datasets with coreferential annotations; direct conversion from existing
datasets, and parsing using multilingual models capable of handling new and
unseen languages. The paper details the current progress on those two fronts,
as well as the challenges the efforts currently face, and our approach to
overcoming these challenges.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2025 14:15:57 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 13:00:05 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Taji",
"Dima",
""
],
[
"Zeman",
"Daniel",
""
]
] | TITLE: Towards Generating Automatic Anaphora Annotations
ABSTRACT: Training models that can perform well on various NLP tasks require large
amounts of data, and this becomes more apparent with nuanced tasks such as
anaphora and conference resolution. To combat the prohibitive costs of creating
manual gold annotated data, this paper explores two methods to automatically
create datasets with coreferential annotations; direct conversion from existing
datasets, and parsing using multilingual models capable of handling new and
unseen languages. The paper details the current progress on those two fronts,
as well as the challenges the efforts currently face, and our approach to
overcoming these challenges.
|
2503.10148 | Jialin Zhu | Jialin Zhu, Jiangbei Yue, Feixiang He, He Wang | 3D Student Splatting and Scooping | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, 3D Gaussian Splatting (3DGS) provides a new framework for novel
view synthesis, and has spiked a new wave of research in neural rendering and
related applications. As 3DGS is becoming a foundational component of many
models, any improvement on 3DGS itself can bring huge benefits. To this end, we
aim to improve the fundamental paradigm and formulation of 3DGS. We argue that
as an unnormalized mixture model, it needs to be neither Gaussians nor
splatting. We subsequently propose a new mixture model consisting of flexible
Student's t distributions, with both positive (splatting) and negative
(scooping) densities. We name our model Student Splatting and Scooping, or SSS.
When providing better expressivity, SSS also poses new challenges in learning.
Therefore, we also propose a new principled sampling approach for optimization.
Through exhaustive evaluation and comparison, across multiple datasets,
settings, and metrics, we demonstrate that SSS outperforms existing methods in
terms of quality and parameter efficiency, e.g. achieving matching or better
quality with similar numbers of components, and obtaining comparable results
while reducing the component number by as much as 82%.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 08:20:54 GMT"
},
{
"version": "v2",
"created": "Sat, 15 Mar 2025 13:33:29 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Mar 2025 02:26:08 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zhu",
"Jialin",
""
],
[
"Yue",
"Jiangbei",
""
],
[
"He",
"Feixiang",
""
],
[
"Wang",
"He",
""
]
] | TITLE: 3D Student Splatting and Scooping
ABSTRACT: Recently, 3D Gaussian Splatting (3DGS) provides a new framework for novel
view synthesis, and has spiked a new wave of research in neural rendering and
related applications. As 3DGS is becoming a foundational component of many
models, any improvement on 3DGS itself can bring huge benefits. To this end, we
aim to improve the fundamental paradigm and formulation of 3DGS. We argue that
as an unnormalized mixture model, it needs to be neither Gaussians nor
splatting. We subsequently propose a new mixture model consisting of flexible
Student's t distributions, with both positive (splatting) and negative
(scooping) densities. We name our model Student Splatting and Scooping, or SSS.
When providing better expressivity, SSS also poses new challenges in learning.
Therefore, we also propose a new principled sampling approach for optimization.
Through exhaustive evaluation and comparison, across multiple datasets,
settings, and metrics, we demonstrate that SSS outperforms existing methods in
terms of quality and parameter efficiency, e.g. achieving matching or better
quality with similar numbers of components, and obtaining comparable results
while reducing the component number by as much as 82%.
|
2503.10705 | Weiran Huang | Haoyuan Gao, Zicong Zhang, Yuqi Wei, Linglan Zhao, Guilin Li, Yexin
Li, Linghe Kong, Weiran Huang | Enhanced Continual Learning of Vision-Language Models with Model Fusion | Accepted by ICLR 2025 workshop | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-Language Models (VLMs) represent a breakthrough in artificial
intelligence by integrating visual and textual modalities to achieve impressive
zero-shot capabilities. However, VLMs are susceptible to catastrophic
forgetting when sequentially fine-tuned on multiple downstream tasks. Existing
continual learning methods for VLMs often rely heavily on additional reference
datasets, compromise zero-shot performance, or are limited to
parameter-efficient fine-tuning scenarios. In this paper, we propose Continual
Decoupling-Unifying (ConDU), a novel approach, by introducing model fusion into
continual learning for VLMs. ConDU maintains a unified model along with task
triggers and prototype sets, employing an iterative process of decoupling
task-specific models for previous tasks and unifying them with the model for
the newly learned task. Additionally, we introduce an inference strategy for
zero-shot scenarios by aggregating predictions from multiple decoupled
task-specific models. Extensive experiments across various settings show that
ConDU achieves up to a 2\% improvement in average performance across all seen
tasks compared to state-of-the-art baselines, while also enhancing zero-shot
capabilities relative to the original VLM.
| [
{
"version": "v1",
"created": "Wed, 12 Mar 2025 15:48:13 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 09:15:37 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Gao",
"Haoyuan",
""
],
[
"Zhang",
"Zicong",
""
],
[
"Wei",
"Yuqi",
""
],
[
"Zhao",
"Linglan",
""
],
[
"Li",
"Guilin",
""
],
[
"Li",
"Yexin",
""
],
[
"Kong",
"Linghe",
""
],
[
"Huang",
"Weiran",
""
]
] | TITLE: Enhanced Continual Learning of Vision-Language Models with Model Fusion
ABSTRACT: Vision-Language Models (VLMs) represent a breakthrough in artificial
intelligence by integrating visual and textual modalities to achieve impressive
zero-shot capabilities. However, VLMs are susceptible to catastrophic
forgetting when sequentially fine-tuned on multiple downstream tasks. Existing
continual learning methods for VLMs often rely heavily on additional reference
datasets, compromise zero-shot performance, or are limited to
parameter-efficient fine-tuning scenarios. In this paper, we propose Continual
Decoupling-Unifying (ConDU), a novel approach, by introducing model fusion into
continual learning for VLMs. ConDU maintains a unified model along with task
triggers and prototype sets, employing an iterative process of decoupling
task-specific models for previous tasks and unifying them with the model for
the newly learned task. Additionally, we introduce an inference strategy for
zero-shot scenarios by aggregating predictions from multiple decoupled
task-specific models. Extensive experiments across various settings show that
ConDU achieves up to a 2\% improvement in average performance across all seen
tasks compared to state-of-the-art baselines, while also enhancing zero-shot
capabilities relative to the original VLM.
|
2503.10838 | Russell Scheinberg | So Young Lee, Russell Scheinberg, Amber Shore, Ameeta Agrawal | Who Relies More on World Knowledge and Bias for Syntactic Ambiguity
Resolution: Humans or LLMs? | Accepted at NAACL 2025 main | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This study explores how recent large language models (LLMs) navigate relative
clause attachment {ambiguity} and use world knowledge biases for disambiguation
in six typologically diverse languages: English, Chinese, Japanese, Korean,
Russian, and Spanish. We describe the process of creating a novel dataset --
MultiWho -- for fine-grained evaluation of relative clause attachment
preferences in ambiguous and unambiguous contexts. Our experiments with three
LLMs indicate that, contrary to humans, LLMs consistently exhibit a preference
for local attachment, displaying limited responsiveness to syntactic variations
or language-specific attachment patterns. Although LLMs performed well in
unambiguous cases, they rigidly prioritized world knowledge biases, lacking the
flexibility of human language processing. These findings highlight the need for
more diverse, pragmatically nuanced multilingual training to improve LLMs'
handling of complex structures and human-like comprehension.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 19:44:15 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 19:35:30 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Lee",
"So Young",
""
],
[
"Scheinberg",
"Russell",
""
],
[
"Shore",
"Amber",
""
],
[
"Agrawal",
"Ameeta",
""
]
] | TITLE: Who Relies More on World Knowledge and Bias for Syntactic Ambiguity
Resolution: Humans or LLMs?
ABSTRACT: This study explores how recent large language models (LLMs) navigate relative
clause attachment {ambiguity} and use world knowledge biases for disambiguation
in six typologically diverse languages: English, Chinese, Japanese, Korean,
Russian, and Spanish. We describe the process of creating a novel dataset --
MultiWho -- for fine-grained evaluation of relative clause attachment
preferences in ambiguous and unambiguous contexts. Our experiments with three
LLMs indicate that, contrary to humans, LLMs consistently exhibit a preference
for local attachment, displaying limited responsiveness to syntactic variations
or language-specific attachment patterns. Although LLMs performed well in
unambiguous cases, they rigidly prioritized world knowledge biases, lacking the
flexibility of human language processing. These findings highlight the need for
more diverse, pragmatically nuanced multilingual training to improve LLMs'
handling of complex structures and human-like comprehension.
|
2503.11919 | Jeonghwan Park | Jeonghwan Park, Kang Li, Huiyu Zhou | k-fold Subsampling based Sequential Backward Feature Elimination | 8 pages | International Conference on Pattern Recognition Applications and
Methods, 2016 | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present a new wrapper feature selection algorithm for human detection.
This algorithm is a hybrid feature selection approach combining the benefits of
filter and wrapper methods. It allows the selection of an optimal feature
vector that well represents the shapes of the subjects in the images. In
detail, the proposed feature selection algorithm adopts the k-fold subsampling
and sequential backward elimination approach, while the standard linear support
vector machine (SVM) is used as the classifier for human detection. We apply
the proposed algorithm to the publicly accessible INRIA and ETH pedestrian full
image datasets with the PASCAL VOC evaluation criteria. Compared to other state
of the arts algorithms, our feature selection based approach can improve the
detection speed of the SVM classifier by over 50% with up to 2% better
detection accuracy. Our algorithm also outperforms the equivalent systems
introduced in the deformable part model approach with around 9% improvement in
the detection accuracy.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 23:10:08 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Park",
"Jeonghwan",
""
],
[
"Li",
"Kang",
""
],
[
"Zhou",
"Huiyu",
""
]
] | TITLE: k-fold Subsampling based Sequential Backward Feature Elimination
ABSTRACT: We present a new wrapper feature selection algorithm for human detection.
This algorithm is a hybrid feature selection approach combining the benefits of
filter and wrapper methods. It allows the selection of an optimal feature
vector that well represents the shapes of the subjects in the images. In
detail, the proposed feature selection algorithm adopts the k-fold subsampling
and sequential backward elimination approach, while the standard linear support
vector machine (SVM) is used as the classifier for human detection. We apply
the proposed algorithm to the publicly accessible INRIA and ETH pedestrian full
image datasets with the PASCAL VOC evaluation criteria. Compared to other state
of the arts algorithms, our feature selection based approach can improve the
detection speed of the SVM classifier by over 50% with up to 2% better
detection accuracy. Our algorithm also outperforms the equivalent systems
introduced in the deformable part model approach with around 9% improvement in
the detection accuracy.
|
2503.11921 | Atoosa Malemir Chegini | Atoosa Malemir Chegini, Keivan Rezaei, Hamid Eghbalzadeh, Soheil Feizi | RePanda: Pandas-powered Tabular Verification and Reasoning | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Fact-checking tabular data is essential for ensuring the accuracy of
structured information. However, existing methods often rely on black-box
models with opaque reasoning. We introduce RePanda, a structured fact
verification approach that translates claims into executable pandas queries,
enabling interpretable and verifiable reasoning.
To train RePanda, we construct PanTabFact, a structured dataset derived from
the TabFact train set, where claims are paired with executable queries
generated using DeepSeek-Chat and refined through automated error correction.
Fine-tuning DeepSeek-coder-7B-instruct-v1.5 on PanTabFact, RePanda achieves
84.09% accuracy on the TabFact test set.
To evaluate Out-of-Distribution (OOD) generalization, we interpret
question-answer pairs from WikiTableQuestions as factual claims and refer to
this dataset as WikiFact. Without additional fine-tuning, RePanda achieves
84.72% accuracy on WikiFact, significantly outperforming all other baselines
and demonstrating strong OOD robustness. Notably, these results closely match
the zero-shot performance of DeepSeek-Chat (671B), indicating that our
fine-tuning approach effectively distills structured reasoning from a much
larger model into a compact, locally executable 7B model.
Beyond fact verification, RePanda extends to tabular question answering by
generating executable queries that retrieve precise answers. To support this,
we introduce PanWiki, a dataset mapping WikiTableQuestions to pandas queries.
Fine-tuning on PanWiki, RePanda achieves 75.1% accuracy in direct answer
retrieval. These results highlight the effectiveness of structured
execution-based reasoning for tabular verification and question answering.
We have publicly released the dataset on Hugging Face at
datasets/AtoosaChegini/PanTabFact.
| [
{
"version": "v1",
"created": "Fri, 14 Mar 2025 23:12:36 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 19:10:27 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Chegini",
"Atoosa Malemir",
""
],
[
"Rezaei",
"Keivan",
""
],
[
"Eghbalzadeh",
"Hamid",
""
],
[
"Feizi",
"Soheil",
""
]
] | TITLE: RePanda: Pandas-powered Tabular Verification and Reasoning
ABSTRACT: Fact-checking tabular data is essential for ensuring the accuracy of
structured information. However, existing methods often rely on black-box
models with opaque reasoning. We introduce RePanda, a structured fact
verification approach that translates claims into executable pandas queries,
enabling interpretable and verifiable reasoning.
To train RePanda, we construct PanTabFact, a structured dataset derived from
the TabFact train set, where claims are paired with executable queries
generated using DeepSeek-Chat and refined through automated error correction.
Fine-tuning DeepSeek-coder-7B-instruct-v1.5 on PanTabFact, RePanda achieves
84.09% accuracy on the TabFact test set.
To evaluate Out-of-Distribution (OOD) generalization, we interpret
question-answer pairs from WikiTableQuestions as factual claims and refer to
this dataset as WikiFact. Without additional fine-tuning, RePanda achieves
84.72% accuracy on WikiFact, significantly outperforming all other baselines
and demonstrating strong OOD robustness. Notably, these results closely match
the zero-shot performance of DeepSeek-Chat (671B), indicating that our
fine-tuning approach effectively distills structured reasoning from a much
larger model into a compact, locally executable 7B model.
Beyond fact verification, RePanda extends to tabular question answering by
generating executable queries that retrieve precise answers. To support this,
we introduce PanWiki, a dataset mapping WikiTableQuestions to pandas queries.
Fine-tuning on PanWiki, RePanda achieves 75.1% accuracy in direct answer
retrieval. These results highlight the effectiveness of structured
execution-based reasoning for tabular verification and question answering.
We have publicly released the dataset on Hugging Face at
datasets/AtoosaChegini/PanTabFact.
|
2503.12261 | Rajasekar Gnana Praveen | R. Gnana Praveen, Jahangir Alam, Eric Charton | United we stand, Divided we fall: Handling Weak Complementary
Relationships for Audio-Visual Emotion Recognition in Valence-Arousal Space | Achieved 2nd place in valence arousal challenge Submission to
CVPR2025 Workshop | null | null | null | cs.CV cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | Audio and visual modalities are two predominant contact-free channels in
videos, which are often expected to carry a complementary relationship with
each other. However, they may not always complement each other, resulting in
poor audio-visual feature representations. In this paper, we introduce Gated
Recursive Joint Cross Attention (GRJCA) using a gating mechanism that can
adaptively choose the most relevant features to effectively capture the
synergic relationships across audio and visual modalities. Specifically, we
improve the performance of Recursive Joint Cross-Attention (RJCA) by
introducing a gating mechanism to control the flow of information between the
input features and the attended features of multiple iterations depending on
the strength of their complementary relationship. For instance, if the
modalities exhibit strong complementary relationships, the gating mechanism
emphasizes cross-attended features, otherwise non-attended features. To further
improve the performance of the system, we also explored a hierarchical gating
approach by introducing a gating mechanism at every iteration, followed by
high-level gating across the gated outputs of each iteration. The proposed
approach improves the performance of RJCA model by adding more flexibility to
deal with weak complementary relationships across audio and visual modalities.
Extensive experiments are conducted on the challenging Affwild2 dataset to
demonstrate the robustness of the proposed approach. By effectively handling
the weak complementary relationships across the audio and visual modalities,
the proposed model achieves a Concordance Correlation Coefficient (CCC) of
0.561 (0.623) and 0.620 (0.660) for valence and arousal respectively on the
test set (validation set).
| [
{
"version": "v1",
"created": "Sat, 15 Mar 2025 21:03:20 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 16:51:33 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Praveen",
"R. Gnana",
""
],
[
"Alam",
"Jahangir",
""
],
[
"Charton",
"Eric",
""
]
] | TITLE: United we stand, Divided we fall: Handling Weak Complementary
Relationships for Audio-Visual Emotion Recognition in Valence-Arousal Space
ABSTRACT: Audio and visual modalities are two predominant contact-free channels in
videos, which are often expected to carry a complementary relationship with
each other. However, they may not always complement each other, resulting in
poor audio-visual feature representations. In this paper, we introduce Gated
Recursive Joint Cross Attention (GRJCA) using a gating mechanism that can
adaptively choose the most relevant features to effectively capture the
synergic relationships across audio and visual modalities. Specifically, we
improve the performance of Recursive Joint Cross-Attention (RJCA) by
introducing a gating mechanism to control the flow of information between the
input features and the attended features of multiple iterations depending on
the strength of their complementary relationship. For instance, if the
modalities exhibit strong complementary relationships, the gating mechanism
emphasizes cross-attended features, otherwise non-attended features. To further
improve the performance of the system, we also explored a hierarchical gating
approach by introducing a gating mechanism at every iteration, followed by
high-level gating across the gated outputs of each iteration. The proposed
approach improves the performance of RJCA model by adding more flexibility to
deal with weak complementary relationships across audio and visual modalities.
Extensive experiments are conducted on the challenging Affwild2 dataset to
demonstrate the robustness of the proposed approach. By effectively handling
the weak complementary relationships across the audio and visual modalities,
the proposed model achieves a Concordance Correlation Coefficient (CCC) of
0.561 (0.623) and 0.620 (0.660) for valence and arousal respectively on the
test set (validation set).
|
2503.12760 | Brian Cho | Brian Cho, Ana-Roxana Pop, Ariel Evnine, Nathan Kallus | SNPL: Simultaneous Policy Learning and Evaluation for Safe
Multi-Objective Policy Improvement | null | null | null | null | stat.ML cs.LG econ.EM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To design effective digital interventions, experimenters face the challenge
of learning decision policies that balance multiple objectives using offline
data. Often, they aim to develop policies that maximize goal outcomes, while
ensuring there are no undesirable changes in guardrail outcomes. To provide
credible recommendations, experimenters must not only identify policies that
satisfy the desired changes in goal and guardrail outcomes, but also offer
probabilistic guarantees about the changes these policies induce. In practice,
however, policy classes are often large, and digital experiments tend to
produce datasets with small effect sizes relative to noise. In this setting,
standard approaches such as data splitting or multiple testing often result in
unstable policy selection and/or insufficient statistical power. In this paper,
we provide safe noisy policy learning (SNPL), a novel approach that leverages
the concept of algorithmic stability to address these challenges. Our method
enables policy learning while simultaneously providing high-confidence
guarantees using the entire dataset, avoiding the need for data-splitting. We
present finite-sample and asymptotic versions of our algorithm that ensure the
recommended policy satisfies high-probability guarantees for avoiding guardrail
regressions and/or achieving goal outcome improvements. We test both variants
of our approach approach empirically on a real-world application of
personalizing SMS delivery. Our results on real-world data suggest that our
approach offers dramatic improvements in settings with large policy classes and
low signal-to-noise across both finite-sample and asymptotic safety guarantees,
offering up to 300\% improvements in detection rates and 150\% improvements in
policy gains at significantly smaller sample sizes.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 02:53:53 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 17:38:14 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Cho",
"Brian",
""
],
[
"Pop",
"Ana-Roxana",
""
],
[
"Evnine",
"Ariel",
""
],
[
"Kallus",
"Nathan",
""
]
] | TITLE: SNPL: Simultaneous Policy Learning and Evaluation for Safe
Multi-Objective Policy Improvement
ABSTRACT: To design effective digital interventions, experimenters face the challenge
of learning decision policies that balance multiple objectives using offline
data. Often, they aim to develop policies that maximize goal outcomes, while
ensuring there are no undesirable changes in guardrail outcomes. To provide
credible recommendations, experimenters must not only identify policies that
satisfy the desired changes in goal and guardrail outcomes, but also offer
probabilistic guarantees about the changes these policies induce. In practice,
however, policy classes are often large, and digital experiments tend to
produce datasets with small effect sizes relative to noise. In this setting,
standard approaches such as data splitting or multiple testing often result in
unstable policy selection and/or insufficient statistical power. In this paper,
we provide safe noisy policy learning (SNPL), a novel approach that leverages
the concept of algorithmic stability to address these challenges. Our method
enables policy learning while simultaneously providing high-confidence
guarantees using the entire dataset, avoiding the need for data-splitting. We
present finite-sample and asymptotic versions of our algorithm that ensure the
recommended policy satisfies high-probability guarantees for avoiding guardrail
regressions and/or achieving goal outcome improvements. We test both variants
of our approach approach empirically on a real-world application of
personalizing SMS delivery. Our results on real-world data suggest that our
approach offers dramatic improvements in settings with large policy classes and
low signal-to-noise across both finite-sample and asymptotic safety guarantees,
offering up to 300\% improvements in detection rates and 150\% improvements in
policy gains at significantly smaller sample sizes.
|
2503.13938 | Qingyao Xu | Qingyao Xu, Siheng Chen, Guang Chen, Yanfeng Wang, Ya Zhang | ChatBEV: A Visual Language Model that Understands BEV Maps | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Traffic scene understanding is essential for intelligent transportation
systems and autonomous driving, ensuring safe and efficient vehicle operation.
While recent advancements in VLMs have shown promise for holistic scene
understanding, the application of VLMs to traffic scenarios, particularly using
BEV maps, remains under explored. Existing methods often suffer from limited
task design and narrow data amount, hindering comprehensive scene
understanding. To address these challenges, we introduce ChatBEV-QA, a novel
BEV VQA benchmark contains over 137k questions, designed to encompass a wide
range of scene understanding tasks, including global scene understanding,
vehicle-lane interactions, and vehicle-vehicle interactions. This benchmark is
constructed using an novel data collection pipeline that generates scalable and
informative VQA data for BEV maps. We further fine-tune a specialized
vision-language model ChatBEV, enabling it to interpret diverse question
prompts and extract relevant context-aware information from BEV maps.
Additionally, we propose a language-driven traffic scene generation pipeline,
where ChatBEV facilitates map understanding and text-aligned navigation
guidance, significantly enhancing the generation of realistic and consistent
traffic scenarios. The dataset, code and the fine-tuned model will be released.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 06:12:38 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 02:17:52 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Xu",
"Qingyao",
""
],
[
"Chen",
"Siheng",
""
],
[
"Chen",
"Guang",
""
],
[
"Wang",
"Yanfeng",
""
],
[
"Zhang",
"Ya",
""
]
] | TITLE: ChatBEV: A Visual Language Model that Understands BEV Maps
ABSTRACT: Traffic scene understanding is essential for intelligent transportation
systems and autonomous driving, ensuring safe and efficient vehicle operation.
While recent advancements in VLMs have shown promise for holistic scene
understanding, the application of VLMs to traffic scenarios, particularly using
BEV maps, remains under explored. Existing methods often suffer from limited
task design and narrow data amount, hindering comprehensive scene
understanding. To address these challenges, we introduce ChatBEV-QA, a novel
BEV VQA benchmark contains over 137k questions, designed to encompass a wide
range of scene understanding tasks, including global scene understanding,
vehicle-lane interactions, and vehicle-vehicle interactions. This benchmark is
constructed using an novel data collection pipeline that generates scalable and
informative VQA data for BEV maps. We further fine-tune a specialized
vision-language model ChatBEV, enabling it to interpret diverse question
prompts and extract relevant context-aware information from BEV maps.
Additionally, we propose a language-driven traffic scene generation pipeline,
where ChatBEV facilitates map understanding and text-aligned navigation
guidance, significantly enhancing the generation of realistic and consistent
traffic scenarios. The dataset, code and the fine-tuned model will be released.
|
2503.14275 | Jiang Qin | Jiang Qin, Senmao Li, Alexandra Gomez-Villa, Shiqi Yang, Yaxing Wang,
Kai Wang, Joost van de Weijer | Free-Lunch Color-Texture Disentanglement for Stylized Image Generation | Code will be available at https://deepffff.github.io/sadis.github.io/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in Text-to-Image (T2I) diffusion models have transformed
image generation, enabling significant progress in stylized generation using
only a few style reference images. However, current diffusion-based methods
struggle with fine-grained style customization due to challenges in controlling
multiple style attributes, such as color and texture. This paper introduces the
first tuning-free approach to achieve free-lunch color-texture disentanglement
in stylized T2I generation, addressing the need for independently controlled
style elements for the Disentangled Stylized Image Generation (DisIG) problem.
Our approach leverages the Image-Prompt Additivity property in the CLIP image
embedding space to develop techniques for separating and extracting
Color-Texture Embeddings (CTE) from individual color and texture reference
images. To ensure that the color palette of the generated image aligns closely
with the color reference, we apply a whitening and coloring transformation to
enhance color consistency. Additionally, to prevent texture loss due to the
signal-leak bias inherent in diffusion training, we introduce a noise term that
preserves textural fidelity during the Regularized Whitening and Coloring
Transformation (RegWCT). Through these methods, our Style Attributes
Disentanglement approach (SADis) delivers a more precise and customizable
solution for stylized image generation. Experiments on images from the WikiArt
and StyleDrop datasets demonstrate that, both qualitatively and quantitatively,
SADis surpasses state-of-the-art stylization methods in the DisIG task.Code
will be released at https://deepffff.github.io/sadis.github.io/.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 14:10:43 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 08:42:51 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Qin",
"Jiang",
""
],
[
"Li",
"Senmao",
""
],
[
"Gomez-Villa",
"Alexandra",
""
],
[
"Yang",
"Shiqi",
""
],
[
"Wang",
"Yaxing",
""
],
[
"Wang",
"Kai",
""
],
[
"van de Weijer",
"Joost",
""
]
] | TITLE: Free-Lunch Color-Texture Disentanglement for Stylized Image Generation
ABSTRACT: Recent advances in Text-to-Image (T2I) diffusion models have transformed
image generation, enabling significant progress in stylized generation using
only a few style reference images. However, current diffusion-based methods
struggle with fine-grained style customization due to challenges in controlling
multiple style attributes, such as color and texture. This paper introduces the
first tuning-free approach to achieve free-lunch color-texture disentanglement
in stylized T2I generation, addressing the need for independently controlled
style elements for the Disentangled Stylized Image Generation (DisIG) problem.
Our approach leverages the Image-Prompt Additivity property in the CLIP image
embedding space to develop techniques for separating and extracting
Color-Texture Embeddings (CTE) from individual color and texture reference
images. To ensure that the color palette of the generated image aligns closely
with the color reference, we apply a whitening and coloring transformation to
enhance color consistency. Additionally, to prevent texture loss due to the
signal-leak bias inherent in diffusion training, we introduce a noise term that
preserves textural fidelity during the Regularized Whitening and Coloring
Transformation (RegWCT). Through these methods, our Style Attributes
Disentanglement approach (SADis) delivers a more precise and customizable
solution for stylized image generation. Experiments on images from the WikiArt
and StyleDrop datasets demonstrate that, both qualitatively and quantitatively,
SADis surpasses state-of-the-art stylization methods in the DisIG task.Code
will be released at https://deepffff.github.io/sadis.github.io/.
|
2503.14878 | Murtaza Zohair | Murtaza Zohair, Vidushi Sharma, Eduardo A. Soares, Khanh Nguyen,
Maxwell Giammona, Linda Sundberg, Andy Tek, Emilio A. V. Vital, Young-Hye La | Chemical Foundation Model Guided Design of High Ionic Conductivity
Electrolyte Formulations | null | null | null | null | cond-mat.mtrl-sci physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Designing optimal formulations is a major challenge in developing
electrolytes for the next generation of rechargeable batteries due to the vast
combinatorial design space and complex interplay between multiple constituents.
Machine learning (ML) offers a powerful tool to uncover underlying chemical
design rules and accelerate the process of formulation discovery. In this work,
we present an approach to design new formulations that can achieve target
performance, using a generalizable chemical foundation model. The chemical
foundation model is fine-tuned on an experimental dataset of 13,666 ionic
conductivity values curated from the lithium-ion battery literature. The
fine-tuned model is used to discover 7 novel high conductivity electrolyte
formulations through generative screening, improving the conductivity of LiFSI
and LiDFOB based electrolytes by 82% and 172%, respectively. These findings
highlight a generalizable workflow that is highly adaptable to the discovery of
chemical mixtures with tailored properties to address challenges in energy
storage and beyond.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 04:14:19 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 18:50:35 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zohair",
"Murtaza",
""
],
[
"Sharma",
"Vidushi",
""
],
[
"Soares",
"Eduardo A.",
""
],
[
"Nguyen",
"Khanh",
""
],
[
"Giammona",
"Maxwell",
""
],
[
"Sundberg",
"Linda",
""
],
[
"Tek",
"Andy",
""
],
[
"Vital",
"Emilio A. V.",
""
],
[
"La",
"Young-Hye",
""
]
] | TITLE: Chemical Foundation Model Guided Design of High Ionic Conductivity
Electrolyte Formulations
ABSTRACT: Designing optimal formulations is a major challenge in developing
electrolytes for the next generation of rechargeable batteries due to the vast
combinatorial design space and complex interplay between multiple constituents.
Machine learning (ML) offers a powerful tool to uncover underlying chemical
design rules and accelerate the process of formulation discovery. In this work,
we present an approach to design new formulations that can achieve target
performance, using a generalizable chemical foundation model. The chemical
foundation model is fine-tuned on an experimental dataset of 13,666 ionic
conductivity values curated from the lithium-ion battery literature. The
fine-tuned model is used to discover 7 novel high conductivity electrolyte
formulations through generative screening, improving the conductivity of LiFSI
and LiDFOB based electrolytes by 82% and 172%, respectively. These findings
highlight a generalizable workflow that is highly adaptable to the discovery of
chemical mixtures with tailored properties to address challenges in energy
storage and beyond.
|
2503.14897 | Vaibhav Rathore | Vaibhav Rathore, Shubhranil B, Saikat Dutta, Sarthak Mehrotra, Zsolt
Kira, Biplab Banerjee | When Domain Generalization meets Generalized Category Discovery: An
Adaptive Task-Arithmetic Driven Approach | Accepted at CVPR 2025 (Main Conference) | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Generalized Class Discovery (GCD) clusters base and novel classes in a target
domain using supervision from a source domain with only base classes. Current
methods often falter with distribution shifts and typically require access to
target data during training, which can sometimes be impractical. To address
this issue, we introduce the novel paradigm of Domain Generalization in GCD
(DG-GCD), where only source data is available for training, while the target
domain, with a distinct data distribution, remains unseen until inference. To
this end, our solution, DG2CD-Net, aims to construct a domain-independent,
discriminative embedding space for GCD. The core innovation is an episodic
training strategy that enhances cross-domain generalization by adapting a base
model on tasks derived from source and synthetic domains generated by a
foundation model. Each episode focuses on a cross-domain GCD task, diversifying
task setups over episodes and combining open-set domain adaptation with a novel
margin loss and representation learning for optimizing the feature space
progressively. To capture the effects of fine-tuning on the base model, we
extend task arithmetic by adaptively weighting the local task vectors
concerning the fine-tuned models based on their GCD performance on a validation
distribution. This episodic update mechanism boosts the adaptability of the
base model to unseen targets. Experiments across three datasets confirm that
DG2CD-Net outperforms existing GCD methods customized for DG-GCD.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 04:48:16 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 14:15:36 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Rathore",
"Vaibhav",
""
],
[
"B",
"Shubhranil",
""
],
[
"Dutta",
"Saikat",
""
],
[
"Mehrotra",
"Sarthak",
""
],
[
"Kira",
"Zsolt",
""
],
[
"Banerjee",
"Biplab",
""
]
] | TITLE: When Domain Generalization meets Generalized Category Discovery: An
Adaptive Task-Arithmetic Driven Approach
ABSTRACT: Generalized Class Discovery (GCD) clusters base and novel classes in a target
domain using supervision from a source domain with only base classes. Current
methods often falter with distribution shifts and typically require access to
target data during training, which can sometimes be impractical. To address
this issue, we introduce the novel paradigm of Domain Generalization in GCD
(DG-GCD), where only source data is available for training, while the target
domain, with a distinct data distribution, remains unseen until inference. To
this end, our solution, DG2CD-Net, aims to construct a domain-independent,
discriminative embedding space for GCD. The core innovation is an episodic
training strategy that enhances cross-domain generalization by adapting a base
model on tasks derived from source and synthetic domains generated by a
foundation model. Each episode focuses on a cross-domain GCD task, diversifying
task setups over episodes and combining open-set domain adaptation with a novel
margin loss and representation learning for optimizing the feature space
progressively. To capture the effects of fine-tuning on the base model, we
extend task arithmetic by adaptively weighting the local task vectors
concerning the fine-tuned models based on their GCD performance on a validation
distribution. This episodic update mechanism boosts the adaptability of the
base model to unseen targets. Experiments across three datasets confirm that
DG2CD-Net outperforms existing GCD methods customized for DG-GCD.
|
2503.15300 | Weixiao Gao | Weixiao Gao, Liangliang Nan, Hugo Ledoux | SUM Parts: Benchmarking Part-Level Semantic Segmentation of Urban Meshes | CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Semantic segmentation in urban scene analysis has mainly focused on images or
point clouds, while textured meshes - offering richer spatial representation -
remain underexplored. This paper introduces SUM Parts, the first large-scale
dataset for urban textured meshes with part-level semantic labels, covering
about 2.5 km2 with 21 classes. The dataset was created using our own annotation
tool, which supports both face- and texture-based annotations with efficient
interactive selection. We also provide a comprehensive evaluation of 3D
semantic segmentation and interactive annotation methods on this dataset. Our
project page is available at https://tudelft3d.github.io/SUMParts/.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 15:22:23 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 13:58:31 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Gao",
"Weixiao",
""
],
[
"Nan",
"Liangliang",
""
],
[
"Ledoux",
"Hugo",
""
]
] | TITLE: SUM Parts: Benchmarking Part-Level Semantic Segmentation of Urban Meshes
ABSTRACT: Semantic segmentation in urban scene analysis has mainly focused on images or
point clouds, while textured meshes - offering richer spatial representation -
remain underexplored. This paper introduces SUM Parts, the first large-scale
dataset for urban textured meshes with part-level semantic labels, covering
about 2.5 km2 with 21 classes. The dataset was created using our own annotation
tool, which supports both face- and texture-based annotations with efficient
interactive selection. We also provide a comprehensive evaluation of 3D
semantic segmentation and interactive annotation methods on this dataset. Our
project page is available at https://tudelft3d.github.io/SUMParts/.
|
2503.15463 | Jianan Li | Jia-Nan Li, Jian Guan, Songhao Wu, Wei Wu, Rui Yan | From 1,000,000 Users to Every User: Scaling Up Personalized Preference
for User-level Alignment | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Large language models (LLMs) have traditionally been aligned through
one-size-fits-all approaches that assume uniform human preferences,
fundamentally overlooking the diversity in user values and needs. This paper
introduces a comprehensive framework for scalable personalized alignment of
LLMs. We establish a systematic preference space characterizing psychological
and behavioral dimensions, alongside diverse persona representations for robust
preference inference in real-world scenarios. Building upon this foundation, we
introduce \textsc{AlignX}, a large-scale dataset of over 1.3 million
personalized preference examples, and develop two complementary alignment
approaches: \textit{in-context alignment} directly conditioning on persona
representations and \textit{preference-bridged alignment} modeling intermediate
preference distributions. Extensive experiments demonstrate substantial
improvements over existing methods, with an average 17.06\% accuracy gain
across four benchmarks while exhibiting a strong adaptation capability to novel
preferences, robustness to limited user data, and precise preference
controllability. These results validate our framework's effectiveness,
advancing toward truly user-adaptive AI systems.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 17:41:46 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 10:33:21 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Jia-Nan",
""
],
[
"Guan",
"Jian",
""
],
[
"Wu",
"Songhao",
""
],
[
"Wu",
"Wei",
""
],
[
"Yan",
"Rui",
""
]
] | TITLE: From 1,000,000 Users to Every User: Scaling Up Personalized Preference
for User-level Alignment
ABSTRACT: Large language models (LLMs) have traditionally been aligned through
one-size-fits-all approaches that assume uniform human preferences,
fundamentally overlooking the diversity in user values and needs. This paper
introduces a comprehensive framework for scalable personalized alignment of
LLMs. We establish a systematic preference space characterizing psychological
and behavioral dimensions, alongside diverse persona representations for robust
preference inference in real-world scenarios. Building upon this foundation, we
introduce \textsc{AlignX}, a large-scale dataset of over 1.3 million
personalized preference examples, and develop two complementary alignment
approaches: \textit{in-context alignment} directly conditioning on persona
representations and \textit{preference-bridged alignment} modeling intermediate
preference distributions. Extensive experiments demonstrate substantial
improvements over existing methods, with an average 17.06\% accuracy gain
across four benchmarks while exhibiting a strong adaptation capability to novel
preferences, robustness to limited user data, and precise preference
controllability. These results validate our framework's effectiveness,
advancing toward truly user-adaptive AI systems.
|
2503.15868 | Soumitri Chattopadhyay | Debabrata Mandal and Soumitri Chattopadhyay and Guansen Tong and
Praneeth Chakravarthula | UniCoRN: Latent Diffusion-based Unified Controllable Image Restoration
Network across Multiple Degradations | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Image restoration is essential for enhancing degraded images across computer
vision tasks. However, most existing methods address only a single type of
degradation (e.g., blur, noise, or haze) at a time, limiting their real-world
applicability where multiple degradations often occur simultaneously. In this
paper, we propose UniCoRN, a unified image restoration approach capable of
handling multiple degradation types simultaneously using a multi-head diffusion
model. Specifically, we uncover the potential of low-level visual cues
extracted from images in guiding a controllable diffusion model for real-world
image restoration and we design a multi-head control network adaptable via a
mixture-of-experts strategy. We train our model without any prior assumption of
specific degradations, through a smartly designed curriculum learning recipe.
Additionally, we also introduce MetaRestore, a metalens imaging benchmark
containing images with multiple degradations and artifacts. Extensive
evaluations on several challenging datasets, including our benchmark,
demonstrate that our method achieves significant performance gains and can
robustly restore images with severe degradations. Project page:
https://codejaeger.github.io/unicorn-gh
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 05:42:13 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 15:24:45 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Mandal",
"Debabrata",
""
],
[
"Chattopadhyay",
"Soumitri",
""
],
[
"Tong",
"Guansen",
""
],
[
"Chakravarthula",
"Praneeth",
""
]
] | TITLE: UniCoRN: Latent Diffusion-based Unified Controllable Image Restoration
Network across Multiple Degradations
ABSTRACT: Image restoration is essential for enhancing degraded images across computer
vision tasks. However, most existing methods address only a single type of
degradation (e.g., blur, noise, or haze) at a time, limiting their real-world
applicability where multiple degradations often occur simultaneously. In this
paper, we propose UniCoRN, a unified image restoration approach capable of
handling multiple degradation types simultaneously using a multi-head diffusion
model. Specifically, we uncover the potential of low-level visual cues
extracted from images in guiding a controllable diffusion model for real-world
image restoration and we design a multi-head control network adaptable via a
mixture-of-experts strategy. We train our model without any prior assumption of
specific degradations, through a smartly designed curriculum learning recipe.
Additionally, we also introduce MetaRestore, a metalens imaging benchmark
containing images with multiple degradations and artifacts. Extensive
evaluations on several challenging datasets, including our benchmark,
demonstrate that our method achieves significant performance gains and can
robustly restore images with severe degradations. Project page:
https://codejaeger.github.io/unicorn-gh
|
2503.15879 | DongGeon Lee | DongGeon Lee, Ahjeong Park, Hyeri Lee, Hyeonseo Nam, Yunho Maeng | Typed-RAG: Type-aware Multi-Aspect Decomposition for Non-Factoid
Question Answering | Accepted to NAACL 2025 SRW | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | Non-factoid question-answering (NFQA) poses a significant challenge due to
its open-ended nature, diverse intents, and the need for multi-aspect
reasoning, which renders conventional factoid QA approaches, including
retrieval-augmented generation (RAG), inadequate. Unlike factoid questions,
non-factoid questions (NFQs) lack definitive answers and require synthesizing
information from multiple sources across various reasoning dimensions. To
address these limitations, we introduce Typed-RAG, a type-aware multi-aspect
decomposition framework within the RAG paradigm for NFQA. Typed-RAG classifies
NFQs into distinct types -- such as debate, experience, and comparison -- and
applies aspect-based decomposition to refine retrieval and generation
strategies. By decomposing multi-aspect NFQs into single-aspect sub-queries and
aggregating the results, Typed-RAG generates more informative and contextually
relevant responses. To evaluate Typed-RAG, we introduce Wiki-NFQA, a benchmark
dataset covering diverse NFQ types. Experimental results demonstrate that
Typed-RAG outperforms baselines, thereby highlighting the importance of
type-aware decomposition for effective retrieval and generation in NFQA. Our
code and dataset are available at https://github.com/TeamNLP/Typed-RAG.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 06:04:12 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 05:50:37 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Lee",
"DongGeon",
""
],
[
"Park",
"Ahjeong",
""
],
[
"Lee",
"Hyeri",
""
],
[
"Nam",
"Hyeonseo",
""
],
[
"Maeng",
"Yunho",
""
]
] | TITLE: Typed-RAG: Type-aware Multi-Aspect Decomposition for Non-Factoid
Question Answering
ABSTRACT: Non-factoid question-answering (NFQA) poses a significant challenge due to
its open-ended nature, diverse intents, and the need for multi-aspect
reasoning, which renders conventional factoid QA approaches, including
retrieval-augmented generation (RAG), inadequate. Unlike factoid questions,
non-factoid questions (NFQs) lack definitive answers and require synthesizing
information from multiple sources across various reasoning dimensions. To
address these limitations, we introduce Typed-RAG, a type-aware multi-aspect
decomposition framework within the RAG paradigm for NFQA. Typed-RAG classifies
NFQs into distinct types -- such as debate, experience, and comparison -- and
applies aspect-based decomposition to refine retrieval and generation
strategies. By decomposing multi-aspect NFQs into single-aspect sub-queries and
aggregating the results, Typed-RAG generates more informative and contextually
relevant responses. To evaluate Typed-RAG, we introduce Wiki-NFQA, a benchmark
dataset covering diverse NFQ types. Experimental results demonstrate that
Typed-RAG outperforms baselines, thereby highlighting the importance of
type-aware decomposition for effective retrieval and generation in NFQA. Our
code and dataset are available at https://github.com/TeamNLP/Typed-RAG.
|
2503.15886 | Hui Liu | Hui Liu, Wenya Wang, Kecheng Chen, Jie Liu, Yibing Liu, Tiexin Qin,
Peisong He, Xinghao Jiang, Haoliang Li | Enhancing Zero-Shot Image Recognition in Vision-Language Models through
Human-like Concept Guidance | 21 pages, 7 figures 7 tables | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | In zero-shot image recognition tasks, humans demonstrate remarkable
flexibility in classifying unseen categories by composing known simpler
concepts. However, existing vision-language models (VLMs), despite achieving
significant progress through large-scale natural language supervision, often
underperform in real-world applications because of sub-optimal prompt
engineering and the inability to adapt effectively to target classes. To
address these issues, we propose a Concept-guided Human-like Bayesian Reasoning
(CHBR) framework. Grounded in Bayes' theorem, CHBR models the concept used in
human image recognition as latent variables and formulates this task by summing
across potential concepts, weighted by a prior distribution and a likelihood
function. To tackle the intractable computation over an infinite concept space,
we introduce an importance sampling algorithm that iteratively prompts large
language models (LLMs) to generate discriminative concepts, emphasizing
inter-class differences. We further propose three heuristic approaches
involving Average Likelihood, Confidence Likelihood, and Test Time Augmentation
(TTA) Likelihood, which dynamically refine the combination of concepts based on
the test image. Extensive evaluations across fifteen datasets demonstrate that
CHBR consistently outperforms existing state-of-the-art zero-shot
generalization methods.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 06:20:13 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 02:55:26 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Liu",
"Hui",
""
],
[
"Wang",
"Wenya",
""
],
[
"Chen",
"Kecheng",
""
],
[
"Liu",
"Jie",
""
],
[
"Liu",
"Yibing",
""
],
[
"Qin",
"Tiexin",
""
],
[
"He",
"Peisong",
""
],
[
"Jiang",
"Xinghao",
""
],
[
"Li",
"Haoliang",
""
]
] | TITLE: Enhancing Zero-Shot Image Recognition in Vision-Language Models through
Human-like Concept Guidance
ABSTRACT: In zero-shot image recognition tasks, humans demonstrate remarkable
flexibility in classifying unseen categories by composing known simpler
concepts. However, existing vision-language models (VLMs), despite achieving
significant progress through large-scale natural language supervision, often
underperform in real-world applications because of sub-optimal prompt
engineering and the inability to adapt effectively to target classes. To
address these issues, we propose a Concept-guided Human-like Bayesian Reasoning
(CHBR) framework. Grounded in Bayes' theorem, CHBR models the concept used in
human image recognition as latent variables and formulates this task by summing
across potential concepts, weighted by a prior distribution and a likelihood
function. To tackle the intractable computation over an infinite concept space,
we introduce an importance sampling algorithm that iteratively prompts large
language models (LLMs) to generate discriminative concepts, emphasizing
inter-class differences. We further propose three heuristic approaches
involving Average Likelihood, Confidence Likelihood, and Test Time Augmentation
(TTA) Likelihood, which dynamically refine the combination of concepts based on
the test image. Extensive evaluations across fifteen datasets demonstrate that
CHBR consistently outperforms existing state-of-the-art zero-shot
generalization methods.
|
2503.16047 | Akinyemi Sadeeq Akintola | Bisola Faith Kayode, Akinyemi Sadeeq Akintola, Oluwole Fagbohun,
Egonna Anaesiuba-Bristol, Onyekachukwu Ojumah, Oluwagbade Odimayo, Toyese
Oloyede, Aniema Inyang, Teslim Kazeem, Habeeb Alli, Udodirim Ibem Offia,
Prisca Chinazor Amajuoyi | Temporal-Spatial Attention Network (TSAN) for DoS Attack Detection in
Network Traffic | 19 Pages, 5 figures | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Denial-of-Service (DoS) attacks remain a critical threat to network security,
disrupting services and causing significant economic losses. Traditional
detection methods, including statistical and rule-based models, struggle to
adapt to evolving attack patterns. To address this challenge, we propose a
novel Temporal-Spatial Attention Network (TSAN) architecture for detecting
Denial of Service (DoS) attacks in network traffic. By leveraging both temporal
and spatial features of network traffic, our approach captures complex traffic
patterns and anomalies that traditional methods might miss. The TSAN model
incorporates transformer-based temporal encoding, convolutional spatial
encoding, and a cross-attention mechanism to fuse these complementary feature
spaces. Additionally, we employ multi-task learning with auxiliary tasks to
enhance the model's robustness. Experimental results on the NSL-KDD dataset
demonstrate that TSAN outperforms state-of-the-art models, achieving superior
accuracy, precision, recall, and F1-score while maintaining computational
efficiency for real-time deployment. The proposed architecture offers an
optimal balance between detection accuracy and computational overhead, making
it highly suitable for real-world network security applications.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 11:31:45 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 17:40:15 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kayode",
"Bisola Faith",
""
],
[
"Akintola",
"Akinyemi Sadeeq",
""
],
[
"Fagbohun",
"Oluwole",
""
],
[
"Anaesiuba-Bristol",
"Egonna",
""
],
[
"Ojumah",
"Onyekachukwu",
""
],
[
"Odimayo",
"Oluwagbade",
""
],
[
"Oloyede",
"Toyese",
""
],
[
"Inyang",
"Aniema",
""
],
[
"Kazeem",
"Teslim",
""
],
[
"Alli",
"Habeeb",
""
],
[
"Offia",
"Udodirim Ibem",
""
],
[
"Amajuoyi",
"Prisca Chinazor",
""
]
] | TITLE: Temporal-Spatial Attention Network (TSAN) for DoS Attack Detection in
Network Traffic
ABSTRACT: Denial-of-Service (DoS) attacks remain a critical threat to network security,
disrupting services and causing significant economic losses. Traditional
detection methods, including statistical and rule-based models, struggle to
adapt to evolving attack patterns. To address this challenge, we propose a
novel Temporal-Spatial Attention Network (TSAN) architecture for detecting
Denial of Service (DoS) attacks in network traffic. By leveraging both temporal
and spatial features of network traffic, our approach captures complex traffic
patterns and anomalies that traditional methods might miss. The TSAN model
incorporates transformer-based temporal encoding, convolutional spatial
encoding, and a cross-attention mechanism to fuse these complementary feature
spaces. Additionally, we employ multi-task learning with auxiliary tasks to
enhance the model's robustness. Experimental results on the NSL-KDD dataset
demonstrate that TSAN outperforms state-of-the-art models, achieving superior
accuracy, precision, recall, and F1-score while maintaining computational
efficiency for real-time deployment. The proposed architecture offers an
optimal balance between detection accuracy and computational overhead, making
it highly suitable for real-world network security applications.
|
2503.16252 | Liwen Zhang | Zhaowei Liu, Xin Guo, Fangqi Lou, Lingfeng Zeng, Jinyi Niu, Zixuan
Wang, Jiajie Xu, Weige Cai, Ziwei Yang, Xueqian Zhao, Chao Li, Sheng Xu,
Dezhi Chen, Yun Chen, Zuo Bai and Liwen Zhang | Fin-R1: A Large Language Model for Financial Reasoning through
Reinforcement Learning | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reasoning large language models are rapidly evolving across various domains.
However, their capabilities in handling complex financial tasks still require
in-depth exploration. In this paper, we introduce Fin-R1, a reasoning large
language model specifically designed for the financial sector. Fin-R1 is built
using a two-stage architecture, leveraging a financial reasoning dataset
distilled and processed based on DeepSeek-R1. Through supervised fine-tuning
(SFT) and reinforcement learning (RL) training, it demonstrates performance
close to DeepSeek-R1 with a parameter size of 7 billion across a range of
financial reasoning tasks. It achieves the state-of-the-art (SOTA) in the FinQA
and ConvFinQA tasks between those LLMs in our evaluation, surpassing larger
models in other tasks as well. Fin-R1 showcases strong reasoning and
decision-making capabilities, providing solutions to various problems
encountered in the financial domain. Our code is available at
https://github.com/SUFE-AIFLM-Lab/Fin-R1.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 15:46:18 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 01:57:58 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Liu",
"Zhaowei",
""
],
[
"Guo",
"Xin",
""
],
[
"Lou",
"Fangqi",
""
],
[
"Zeng",
"Lingfeng",
""
],
[
"Niu",
"Jinyi",
""
],
[
"Wang",
"Zixuan",
""
],
[
"Xu",
"Jiajie",
""
],
[
"Cai",
"Weige",
""
],
[
"Yang",
"Ziwei",
""
],
[
"Zhao",
"Xueqian",
""
],
[
"Li",
"Chao",
""
],
[
"Xu",
"Sheng",
""
],
[
"Chen",
"Dezhi",
""
],
[
"Chen",
"Yun",
""
],
[
"Bai",
"Zuo",
""
],
[
"Zhang",
"Liwen",
""
]
] | TITLE: Fin-R1: A Large Language Model for Financial Reasoning through
Reinforcement Learning
ABSTRACT: Reasoning large language models are rapidly evolving across various domains.
However, their capabilities in handling complex financial tasks still require
in-depth exploration. In this paper, we introduce Fin-R1, a reasoning large
language model specifically designed for the financial sector. Fin-R1 is built
using a two-stage architecture, leveraging a financial reasoning dataset
distilled and processed based on DeepSeek-R1. Through supervised fine-tuning
(SFT) and reinforcement learning (RL) training, it demonstrates performance
close to DeepSeek-R1 with a parameter size of 7 billion across a range of
financial reasoning tasks. It achieves the state-of-the-art (SOTA) in the FinQA
and ConvFinQA tasks between those LLMs in our evaluation, surpassing larger
models in other tasks as well. Fin-R1 showcases strong reasoning and
decision-making capabilities, providing solutions to various problems
encountered in the financial domain. Our code is available at
https://github.com/SUFE-AIFLM-Lab/Fin-R1.
|
2503.16454 | Haidong Wang | Haidong Wang, Qia Shan, JianHua Zhang, PengFei Xiao, Ao Liu | An Audio-Visual Fusion Emotion Generation Model Based on Neuroanatomical
Alignment | null | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by/4.0/ | In the field of affective computing, traditional methods for generating
emotions predominantly rely on deep learning techniques and large-scale emotion
datasets. However, deep learning techniques are often complex and difficult to
interpret, and standardizing large-scale emotional datasets are difficult and
costly to establish. To tackle these challenges, we introduce a novel framework
named Audio-Visual Fusion for Brain-like Emotion Learning(AVF-BEL). In contrast
to conventional brain-inspired emotion learning methods, this approach improves
the audio-visual emotion fusion and generation model through the integration of
modular components, thereby enabling more lightweight and interpretable emotion
learning and generation processes. The framework simulates the integration of
the visual, auditory, and emotional pathways of the brain, optimizes the fusion
of emotional features across visual and auditory modalities, and improves upon
the traditional Brain Emotional Learning (BEL) model. The experimental results
indicate a significant improvement in the similarity of the audio-visual fusion
emotion learning generation model compared to single-modality visual and
auditory emotion learning and generation model. Ultimately, this aligns with
the fundamental phenomenon of heightened emotion generation facilitated by the
integrated impact of visual and auditory stimuli. This contribution not only
enhances the interpretability and efficiency of affective intelligence but also
provides new insights and pathways for advancing affective computing
technology. Our source code can be accessed here:
https://github.com/OpenHUTB/emotion}{https://github.com/OpenHUTB/emotion.
| [
{
"version": "v1",
"created": "Fri, 21 Feb 2025 14:26:58 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Wang",
"Haidong",
""
],
[
"Shan",
"Qia",
""
],
[
"Zhang",
"JianHua",
""
],
[
"Xiao",
"PengFei",
""
],
[
"Liu",
"Ao",
""
]
] | TITLE: An Audio-Visual Fusion Emotion Generation Model Based on Neuroanatomical
Alignment
ABSTRACT: In the field of affective computing, traditional methods for generating
emotions predominantly rely on deep learning techniques and large-scale emotion
datasets. However, deep learning techniques are often complex and difficult to
interpret, and standardizing large-scale emotional datasets are difficult and
costly to establish. To tackle these challenges, we introduce a novel framework
named Audio-Visual Fusion for Brain-like Emotion Learning(AVF-BEL). In contrast
to conventional brain-inspired emotion learning methods, this approach improves
the audio-visual emotion fusion and generation model through the integration of
modular components, thereby enabling more lightweight and interpretable emotion
learning and generation processes. The framework simulates the integration of
the visual, auditory, and emotional pathways of the brain, optimizes the fusion
of emotional features across visual and auditory modalities, and improves upon
the traditional Brain Emotional Learning (BEL) model. The experimental results
indicate a significant improvement in the similarity of the audio-visual fusion
emotion learning generation model compared to single-modality visual and
auditory emotion learning and generation model. Ultimately, this aligns with
the fundamental phenomenon of heightened emotion generation facilitated by the
integrated impact of visual and auditory stimuli. This contribution not only
enhances the interpretability and efficiency of affective intelligence but also
provides new insights and pathways for advancing affective computing
technology. Our source code can be accessed here:
https://github.com/OpenHUTB/emotion}{https://github.com/OpenHUTB/emotion.
|
2503.16465 | Pengzhou Cheng | Pengzhou Cheng, Zheng Wu, Zongru Wu, Aston Zhang, Zhuosheng Zhang,
Gongshen Liu | OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents | 25 pages, 24 figures, 11 tables | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by/4.0/ | Autonomous graphical user interface (GUI) agents powered by multimodal large
language models have shown great promise. However, a critical yet underexplored
issue persists: over-execution, where the agent executes tasks in a fully
autonomous way, without adequate assessment of its action confidence to
compromise an adaptive human-agent collaboration. This poses substantial risks
in complex scenarios, such as those involving ambiguous user instructions,
unexpected interruptions, and environmental hijacks. To address the issue, we
introduce OS-Kairos, an adaptive GUI agent capable of predicting confidence
levels at each interaction step and efficiently deciding whether to act
autonomously or seek human intervention. OS-Kairos is developed through two key
mechanisms: (i) collaborative probing that annotates confidence scores at each
interaction step; (ii) confidence-driven interaction that leverages these
confidence scores to elicit the ability of adaptive interaction. Experimental
results show that OS-Kairos substantially outperforms existing models on our
curated dataset featuring complex scenarios, as well as on established
benchmarks such as AITZ and Meta-GUI, with 24.59\%$\sim$87.29\% improvements in
task success rate. OS-Kairos facilitates an adaptive human-agent collaboration,
prioritizing effectiveness, generality, scalability, and efficiency for
real-world GUI interaction. The dataset and codes are available at
https://github.com/Wuzheng02/OS-Kairos.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2025 12:31:16 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Cheng",
"Pengzhou",
""
],
[
"Wu",
"Zheng",
""
],
[
"Wu",
"Zongru",
""
],
[
"Zhang",
"Aston",
""
],
[
"Zhang",
"Zhuosheng",
""
],
[
"Liu",
"Gongshen",
""
]
] | TITLE: OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents
ABSTRACT: Autonomous graphical user interface (GUI) agents powered by multimodal large
language models have shown great promise. However, a critical yet underexplored
issue persists: over-execution, where the agent executes tasks in a fully
autonomous way, without adequate assessment of its action confidence to
compromise an adaptive human-agent collaboration. This poses substantial risks
in complex scenarios, such as those involving ambiguous user instructions,
unexpected interruptions, and environmental hijacks. To address the issue, we
introduce OS-Kairos, an adaptive GUI agent capable of predicting confidence
levels at each interaction step and efficiently deciding whether to act
autonomously or seek human intervention. OS-Kairos is developed through two key
mechanisms: (i) collaborative probing that annotates confidence scores at each
interaction step; (ii) confidence-driven interaction that leverages these
confidence scores to elicit the ability of adaptive interaction. Experimental
results show that OS-Kairos substantially outperforms existing models on our
curated dataset featuring complex scenarios, as well as on established
benchmarks such as AITZ and Meta-GUI, with 24.59\%$\sim$87.29\% improvements in
task success rate. OS-Kairos facilitates an adaptive human-agent collaboration,
prioritizing effectiveness, generality, scalability, and efficiency for
real-world GUI interaction. The dataset and codes are available at
https://github.com/Wuzheng02/OS-Kairos.
|
2503.16478 | Rishabh Vishwakarma | Rishabh Vishwakarma, Caroline Brophy, and Catherine Hurley | PieGlyph: An R package for creating axis invariant pie-glyphs for 2d
plots | PieGlyph is available on CRAN here:
https://CRAN.R-project.org/package=PieGlyph While the development version can
be downloaded from Github at: https://github.com/rishvish/PieGlyph Package
vignettes are available here: https://rishvish.github.io/PieGlyph/ | null | null | null | cs.HC stat.AP | http://creativecommons.org/licenses/by-sa/4.0/ | Effective visualisation of multidimensional data is crucial for generating
insights. Glyph-based visualisations, which encode data dimensions onto
multiple visual channels such as colour, shape, and size, provide an effective
means of representing complex datasets. Pie-chart glyphs (pie-glyphs) are one
such approach, where multiple data attributes are mapped to slices within a pie
chart. This paper introduces the PieGlyph R package, which enables users to
overlay any 2D plot with axis-invariant pie-glyphs, offering a compact and
intuitive representation of multidimensional data. Unlike existing R packages
such as scatterpie or ggforce, PieGlyph generates pie-glyphs independently of
the plot axes by employing a nested coordinate system, ensuring they remain
circular regardless of changes to the underlying coordinate system. This
enhances interpretability, particularly in when visualising spatial data, as
users can select the most appropriate map projection without distorting the
glyphs' shape. Pie-glyphs are also particularly well-suited for visualising
compositional data, where there is a natural sum-to-one constraint on the data
attributes. PieGlyph is developed under the Grammar of Graphics paradigm using
the ggplot2 framework and supports the generation of interactive pie-glyphs
through the ggiraph package. Designed to integrate seamlessly with all features
and extensions offered by ggplot2 and ggiraph, PieGlyph provides users with
full flexibility in customising every aspect of the visualisation. This paper
outlines the conceptual framework of PieGlyph, compares it with existing
alternatives, and demonstrates its applications through example visualisations.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2025 10:53:04 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Vishwakarma",
"Rishabh",
""
],
[
"Brophy",
"Caroline",
""
],
[
"Hurley",
"Catherine",
""
]
] | TITLE: PieGlyph: An R package for creating axis invariant pie-glyphs for 2d
plots
ABSTRACT: Effective visualisation of multidimensional data is crucial for generating
insights. Glyph-based visualisations, which encode data dimensions onto
multiple visual channels such as colour, shape, and size, provide an effective
means of representing complex datasets. Pie-chart glyphs (pie-glyphs) are one
such approach, where multiple data attributes are mapped to slices within a pie
chart. This paper introduces the PieGlyph R package, which enables users to
overlay any 2D plot with axis-invariant pie-glyphs, offering a compact and
intuitive representation of multidimensional data. Unlike existing R packages
such as scatterpie or ggforce, PieGlyph generates pie-glyphs independently of
the plot axes by employing a nested coordinate system, ensuring they remain
circular regardless of changes to the underlying coordinate system. This
enhances interpretability, particularly in when visualising spatial data, as
users can select the most appropriate map projection without distorting the
glyphs' shape. Pie-glyphs are also particularly well-suited for visualising
compositional data, where there is a natural sum-to-one constraint on the data
attributes. PieGlyph is developed under the Grammar of Graphics paradigm using
the ggplot2 framework and supports the generation of interactive pie-glyphs
through the ggiraph package. Designed to integrate seamlessly with all features
and extensions offered by ggplot2 and ggiraph, PieGlyph provides users with
full flexibility in customising every aspect of the visualisation. This paper
outlines the conceptual framework of PieGlyph, compares it with existing
alternatives, and demonstrates its applications through example visualisations.
|
2503.16480 | Ramit Debnath | Yara Kyrychenko, Jon Roozenbeek, Brandon Davidson, Sander van der
Linden and Ramit Debnath | Human Preferences for Constructive Interactions in Language Model
Alignment | 1 Figure, 1 Table, 11 pages | null | null | null | cs.HC cs.AI cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | As large language models (LLMs) enter the mainstream, aligning them to foster
constructive dialogue rather than exacerbate societal divisions is critical.
Using an individualized and multicultural alignment dataset of over 7,500
conversations of individuals from 74 countries engaging with 21 LLMs, we
examined how linguistic attributes linked to constructive interactions are
reflected in human preference data used for training AI. We found that users
consistently preferred well-reasoned and nuanced responses while rejecting
those high in personal storytelling. However, users who believed that AI should
reflect their values tended to place less preference on reasoning in LLM
responses and more on curiosity. Encouragingly, we observed that users could
set the tone for how constructive their conversation would be, as LLMs mirrored
linguistic attributes, including toxicity, in user queries.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2025 15:08:41 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kyrychenko",
"Yara",
""
],
[
"Roozenbeek",
"Jon",
""
],
[
"Davidson",
"Brandon",
""
],
[
"van der Linden",
"Sander",
""
],
[
"Debnath",
"Ramit",
""
]
] | TITLE: Human Preferences for Constructive Interactions in Language Model
Alignment
ABSTRACT: As large language models (LLMs) enter the mainstream, aligning them to foster
constructive dialogue rather than exacerbate societal divisions is critical.
Using an individualized and multicultural alignment dataset of over 7,500
conversations of individuals from 74 countries engaging with 21 LLMs, we
examined how linguistic attributes linked to constructive interactions are
reflected in human preference data used for training AI. We found that users
consistently preferred well-reasoned and nuanced responses while rejecting
those high in personal storytelling. However, users who believed that AI should
reflect their values tended to place less preference on reasoning in LLM
responses and more on curiosity. Encouragingly, we observed that users could
set the tone for how constructive their conversation would be, as LLMs mirrored
linguistic attributes, including toxicity, in user queries.
|
2503.16481 | Subham Agrawal | Subham Agrawal and Nico Ostermann-Myrau and Nils Dengler and Maren
Bennewitz | Pedestrians and Robots: A Novel Dataset for Learning Distinct Social
Navigation Forces | null | null | null | null | cs.HC cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing use of robots in human-centric public spaces such as shopping
malls, sidewalks, and hospitals, requires understanding of how pedestrians
respond to their presence. However, existing research lacks comprehensive
datasets that capture the full range of pedestrian behaviors, e.g., including
avoidance, neutrality, and attraction in the presence of robots. Such datasets
can be used to effectively learn models capable of accurately predicting
diverse responses of pedestrians to robot presence, which are crucial for
advancing robot navigation strategies and optimizing pedestrian-aware motion
planning. In this paper, we address these challenges by collecting a novel
dataset of pedestrian motion in two outdoor locations under three distinct
conditions, i.e., no robot presence, a stationary robot, and a moving robot.
Thus, unlike existing datasets, ours explicitly encapsulates variations in
pedestrian behavior across the different robot conditions. Using our dataset,
we propose a novel Neural Social Robot Force Model (NSRFM), an extension of the
traditional Social Force Model that integrates neural networks and
robot-induced forces to better predict pedestrian behavior in the presence of
robots. We validate the NSRFM by comparing its generated trajectories on
different real-world datasets. Furthermore, we implemented it in simulation to
enable the learning and benchmarking of robot navigation strategies based on
their impact on pedestrian movement. Our results demonstrate the model's
effectiveness in replicating real-world pedestrian reactions and its its
utility in developing, evaluating, and benchmarking social robot navigation
algorithms.
| [
{
"version": "v1",
"created": "Wed, 5 Mar 2025 17:02:29 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Agrawal",
"Subham",
""
],
[
"Ostermann-Myrau",
"Nico",
""
],
[
"Dengler",
"Nils",
""
],
[
"Bennewitz",
"Maren",
""
]
] | TITLE: Pedestrians and Robots: A Novel Dataset for Learning Distinct Social
Navigation Forces
ABSTRACT: The increasing use of robots in human-centric public spaces such as shopping
malls, sidewalks, and hospitals, requires understanding of how pedestrians
respond to their presence. However, existing research lacks comprehensive
datasets that capture the full range of pedestrian behaviors, e.g., including
avoidance, neutrality, and attraction in the presence of robots. Such datasets
can be used to effectively learn models capable of accurately predicting
diverse responses of pedestrians to robot presence, which are crucial for
advancing robot navigation strategies and optimizing pedestrian-aware motion
planning. In this paper, we address these challenges by collecting a novel
dataset of pedestrian motion in two outdoor locations under three distinct
conditions, i.e., no robot presence, a stationary robot, and a moving robot.
Thus, unlike existing datasets, ours explicitly encapsulates variations in
pedestrian behavior across the different robot conditions. Using our dataset,
we propose a novel Neural Social Robot Force Model (NSRFM), an extension of the
traditional Social Force Model that integrates neural networks and
robot-induced forces to better predict pedestrian behavior in the presence of
robots. We validate the NSRFM by comparing its generated trajectories on
different real-world datasets. Furthermore, we implemented it in simulation to
enable the learning and benchmarking of robot navigation strategies based on
their impact on pedestrian movement. Our results demonstrate the model's
effectiveness in replicating real-world pedestrian reactions and its its
utility in developing, evaluating, and benchmarking social robot navigation
algorithms.
|
2503.16500 | Jorge de Heuvel | Jorge de Heuvel, Daniel Marta, Simon Holk, Iolanda Leite, Maren
Bennewitz | The Impact of VR and 2D Interfaces on Human Feedback in Preference-Based
Robot Learning | null | null | null | null | cs.HC cs.RO | http://creativecommons.org/licenses/by/4.0/ | Aligning robot navigation with human preferences is essential for ensuring
comfortable and predictable robot movement in shared spaces, facilitating
seamless human-robot coexistence. While preference-based learning methods, such
as reinforcement learning from human feedback (RLHF), enable this alignment,
the choice of the preference collection interface may influence the process.
Traditional 2D interfaces provide structured views but lack spatial depth,
whereas immersive VR offers richer perception, potentially affecting preference
articulation. This study systematically examines how the interface modality
impacts human preference collection and navigation policy alignment. We
introduce a novel dataset of 2,325 human preference queries collected through
both VR and 2D interfaces, revealing significant differences in user
experience, preference consistency, and policy outcomes. Our findings highlight
the trade-offs between immersion, perception, and preference reliability,
emphasizing the importance of interface selection in preference-based robot
learning. The dataset will be publicly released to support future research.
| [
{
"version": "v1",
"created": "Tue, 11 Mar 2025 21:02:47 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"de Heuvel",
"Jorge",
""
],
[
"Marta",
"Daniel",
""
],
[
"Holk",
"Simon",
""
],
[
"Leite",
"Iolanda",
""
],
[
"Bennewitz",
"Maren",
""
]
] | TITLE: The Impact of VR and 2D Interfaces on Human Feedback in Preference-Based
Robot Learning
ABSTRACT: Aligning robot navigation with human preferences is essential for ensuring
comfortable and predictable robot movement in shared spaces, facilitating
seamless human-robot coexistence. While preference-based learning methods, such
as reinforcement learning from human feedback (RLHF), enable this alignment,
the choice of the preference collection interface may influence the process.
Traditional 2D interfaces provide structured views but lack spatial depth,
whereas immersive VR offers richer perception, potentially affecting preference
articulation. This study systematically examines how the interface modality
impacts human preference collection and navigation policy alignment. We
introduce a novel dataset of 2,325 human preference queries collected through
both VR and 2D interfaces, revealing significant differences in user
experience, preference consistency, and policy outcomes. Our findings highlight
the trade-offs between immersion, perception, and preference reliability,
emphasizing the importance of interface selection in preference-based robot
learning. The dataset will be publicly released to support future research.
|
2503.16505 | Dimitrios Tsirmpas | Dimitris Tsirmpas and Ion Androutsopoulos and John Pavlopoulos | Scalable Evaluation of Online Moderation Strategies via Synthetic
Simulations | 25 pages, 6 tables, 9 figures | null | null | null | cs.HC cs.CL cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Despite the ever-growing importance of online moderation, there has been no
large-scale study evaluating the effectiveness of alternative moderation
strategies. This is largely due to the lack of appropriate datasets, and the
difficulty of getting human discussants, moderators, and evaluators involved in
multiple experiments. In this paper, we propose a methodology for leveraging
synthetic experiments performed exclusively by Large Language Models (LLMs) to
initially bypass the need for human participation in experiments involving
online moderation. We evaluate six LLM moderation configurations; two currently
used real-life moderation strategies (guidelines issued for human moderators
for online moderation and real-life facilitation), two baseline strategies
(guidelines elicited for LLM alignment work, and LLM moderation with minimal
prompting) a baseline with no moderator at all, as well as our own proposed
strategy inspired by a Reinforcement Learning (RL) formulation of the problem.
We find that our own moderation strategy significantly outperforms established
moderation guidelines, as well as out-of-the-box LLM moderation. We also find
that smaller LLMs, with less intensive instruction-tuning, can create more
varied discussions than larger models. In order to run these experiments, we
create and release an efficient, purpose-built, open-source Python framework,
dubbed "SynDisco" to easily simulate hundreds of discussions using LLM
user-agents and moderators. Additionally, we release the Virtual Moderation
Dataset (VMD), a large dataset of LLM-generated and LLM-annotated discussions,
generated by three families of open-source LLMs accompanied by an exploratory
analysis of the dataset.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2025 08:13:07 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Tsirmpas",
"Dimitris",
""
],
[
"Androutsopoulos",
"Ion",
""
],
[
"Pavlopoulos",
"John",
""
]
] | TITLE: Scalable Evaluation of Online Moderation Strategies via Synthetic
Simulations
ABSTRACT: Despite the ever-growing importance of online moderation, there has been no
large-scale study evaluating the effectiveness of alternative moderation
strategies. This is largely due to the lack of appropriate datasets, and the
difficulty of getting human discussants, moderators, and evaluators involved in
multiple experiments. In this paper, we propose a methodology for leveraging
synthetic experiments performed exclusively by Large Language Models (LLMs) to
initially bypass the need for human participation in experiments involving
online moderation. We evaluate six LLM moderation configurations; two currently
used real-life moderation strategies (guidelines issued for human moderators
for online moderation and real-life facilitation), two baseline strategies
(guidelines elicited for LLM alignment work, and LLM moderation with minimal
prompting) a baseline with no moderator at all, as well as our own proposed
strategy inspired by a Reinforcement Learning (RL) formulation of the problem.
We find that our own moderation strategy significantly outperforms established
moderation guidelines, as well as out-of-the-box LLM moderation. We also find
that smaller LLMs, with less intensive instruction-tuning, can create more
varied discussions than larger models. In order to run these experiments, we
create and release an efficient, purpose-built, open-source Python framework,
dubbed "SynDisco" to easily simulate hundreds of discussions using LLM
user-agents and moderators. Additionally, we release the Virtual Moderation
Dataset (VMD), a large dataset of LLM-generated and LLM-annotated discussions,
generated by three families of open-source LLMs accompanied by an exploratory
analysis of the dataset.
|
2503.16511 | Tingkai Liu | Tingkai Liu, Ari S. Benjamin, Anthony M. Zador | Token-Level Uncertainty-Aware Objective for Language Model Post-Training | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the current work, we connect token-level uncertainty in causal language
modeling to two types of training objectives: 1) masked maximum likelihood
(MLE), 2) self-distillation. We show that masked MLE is effective in reducing
epistemic uncertainty, and serve as an effective token-level automatic
curriculum learning technique. However, masked MLE is prone to overfitting and
requires self-distillation regularization to improve or maintain performance on
out-of-distribution tasks. We demonstrate significant performance gain via the
proposed training objective - combined masked MLE and self-distillation -
across multiple architectures (Gemma, LLaMA, Phi) and datasets (Alpaca,
ShareGPT, GSM8K), mitigating overfitting while maintaining adaptability during
post-training. Our findings suggest that uncertainty-aware training provides an
effective mechanism for enhancing language model training.
| [
{
"version": "v1",
"created": "Sat, 15 Mar 2025 00:32:14 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Liu",
"Tingkai",
""
],
[
"Benjamin",
"Ari S.",
""
],
[
"Zador",
"Anthony M.",
""
]
] | TITLE: Token-Level Uncertainty-Aware Objective for Language Model Post-Training
ABSTRACT: In the current work, we connect token-level uncertainty in causal language
modeling to two types of training objectives: 1) masked maximum likelihood
(MLE), 2) self-distillation. We show that masked MLE is effective in reducing
epistemic uncertainty, and serve as an effective token-level automatic
curriculum learning technique. However, masked MLE is prone to overfitting and
requires self-distillation regularization to improve or maintain performance on
out-of-distribution tasks. We demonstrate significant performance gain via the
proposed training objective - combined masked MLE and self-distillation -
across multiple architectures (Gemma, LLaMA, Phi) and datasets (Alpaca,
ShareGPT, GSM8K), mitigating overfitting while maintaining adaptability during
post-training. Our findings suggest that uncertainty-aware training provides an
effective mechanism for enhancing language model training.
|
2503.16520 | Ji-Eun Han | Ji-Eun Han, Yoonseok Heo | Not All Personas Are Worth It: Culture-Reflective Persona Data
Augmentation | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Incorporating personas into conversational AI models is crucial for achieving
authentic and engaging interactions. However, the cultural diversity and
adaptability of existing persona datasets is often overlooked, reducing their
efficacy in building culturally aware AI systems. To address this issue, we
propose a two-step pipeline for generating culture-specific personas and
introduce KoPersona, a dataset comprising 200,000 personas designed to capture
Korean cultural values, behaviors, and social nuances. A comprehensive
evaluation through various metrics validates the quality of KoPersona and its
relevance to Korean culture. This work not only contributes to persona-based
research, but also establishes a scalable approach for creating culturally
relevant personas adaptable to various languages and cultural contexts.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 01:23:57 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Han",
"Ji-Eun",
""
],
[
"Heo",
"Yoonseok",
""
]
] | TITLE: Not All Personas Are Worth It: Culture-Reflective Persona Data
Augmentation
ABSTRACT: Incorporating personas into conversational AI models is crucial for achieving
authentic and engaging interactions. However, the cultural diversity and
adaptability of existing persona datasets is often overlooked, reducing their
efficacy in building culturally aware AI systems. To address this issue, we
propose a two-step pipeline for generating culture-specific personas and
introduce KoPersona, a dataset comprising 200,000 personas designed to capture
Korean cultural values, behaviors, and social nuances. A comprehensive
evaluation through various metrics validates the quality of KoPersona and its
relevance to Korean culture. This work not only contributes to persona-based
research, but also establishes a scalable approach for creating culturally
relevant personas adaptable to various languages and cultural contexts.
|
2503.16522 | Yongjia Ma | Yongjia Ma, Donglin Di, Xuan Liu, Xiaokai Chen, Lei Fan, Wei Chen,
Tonghua Su | Adams Bashforth Moulton Solver for Inversion and Editing in Rectified
Flow | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Rectified flow models have achieved remarkable performance in image and video
generation tasks. However, existing numerical solvers face a trade-off between
fast sampling and high-accuracy solutions, limiting their effectiveness in
downstream applications such as reconstruction and editing. To address this
challenge, we propose leveraging the Adams-Bashforth-Moulton (ABM)
predictor-corrector method to enhance the accuracy of ODE solving in rectified
flow models. Specifically, we introduce ABM-Solver, which integrates a multi
step predictor corrector approach to reduce local truncation errors and employs
Adaptive Step Size Adjustment to improve sampling speed. Furthermore, to
effectively preserve non edited regions while facilitating semantic
modifications, we introduce a Mask Guided Feature Injection module. We estimate
self-similarity to generate a spatial mask that differentiates preserved
regions from those available for editing. Extensive experiments on multiple
high-resolution image datasets validate that ABM-Solver significantly improves
inversion precision and editing quality, outperforming existing solvers without
requiring additional training or optimization.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 02:17:33 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ma",
"Yongjia",
""
],
[
"Di",
"Donglin",
""
],
[
"Liu",
"Xuan",
""
],
[
"Chen",
"Xiaokai",
""
],
[
"Fan",
"Lei",
""
],
[
"Chen",
"Wei",
""
],
[
"Su",
"Tonghua",
""
]
] | TITLE: Adams Bashforth Moulton Solver for Inversion and Editing in Rectified
Flow
ABSTRACT: Rectified flow models have achieved remarkable performance in image and video
generation tasks. However, existing numerical solvers face a trade-off between
fast sampling and high-accuracy solutions, limiting their effectiveness in
downstream applications such as reconstruction and editing. To address this
challenge, we propose leveraging the Adams-Bashforth-Moulton (ABM)
predictor-corrector method to enhance the accuracy of ODE solving in rectified
flow models. Specifically, we introduce ABM-Solver, which integrates a multi
step predictor corrector approach to reduce local truncation errors and employs
Adaptive Step Size Adjustment to improve sampling speed. Furthermore, to
effectively preserve non edited regions while facilitating semantic
modifications, we introduce a Mask Guided Feature Injection module. We estimate
self-similarity to generate a spatial mask that differentiates preserved
regions from those available for editing. Extensive experiments on multiple
high-resolution image datasets validate that ABM-Solver significantly improves
inversion precision and editing quality, outperforming existing solvers without
requiring additional training or optimization.
|
2503.16525 | Huan Yang | Huan Yang, Renji Zhang and Deyu Zhang | KVShare: Semantic-Aware Key-Value Cache Sharing for Efficient Large
Language Model Inference | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents KVShare, a multi-user Key-Value (KV) Cache sharing
technology based on semantic similarity, designed to enhance the inference
efficiency of Large Language Models (LLMs) and Multimodal Large Language Models
(MLLMs). Addressing the limitations of existing prefix caching (strict text
prefix matching) and semantic caching (loss of response diversity), KVShare
achieves fine-grained KV cache reuse through semantic alignment algorithms and
differential editing operations. Experiments on real-world user conversation
datasets demonstrate that KVShare improves KV cache hit rates by over 60%,
while maintaining output quality comparable to full computation (no significant
degradation in BLEU and Rouge-L metrics). This approach effectively reduces GPU
resource consumption and is applicable to scenarios with repetitive queries,
such as healthcare and education.
| [
{
"version": "v1",
"created": "Mon, 17 Mar 2025 16:43:35 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Yang",
"Huan",
""
],
[
"Zhang",
"Renji",
""
],
[
"Zhang",
"Deyu",
""
]
] | TITLE: KVShare: Semantic-Aware Key-Value Cache Sharing for Efficient Large
Language Model Inference
ABSTRACT: This paper presents KVShare, a multi-user Key-Value (KV) Cache sharing
technology based on semantic similarity, designed to enhance the inference
efficiency of Large Language Models (LLMs) and Multimodal Large Language Models
(MLLMs). Addressing the limitations of existing prefix caching (strict text
prefix matching) and semantic caching (loss of response diversity), KVShare
achieves fine-grained KV cache reuse through semantic alignment algorithms and
differential editing operations. Experiments on real-world user conversation
datasets demonstrate that KVShare improves KV cache hit rates by over 60%,
while maintaining output quality comparable to full computation (no significant
degradation in BLEU and Rouge-L metrics). This approach effectively reduces GPU
resource consumption and is applicable to scenarios with repetitive queries,
such as healthcare and education.
|
2503.16527 | Ang Li | Ang Li, Haozhe Chen, Hongseok Namkoong, Tianyi Peng | LLM Generated Persona is a Promise with a Catch | null | null | null | null | cs.CL cs.AI cs.CY cs.SI | http://creativecommons.org/licenses/by/4.0/ | The use of large language models (LLMs) to simulate human behavior has gained
significant attention, particularly through personas that approximate
individual characteristics. Persona-based simulations hold promise for
transforming disciplines that rely on population-level feedback, including
social science, economic analysis, marketing research, and business operations.
Traditional methods to collect realistic persona data face significant
challenges. They are prohibitively expensive and logistically challenging due
to privacy constraints, and often fail to capture multi-dimensional attributes,
particularly subjective qualities. Consequently, synthetic persona generation
with LLMs offers a scalable, cost-effective alternative. However, current
approaches rely on ad hoc and heuristic generation techniques that do not
guarantee methodological rigor or simulation precision, resulting in systematic
biases in downstream tasks. Through extensive large-scale experiments including
presidential election forecasts and general opinion surveys of the U.S.
population, we reveal that these biases can lead to significant deviations from
real-world outcomes. Our findings underscore the need to develop a rigorous
science of persona generation and outline the methodological innovations,
organizational and institutional support, and empirical foundations required to
enhance the reliability and scalability of LLM-driven persona simulations. To
support further research and development in this area, we have open-sourced
approximately one million generated personas, available for public access and
analysis at https://huggingface.co/datasets/Tianyi-Lab/Personas.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 03:11:27 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Ang",
""
],
[
"Chen",
"Haozhe",
""
],
[
"Namkoong",
"Hongseok",
""
],
[
"Peng",
"Tianyi",
""
]
] | TITLE: LLM Generated Persona is a Promise with a Catch
ABSTRACT: The use of large language models (LLMs) to simulate human behavior has gained
significant attention, particularly through personas that approximate
individual characteristics. Persona-based simulations hold promise for
transforming disciplines that rely on population-level feedback, including
social science, economic analysis, marketing research, and business operations.
Traditional methods to collect realistic persona data face significant
challenges. They are prohibitively expensive and logistically challenging due
to privacy constraints, and often fail to capture multi-dimensional attributes,
particularly subjective qualities. Consequently, synthetic persona generation
with LLMs offers a scalable, cost-effective alternative. However, current
approaches rely on ad hoc and heuristic generation techniques that do not
guarantee methodological rigor or simulation precision, resulting in systematic
biases in downstream tasks. Through extensive large-scale experiments including
presidential election forecasts and general opinion surveys of the U.S.
population, we reveal that these biases can lead to significant deviations from
real-world outcomes. Our findings underscore the need to develop a rigorous
science of persona generation and outline the methodological innovations,
organizational and institutional support, and empirical foundations required to
enhance the reliability and scalability of LLM-driven persona simulations. To
support further research and development in this area, we have open-sourced
approximately one million generated personas, available for public access and
analysis at https://huggingface.co/datasets/Tianyi-Lab/Personas.
|
2503.16530 | Chengfeng Dou | Chengfeng Dou, Ying Zhang, Zhi Jin, Wenpin Jiao, Haiyan Zhao,
Yongqiang Zhao, Zhengwei Tao | Enhancing LLM Generation with Knowledge Hypergraph for Evidence-Based
Medicine | null | null | null | null | cs.CL cs.AI cs.IR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Evidence-based medicine (EBM) plays a crucial role in the application of
large language models (LLMs) in healthcare, as it provides reliable support for
medical decision-making processes. Although it benefits from current
retrieval-augmented generation~(RAG) technologies, it still faces two
significant challenges: the collection of dispersed evidence and the efficient
organization of this evidence to support the complex queries necessary for EBM.
To tackle these issues, we propose using LLMs to gather scattered evidence from
multiple sources and present a knowledge hypergraph-based evidence management
model to integrate these evidence while capturing intricate relationships.
Furthermore, to better support complex queries, we have developed an
Importance-Driven Evidence Prioritization (IDEP) algorithm that utilizes the
LLM to generate multiple evidence features, each with an associated importance
score, which are then used to rank the evidence and produce the final retrieval
results. Experimental results from six datasets demonstrate that our approach
outperforms existing RAG techniques in application domains of interest to EBM,
such as medical quizzing, hallucination detection, and decision support.
Testsets and the constructed knowledge graph can be accessed at
\href{https://drive.google.com/file/d/1WJ9QTokK3MdkjEmwuFQxwH96j_Byawj_/view?usp=drive_link}{https://drive.google.com/rag4ebm}.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 09:17:31 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Dou",
"Chengfeng",
""
],
[
"Zhang",
"Ying",
""
],
[
"Jin",
"Zhi",
""
],
[
"Jiao",
"Wenpin",
""
],
[
"Zhao",
"Haiyan",
""
],
[
"Zhao",
"Yongqiang",
""
],
[
"Tao",
"Zhengwei",
""
]
] | TITLE: Enhancing LLM Generation with Knowledge Hypergraph for Evidence-Based
Medicine
ABSTRACT: Evidence-based medicine (EBM) plays a crucial role in the application of
large language models (LLMs) in healthcare, as it provides reliable support for
medical decision-making processes. Although it benefits from current
retrieval-augmented generation~(RAG) technologies, it still faces two
significant challenges: the collection of dispersed evidence and the efficient
organization of this evidence to support the complex queries necessary for EBM.
To tackle these issues, we propose using LLMs to gather scattered evidence from
multiple sources and present a knowledge hypergraph-based evidence management
model to integrate these evidence while capturing intricate relationships.
Furthermore, to better support complex queries, we have developed an
Importance-Driven Evidence Prioritization (IDEP) algorithm that utilizes the
LLM to generate multiple evidence features, each with an associated importance
score, which are then used to rank the evidence and produce the final retrieval
results. Experimental results from six datasets demonstrate that our approach
outperforms existing RAG techniques in application domains of interest to EBM,
such as medical quizzing, hallucination detection, and decision support.
Testsets and the constructed knowledge graph can be accessed at
\href{https://drive.google.com/file/d/1WJ9QTokK3MdkjEmwuFQxwH96j_Byawj_/view?usp=drive_link}{https://drive.google.com/rag4ebm}.
|
2503.16532 | Meisam Jamshidi Seikavandi | Meisam Jamshidi Seikavandi, Jostein Fimland, Maria Barrett, Paolo
Burelli | Modelling Emotions in Face-to-Face Setting: The Interplay of
Eye-Tracking, Personality, and Temporal Dynamics | null | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by/4.0/ | Accurate emotion recognition is pivotal for nuanced and engaging
human-computer interactions, yet remains difficult to achieve, especially in
dynamic, conversation-like settings. In this study, we showcase how integrating
eye-tracking data, temporal dynamics, and personality traits can substantially
enhance the detection of both perceived and felt emotions. Seventy-three
participants viewed short, speech-containing videos from the CREMA-D dataset,
while being recorded for eye-tracking signals (pupil size, fixation patterns),
Big Five personality assessments, and self-reported emotional states. Our
neural network models combined these diverse inputs including stimulus emotion
labels for contextual cues and yielded marked performance gains compared to the
state-of-the-art. Specifically, perceived valence predictions reached a macro
F1-score of 0.76, and models incorporating personality traits and stimulus
information demonstrated significant improvements in felt emotion accuracy.
These results highlight the benefit of unifying physiological, individual and
contextual factors to address the subjectivity and complexity of emotional
expression. Beyond validating the role of user-specific data in capturing
subtle internal states, our findings inform the design of future affective
computing and human-agent systems, paving the way for more adaptive and
cross-individual emotional intelligence in real-world interactions.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 13:15:32 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Seikavandi",
"Meisam Jamshidi",
""
],
[
"Fimland",
"Jostein",
""
],
[
"Barrett",
"Maria",
""
],
[
"Burelli",
"Paolo",
""
]
] | TITLE: Modelling Emotions in Face-to-Face Setting: The Interplay of
Eye-Tracking, Personality, and Temporal Dynamics
ABSTRACT: Accurate emotion recognition is pivotal for nuanced and engaging
human-computer interactions, yet remains difficult to achieve, especially in
dynamic, conversation-like settings. In this study, we showcase how integrating
eye-tracking data, temporal dynamics, and personality traits can substantially
enhance the detection of both perceived and felt emotions. Seventy-three
participants viewed short, speech-containing videos from the CREMA-D dataset,
while being recorded for eye-tracking signals (pupil size, fixation patterns),
Big Five personality assessments, and self-reported emotional states. Our
neural network models combined these diverse inputs including stimulus emotion
labels for contextual cues and yielded marked performance gains compared to the
state-of-the-art. Specifically, perceived valence predictions reached a macro
F1-score of 0.76, and models incorporating personality traits and stimulus
information demonstrated significant improvements in felt emotion accuracy.
These results highlight the benefit of unifying physiological, individual and
contextual factors to address the subjectivity and complexity of emotional
expression. Beyond validating the role of user-specific data in capturing
subtle internal states, our findings inform the design of future affective
computing and human-agent systems, paving the way for more adaptive and
cross-individual emotional intelligence in real-world interactions.
|
2503.16538 | Bastian P\"atzold | Bastian P\"atzold, Jan Nogga, Sven Behnke | Leveraging Vision-Language Models for Open-Vocabulary Instance
Segmentation and Tracking | Submitted to IEEE Robotics and Automation Letters (RA-L) | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a novel approach that leverages the capabilities of
vision-language models (VLMs) by integrating them with established approaches
for open-vocabulary detection (OVD), instance segmentation, and tracking. We
utilize VLM-generated structured descriptions to identify visible object
instances, collect application-relevant attributes, and inform an
open-vocabulary detector to extract corresponding bounding boxes that are
passed to a video segmentation model providing precise segmentation masks and
tracking capabilities. Once initialized, this model can then directly extract
segmentation masks, allowing processing of image streams in real time with
minimal computational overhead. Tracks can be updated online as needed by
generating new structured descriptions and corresponding open-vocabulary
detections. This combines the descriptive power of VLMs with the grounding
capability of OVD and the pixel-level understanding and speed of video
segmentation. Our evaluation across datasets and robotics platforms
demonstrates the broad applicability of this approach, showcasing its ability
to extract task-specific attributes from non-standard objects in dynamic
environments.
| [
{
"version": "v1",
"created": "Tue, 18 Mar 2025 20:18:42 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Pätzold",
"Bastian",
""
],
[
"Nogga",
"Jan",
""
],
[
"Behnke",
"Sven",
""
]
] | TITLE: Leveraging Vision-Language Models for Open-Vocabulary Instance
Segmentation and Tracking
ABSTRACT: This paper introduces a novel approach that leverages the capabilities of
vision-language models (VLMs) by integrating them with established approaches
for open-vocabulary detection (OVD), instance segmentation, and tracking. We
utilize VLM-generated structured descriptions to identify visible object
instances, collect application-relevant attributes, and inform an
open-vocabulary detector to extract corresponding bounding boxes that are
passed to a video segmentation model providing precise segmentation masks and
tracking capabilities. Once initialized, this model can then directly extract
segmentation masks, allowing processing of image streams in real time with
minimal computational overhead. Tracks can be updated online as needed by
generating new structured descriptions and corresponding open-vocabulary
detections. This combines the descriptive power of VLMs with the grounding
capability of OVD and the pixel-level understanding and speed of video
segmentation. Our evaluation across datasets and robotics platforms
demonstrates the broad applicability of this approach, showcasing its ability
to extract task-specific attributes from non-standard objects in dynamic
environments.
|
2503.16542 | Shiyi Jiang | Shiyi Jiang, Farshad Firouzi, Krishnendu Chakrabarty | Defending Against Gradient Inversion Attacks for Biomedical Images via
Learnable Data Perturbation | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The increasing need for sharing healthcare data and collaborating on clinical
research has raised privacy concerns. Health information leakage due to
malicious attacks can lead to serious problems such as misdiagnoses and patient
identification issues. Privacy-preserving machine learning (PPML) and
privacy-enhancing technologies, particularly federated learning (FL), have
emerged in recent years as innovative solutions to balance privacy protection
with data utility; however, they also suffer from inherent privacy
vulnerabilities. Gradient inversion attacks constitute major threats to data
sharing in federated learning. Researchers have proposed many defenses against
gradient inversion attacks. However, current defense methods for healthcare
data lack generalizability, i.e., existing solutions may not be applicable to
data from a broader range of populations. In addition, most existing defense
methods are tested using non-healthcare data, which raises concerns about their
applicability to real-world healthcare systems. In this study, we present a
defense against gradient inversion attacks in federated learning. We achieve
this using latent data perturbation and minimax optimization, utilizing both
general and medical image datasets. Our method is compared to two baselines,
and the results show that our approach can outperform the baselines with a
reduction of 12.5% in the attacker's accuracy in classifying reconstructed
images. The proposed method also yields an increase of over 12.4% in Mean
Squared Error (MSE) between the original and reconstructed images at the same
level of model utility of around 90% client classification accuracy. The
results suggest the potential of a generalizable defense for healthcare data.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 01:53:23 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Jiang",
"Shiyi",
""
],
[
"Firouzi",
"Farshad",
""
],
[
"Chakrabarty",
"Krishnendu",
""
]
] | TITLE: Defending Against Gradient Inversion Attacks for Biomedical Images via
Learnable Data Perturbation
ABSTRACT: The increasing need for sharing healthcare data and collaborating on clinical
research has raised privacy concerns. Health information leakage due to
malicious attacks can lead to serious problems such as misdiagnoses and patient
identification issues. Privacy-preserving machine learning (PPML) and
privacy-enhancing technologies, particularly federated learning (FL), have
emerged in recent years as innovative solutions to balance privacy protection
with data utility; however, they also suffer from inherent privacy
vulnerabilities. Gradient inversion attacks constitute major threats to data
sharing in federated learning. Researchers have proposed many defenses against
gradient inversion attacks. However, current defense methods for healthcare
data lack generalizability, i.e., existing solutions may not be applicable to
data from a broader range of populations. In addition, most existing defense
methods are tested using non-healthcare data, which raises concerns about their
applicability to real-world healthcare systems. In this study, we present a
defense against gradient inversion attacks in federated learning. We achieve
this using latent data perturbation and minimax optimization, utilizing both
general and medical image datasets. Our method is compared to two baselines,
and the results show that our approach can outperform the baselines with a
reduction of 12.5% in the attacker's accuracy in classifying reconstructed
images. The proposed method also yields an increase of over 12.4% in Mean
Squared Error (MSE) between the original and reconstructed images at the same
level of model utility of around 90% client classification accuracy. The
results suggest the potential of a generalizable defense for healthcare data.
|
2503.16544 | Donghuo Zeng | Donghuo Zeng, Roberto Legaspi, Yuewen Sun, Xinshuai Dong, Kazushi
Ikeda, Peter Spirtes, Kun Zhang | Causal Discovery and Counterfactual Reasoning to Optimize Persuasive
Dialogue Policies | 21 pages, 8 figures | null | null | null | cs.CL cs.AI cs.HC | http://creativecommons.org/licenses/by/4.0/ | Tailoring persuasive conversations to users leads to more effective
persuasion. However, existing dialogue systems often struggle to adapt to
dynamically evolving user states. This paper presents a novel method that
leverages causal discovery and counterfactual reasoning for optimizing system
persuasion capability and outcomes. We employ the Greedy Relaxation of the
Sparsest Permutation (GRaSP) algorithm to identify causal relationships between
user and system utterance strategies, treating user strategies as states and
system strategies as actions. GRaSP identifies user strategies as causal
factors influencing system responses, which inform Bidirectional Conditional
Generative Adversarial Networks (BiCoGAN) in generating counterfactual
utterances for the system. Subsequently, we use the Dueling Double Deep
Q-Network (D3QN) model to utilize counterfactual data to determine the best
policy for selecting system utterances. Our experiments with the
PersuasionForGood dataset show measurable improvements in persuasion outcomes
using our approach over baseline methods. The observed increase in cumulative
rewards and Q-values highlights the effectiveness of causal discovery in
enhancing counterfactual reasoning and optimizing reinforcement learning
policies for online dialogue systems.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 06:06:10 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zeng",
"Donghuo",
""
],
[
"Legaspi",
"Roberto",
""
],
[
"Sun",
"Yuewen",
""
],
[
"Dong",
"Xinshuai",
""
],
[
"Ikeda",
"Kazushi",
""
],
[
"Spirtes",
"Peter",
""
],
[
"Zhang",
"Kun",
""
]
] | TITLE: Causal Discovery and Counterfactual Reasoning to Optimize Persuasive
Dialogue Policies
ABSTRACT: Tailoring persuasive conversations to users leads to more effective
persuasion. However, existing dialogue systems often struggle to adapt to
dynamically evolving user states. This paper presents a novel method that
leverages causal discovery and counterfactual reasoning for optimizing system
persuasion capability and outcomes. We employ the Greedy Relaxation of the
Sparsest Permutation (GRaSP) algorithm to identify causal relationships between
user and system utterance strategies, treating user strategies as states and
system strategies as actions. GRaSP identifies user strategies as causal
factors influencing system responses, which inform Bidirectional Conditional
Generative Adversarial Networks (BiCoGAN) in generating counterfactual
utterances for the system. Subsequently, we use the Dueling Double Deep
Q-Network (D3QN) model to utilize counterfactual data to determine the best
policy for selecting system utterances. Our experiments with the
PersuasionForGood dataset show measurable improvements in persuasion outcomes
using our approach over baseline methods. The observed increase in cumulative
rewards and Q-values highlights the effectiveness of causal discovery in
enhancing counterfactual reasoning and optimizing reinforcement learning
policies for online dialogue systems.
|
2503.16550 | Sun Yudao | Yudao Sun, Juan Yin, Juan Zhao, Fan Zhang, Yongheng Liu, Hongji Chen | Unified Enhancement of the Generalization and Robustness of Language
Models via Bi-Stage Optimization | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Neural network language models (LMs) are confronted with significant
challenges in generalization and robustness. Currently, many studies focus on
improving either generalization or robustness in isolation, without methods
addressing both aspects simultaneously, which presents a significant challenge
in developing LMs that are both robust and generalized. In this paper, we
propose a bi-stage optimization framework to uniformly enhance both the
generalization and robustness of LMs, termed UEGR. Specifically, during the
forward propagation stage, we enrich the output probability distributions of
adversarial samples by adaptive dropout to generate diverse sub models, and
incorporate JS divergence and adversarial losses of these output distributions
to reinforce output stability. During backward propagation stage, we compute
parameter saliency scores and selectively update only the most critical
parameters to minimize unnecessary deviations and consolidate the model's
resilience. Theoretical analysis shows that our framework includes gradient
regularization to limit the model's sensitivity to input perturbations and
selective parameter updates to flatten the loss landscape, thus improving both
generalization and robustness. The experimental results show that our method
significantly improves the generalization and robustness of LMs compared to
other existing methods across 13 publicly available language datasets,
achieving state-of-the-art (SOTA) performance.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 13:50:36 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Sun",
"Yudao",
""
],
[
"Yin",
"Juan",
""
],
[
"Zhao",
"Juan",
""
],
[
"Zhang",
"Fan",
""
],
[
"Liu",
"Yongheng",
""
],
[
"Chen",
"Hongji",
""
]
] | TITLE: Unified Enhancement of the Generalization and Robustness of Language
Models via Bi-Stage Optimization
ABSTRACT: Neural network language models (LMs) are confronted with significant
challenges in generalization and robustness. Currently, many studies focus on
improving either generalization or robustness in isolation, without methods
addressing both aspects simultaneously, which presents a significant challenge
in developing LMs that are both robust and generalized. In this paper, we
propose a bi-stage optimization framework to uniformly enhance both the
generalization and robustness of LMs, termed UEGR. Specifically, during the
forward propagation stage, we enrich the output probability distributions of
adversarial samples by adaptive dropout to generate diverse sub models, and
incorporate JS divergence and adversarial losses of these output distributions
to reinforce output stability. During backward propagation stage, we compute
parameter saliency scores and selectively update only the most critical
parameters to minimize unnecessary deviations and consolidate the model's
resilience. Theoretical analysis shows that our framework includes gradient
regularization to limit the model's sensitivity to input perturbations and
selective parameter updates to flatten the loss landscape, thus improving both
generalization and robustness. The experimental results show that our method
significantly improves the generalization and robustness of LMs compared to
other existing methods across 13 publicly available language datasets,
achieving state-of-the-art (SOTA) performance.
|
2503.16553 | Zhenlin Qin | Zhenlin Qin, Leizhen Wang, Francisco Camara Pereira, Zhenlinag Ma | A Foundational individual Mobility Prediction Model based on Open-Source
Large Language Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) are widely applied to domain-specific tasks due
to their massive general knowledge and remarkable inference capacities. Current
studies on LLMs have shown immense potential in applying LLMs to model
individual mobility prediction problems. However, most LLM-based mobility
prediction models only train on specific datasets or use single well-designed
prompts, leading to difficulty in adapting to different cities and users with
diverse contexts. To fill these gaps, this paper proposes a unified fine-tuning
framework to train a foundational open source LLM-based mobility prediction
model. We conducted extensive experiments on six real-world mobility datasets
to validate the proposed model. The results showed that the proposed model
achieved the best performance in prediction accuracy and transferability over
state-of-the-art models based on deep learning and LLMs.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 15:08:37 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Qin",
"Zhenlin",
""
],
[
"Wang",
"Leizhen",
""
],
[
"Pereira",
"Francisco Camara",
""
],
[
"Ma",
"Zhenlinag",
""
]
] | TITLE: A Foundational individual Mobility Prediction Model based on Open-Source
Large Language Models
ABSTRACT: Large Language Models (LLMs) are widely applied to domain-specific tasks due
to their massive general knowledge and remarkable inference capacities. Current
studies on LLMs have shown immense potential in applying LLMs to model
individual mobility prediction problems. However, most LLM-based mobility
prediction models only train on specific datasets or use single well-designed
prompts, leading to difficulty in adapting to different cities and users with
diverse contexts. To fill these gaps, this paper proposes a unified fine-tuning
framework to train a foundational open source LLM-based mobility prediction
model. We conducted extensive experiments on six real-world mobility datasets
to validate the proposed model. The results showed that the proposed model
achieved the best performance in prediction accuracy and transferability over
state-of-the-art models based on deep learning and LLMs.
|
2503.16556 | Sabeen Ahmed | Sabeen Ahmed, Nathan Parker, Margaret Park, Daniel Jeong, Lauren
Peres, Evan W. Davis, Jennifer B. Permuth, Erin Siegel, Matthew B. Schabath,
Yasin Yilmaz, and Ghulam Rasool | Reliable Radiologic Skeletal Muscle Area Assessment -- A Biomarker for
Cancer Cachexia Diagnosis | 47 pages, 19 figures, 9 Tables | null | null | null | eess.IV cs.AI cs.CE cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Cancer cachexia is a common metabolic disorder characterized by severe muscle
atrophy which is associated with poor prognosis and quality of life. Monitoring
skeletal muscle area (SMA) longitudinally through computed tomography (CT)
scans, an imaging modality routinely acquired in cancer care, is an effective
way to identify and track this condition. However, existing tools often lack
full automation and exhibit inconsistent accuracy, limiting their potential for
integration into clinical workflows. To address these challenges, we developed
SMAART-AI (Skeletal Muscle Assessment-Automated and Reliable Tool-based on AI),
an end-to-end automated pipeline powered by deep learning models (nnU-Net 2D)
trained on mid-third lumbar level CT images with 5-fold cross-validation,
ensuring generalizability and robustness. SMAART-AI incorporates an
uncertainty-based mechanism to flag high-error SMA predictions for expert
review, enhancing reliability. We combined the SMA, skeletal muscle index, BMI,
and clinical data to train a multi-layer perceptron (MLP) model designed to
predict cachexia at the time of cancer diagnosis. Tested on the
gastroesophageal cancer dataset, SMAART-AI achieved a Dice score of 97.80% +/-
0.93%, with SMA estimated across all four datasets in this study at a median
absolute error of 2.48% compared to manual annotations with SliceOmatic.
Uncertainty metrics-variance, entropy, and coefficient of variation-strongly
correlated with SMA prediction errors (0.83, 0.76, and 0.73 respectively). The
MLP model predicts cachexia with 79% precision, providing clinicians with a
reliable tool for early diagnosis and intervention. By combining automation,
accuracy, and uncertainty awareness, SMAART-AI bridges the gap between research
and clinical application, offering a transformative approach to managing cancer
cachexia.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 19:07:59 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ahmed",
"Sabeen",
""
],
[
"Parker",
"Nathan",
""
],
[
"Park",
"Margaret",
""
],
[
"Jeong",
"Daniel",
""
],
[
"Peres",
"Lauren",
""
],
[
"Davis",
"Evan W.",
""
],
[
"Permuth",
"Jennifer B.",
""
],
[
"Siegel",
"Erin",
""
],
[
"Schabath",
"Matthew B.",
""
],
[
"Yilmaz",
"Yasin",
""
],
[
"Rasool",
"Ghulam",
""
]
] | TITLE: Reliable Radiologic Skeletal Muscle Area Assessment -- A Biomarker for
Cancer Cachexia Diagnosis
ABSTRACT: Cancer cachexia is a common metabolic disorder characterized by severe muscle
atrophy which is associated with poor prognosis and quality of life. Monitoring
skeletal muscle area (SMA) longitudinally through computed tomography (CT)
scans, an imaging modality routinely acquired in cancer care, is an effective
way to identify and track this condition. However, existing tools often lack
full automation and exhibit inconsistent accuracy, limiting their potential for
integration into clinical workflows. To address these challenges, we developed
SMAART-AI (Skeletal Muscle Assessment-Automated and Reliable Tool-based on AI),
an end-to-end automated pipeline powered by deep learning models (nnU-Net 2D)
trained on mid-third lumbar level CT images with 5-fold cross-validation,
ensuring generalizability and robustness. SMAART-AI incorporates an
uncertainty-based mechanism to flag high-error SMA predictions for expert
review, enhancing reliability. We combined the SMA, skeletal muscle index, BMI,
and clinical data to train a multi-layer perceptron (MLP) model designed to
predict cachexia at the time of cancer diagnosis. Tested on the
gastroesophageal cancer dataset, SMAART-AI achieved a Dice score of 97.80% +/-
0.93%, with SMA estimated across all four datasets in this study at a median
absolute error of 2.48% compared to manual annotations with SliceOmatic.
Uncertainty metrics-variance, entropy, and coefficient of variation-strongly
correlated with SMA prediction errors (0.83, 0.76, and 0.73 respectively). The
MLP model predicts cachexia with 79% precision, providing clinicians with a
reliable tool for early diagnosis and intervention. By combining automation,
accuracy, and uncertainty awareness, SMAART-AI bridges the gap between research
and clinical application, offering a transformative approach to managing cancer
cachexia.
|
2503.16557 | Abdulaziz Al Mannai | Abdulaziz Al Mannai | Investigating Cultural Dimensions and Technological Acceptance: The
Adoption of Electronic Performance and Tracking Systems in Qatar's Football
Sector | null | null | null | null | cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | Qatar's football sector has undergone a substantial technological
transformation with the implementation of Electronic Performance and Tracking
Systems (EPTS). This study examines the impact of cultural and technological
factors on EPTS adoption, using Hofstede's Cultural Dimensions Theory and the
Technology Acceptance Model (TAM) as theoretical frameworks. An initial
exploratory study involved ten participants, followed by an expanded dataset
comprising thirty stakeholders, including players, coaches, and staff from
Qatari football organizations. Multiple regression analysis was conducted to
evaluate the relationships between perceived usefulness, perceived ease of use,
power distance, innovation receptiveness, integration complexity, and overall
adoption. The results indicate that perceived usefulness, innovation
receptiveness, and lower power distance significantly drive EPTS adoption,
while ease of use is marginally significant and integration complexity is
non-significant in this sample. These findings provide practical insights for
sports technology stakeholders in Qatar and emphasize the importance of
aligning cultural considerations with technological readiness for successful
EPTS integration.
| [
{
"version": "v1",
"created": "Wed, 19 Mar 2025 19:50:09 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Mannai",
"Abdulaziz Al",
""
]
] | TITLE: Investigating Cultural Dimensions and Technological Acceptance: The
Adoption of Electronic Performance and Tracking Systems in Qatar's Football
Sector
ABSTRACT: Qatar's football sector has undergone a substantial technological
transformation with the implementation of Electronic Performance and Tracking
Systems (EPTS). This study examines the impact of cultural and technological
factors on EPTS adoption, using Hofstede's Cultural Dimensions Theory and the
Technology Acceptance Model (TAM) as theoretical frameworks. An initial
exploratory study involved ten participants, followed by an expanded dataset
comprising thirty stakeholders, including players, coaches, and staff from
Qatari football organizations. Multiple regression analysis was conducted to
evaluate the relationships between perceived usefulness, perceived ease of use,
power distance, innovation receptiveness, integration complexity, and overall
adoption. The results indicate that perceived usefulness, innovation
receptiveness, and lower power distance significantly drive EPTS adoption,
while ease of use is marginally significant and integration complexity is
non-significant in this sample. These findings provide practical insights for
sports technology stakeholders in Qatar and emphasize the importance of
aligning cultural considerations with technological readiness for successful
EPTS integration.
|
2503.16561 | Ibrahim Al Azhar | Ibrahim Al Azher, Miftahul Jannat Mokarrama, Zhishuai Guo, Sagnik Ray
Choudhury, Hamed Alhoori | FutureGen: LLM-RAG Approach to Generate the Future Work of Scientific
Article | 19 pages, 5 figures | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The future work section of a scientific article outlines potential research
directions by identifying gaps and limitations of a current study. This section
serves as a valuable resource for early-career researchers seeking unexplored
areas and experienced researchers looking for new projects or collaborations.
In this study, we generate future work suggestions from key sections of a
scientific article alongside related papers and analyze how the trends have
evolved. We experimented with various Large Language Models (LLMs) and
integrated Retrieval-Augmented Generation (RAG) to enhance the generation
process. We incorporate a LLM feedback mechanism to improve the quality of the
generated content and propose an LLM-as-a-judge approach for evaluation. Our
results demonstrated that the RAG-based approach with LLM feedback outperforms
other methods evaluated through qualitative and quantitative metrics. Moreover,
we conduct a human evaluation to assess the LLM as an extractor and judge. The
code and dataset for this project are here, code: HuggingFace
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 06:14:02 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Azher",
"Ibrahim Al",
""
],
[
"Mokarrama",
"Miftahul Jannat",
""
],
[
"Guo",
"Zhishuai",
""
],
[
"Choudhury",
"Sagnik Ray",
""
],
[
"Alhoori",
"Hamed",
""
]
] | TITLE: FutureGen: LLM-RAG Approach to Generate the Future Work of Scientific
Article
ABSTRACT: The future work section of a scientific article outlines potential research
directions by identifying gaps and limitations of a current study. This section
serves as a valuable resource for early-career researchers seeking unexplored
areas and experienced researchers looking for new projects or collaborations.
In this study, we generate future work suggestions from key sections of a
scientific article alongside related papers and analyze how the trends have
evolved. We experimented with various Large Language Models (LLMs) and
integrated Retrieval-Augmented Generation (RAG) to enhance the generation
process. We incorporate a LLM feedback mechanism to improve the quality of the
generated content and propose an LLM-as-a-judge approach for evaluation. Our
results demonstrated that the RAG-based approach with LLM feedback outperforms
other methods evaluated through qualitative and quantitative metrics. Moreover,
we conduct a human evaluation to assess the LLM as an extractor and judge. The
code and dataset for this project are here, code: HuggingFace
|
2503.16567 | Paolo Burelli | Laurits Dixen, Stefan Heinrich and Paolo Burelli | Exploring Deep Learning Models for EEG Neural Decoding | null | null | 0.1007/978-3-031-82487-6_12 | null | cs.LG eess.SP q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Neural decoding is an important method in cognitive neuroscience that aims to
decode brain representations from recorded neural activity using a multivariate
machine learning model. The THINGS initiative provides a large EEG dataset of
46 subjects watching rapidly shown images. Here, we test the feasibility of
using this method for decoding high-level object features using recent deep
learning models. We create a derivative dataset from this of living vs
non-living entities test 15 different deep learning models with 5 different
architectures and compare to a SOTA linear model. We show that the linear model
is not able to solve the decoding task, while almost all the deep learning
models are successful, suggesting that in some cases non-linear models are
needed to decode neural representations. We also run a comparative study of the
models' performance on individual object categories, and suggest how artificial
neural networks can be used to study brain activity.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 08:02:09 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Dixen",
"Laurits",
""
],
[
"Heinrich",
"Stefan",
""
],
[
"Burelli",
"Paolo",
""
]
] | TITLE: Exploring Deep Learning Models for EEG Neural Decoding
ABSTRACT: Neural decoding is an important method in cognitive neuroscience that aims to
decode brain representations from recorded neural activity using a multivariate
machine learning model. The THINGS initiative provides a large EEG dataset of
46 subjects watching rapidly shown images. Here, we test the feasibility of
using this method for decoding high-level object features using recent deep
learning models. We create a derivative dataset from this of living vs
non-living entities test 15 different deep learning models with 5 different
architectures and compare to a SOTA linear model. We show that the linear model
is not able to solve the decoding task, while almost all the deep learning
models are successful, suggesting that in some cases non-linear models are
needed to decode neural representations. We also run a comparative study of the
models' performance on individual object categories, and suggest how artificial
neural networks can be used to study brain activity.
|
2503.16572 | Shu Yang | Shu Yang, Chengting Yu, Lei Liu, Hanzhi Ma, Aili Wang, Erping Li | Efficient ANN-Guided Distillation: Aligning Rate-based Features of
Spiking Neural Networks through Hybrid Block-wise Replacement | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spiking Neural Networks (SNNs) have garnered considerable attention as a
potential alternative to Artificial Neural Networks (ANNs). Recent studies have
highlighted SNNs' potential on large-scale datasets. For SNN training, two main
approaches exist: direct training and ANN-to-SNN (ANN2SNN) conversion. To fully
leverage existing ANN models in guiding SNN learning, either direct ANN-to-SNN
conversion or ANN-SNN distillation training can be employed. In this paper, we
propose an ANN-SNN distillation framework from the ANN-to-SNN perspective,
designed with a block-wise replacement strategy for ANN-guided learning. By
generating intermediate hybrid models that progressively align SNN feature
spaces to those of ANN through rate-based features, our framework naturally
incorporates rate-based backpropagation as a training method. Our approach
achieves results comparable to or better than state-of-the-art SNN distillation
methods, showing both training and learning efficiency.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 09:04:38 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Yang",
"Shu",
""
],
[
"Yu",
"Chengting",
""
],
[
"Liu",
"Lei",
""
],
[
"Ma",
"Hanzhi",
""
],
[
"Wang",
"Aili",
""
],
[
"Li",
"Erping",
""
]
] | TITLE: Efficient ANN-Guided Distillation: Aligning Rate-based Features of
Spiking Neural Networks through Hybrid Block-wise Replacement
ABSTRACT: Spiking Neural Networks (SNNs) have garnered considerable attention as a
potential alternative to Artificial Neural Networks (ANNs). Recent studies have
highlighted SNNs' potential on large-scale datasets. For SNN training, two main
approaches exist: direct training and ANN-to-SNN (ANN2SNN) conversion. To fully
leverage existing ANN models in guiding SNN learning, either direct ANN-to-SNN
conversion or ANN-SNN distillation training can be employed. In this paper, we
propose an ANN-SNN distillation framework from the ANN-to-SNN perspective,
designed with a block-wise replacement strategy for ANN-guided learning. By
generating intermediate hybrid models that progressively align SNN feature
spaces to those of ANN through rate-based features, our framework naturally
incorporates rate-based backpropagation as a training method. Our approach
achieves results comparable to or better than state-of-the-art SNN distillation
methods, showing both training and learning efficiency.
|
2503.16575 | Han Yuan | Bo Hu, Han Yuan, Vlad Pandelea, Wuqiong Luo, Yingzhu Zhao, Zheng Ma | Extract, Match, and Score: An Evaluation Paradigm for Long
Question-context-answer Triplets in Financial Analysis | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The rapid advancement of large language models (LLMs) has sparked widespread
adoption across diverse applications, making robust evaluation frameworks
crucial for assessing their performance. While conventional evaluation metrics
remain applicable for shorter texts, their efficacy diminishes when evaluating
the quality of long-form answers. This limitation is particularly critical in
real-world scenarios involving extended questions, extensive context, and
long-form answers, such as financial analysis or regulatory compliance. In this
paper, we use a practical financial use case to illustrate applications that
handle "long question-context-answer triplets". We construct a real-world
financial dataset comprising long triplets and demonstrate the inadequacies of
traditional metrics. To address this, we propose an effective Extract, Match,
and Score (EMS) evaluation approach tailored to the complexities of long-form
LLMs' outputs, providing practitioners with a reliable methodology for
assessing LLMs' performance in complex real-world scenarios.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 09:38:44 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Hu",
"Bo",
""
],
[
"Yuan",
"Han",
""
],
[
"Pandelea",
"Vlad",
""
],
[
"Luo",
"Wuqiong",
""
],
[
"Zhao",
"Yingzhu",
""
],
[
"Ma",
"Zheng",
""
]
] | TITLE: Extract, Match, and Score: An Evaluation Paradigm for Long
Question-context-answer Triplets in Financial Analysis
ABSTRACT: The rapid advancement of large language models (LLMs) has sparked widespread
adoption across diverse applications, making robust evaluation frameworks
crucial for assessing their performance. While conventional evaluation metrics
remain applicable for shorter texts, their efficacy diminishes when evaluating
the quality of long-form answers. This limitation is particularly critical in
real-world scenarios involving extended questions, extensive context, and
long-form answers, such as financial analysis or regulatory compliance. In this
paper, we use a practical financial use case to illustrate applications that
handle "long question-context-answer triplets". We construct a real-world
financial dataset comprising long triplets and demonstrate the inadequacies of
traditional metrics. To address this, we propose an effective Extract, Match,
and Score (EMS) evaluation approach tailored to the complexities of long-form
LLMs' outputs, providing practitioners with a reliable methodology for
assessing LLMs' performance in complex real-world scenarios.
|
2503.16577 | Bilal Ahmad | Bilal Ahmad, Jinfu Chen, Haibao Chen | Feature selection strategies for optimized heart disease diagnosis using
ML and DL models | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Heart disease remains one of the leading causes of morbidity and mortality
worldwide, necessitating the development of effective diagnostic tools to
enable early diagnosis and clinical decision-making. This study evaluates the
impact of feature selection techniques Mutual Information (MI), Analysis of
Variance (ANOVA), and Chi-Square on the predictive performance of various
machine learning (ML) and deep learning (DL) models using a dataset of clinical
indicators for heart disease. Eleven ML/DL models were assessed using metrics
such as precision, recall, AUC score, F1-score, and accuracy. Results indicate
that MI outperformed other methods, particularly for advanced models like
neural networks, achieving the highest accuracy of 82.3% and recall score of
0.94. Logistic regression (accuracy 82.1%) and random forest (accuracy 80.99%)
also demonstrated improved performance with MI. Simpler models such as Naive
Bayes and decision trees achieved comparable results with ANOVA and Chi-Square,
yielding accuracies of 76.45% and 75.99%, respectively, making them
computationally efficient alternatives. Conversely, k Nearest Neighbors (KNN)
and Support Vector Machines (SVM) exhibited lower performance, with accuracies
ranging between 51.52% and 54.43%, regardless of the feature selection method.
This study provides a comprehensive comparison of feature selection methods for
heart disease prediction, demonstrating the critical role of feature selection
in optimizing model performance. The results offer practical guidance for
selecting appropriate feature selection techniques based on the chosen
classification algorithm, contributing to the development of more accurate and
efficient diagnostic tools for enhanced clinical decision-making in cardiology.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 09:59:01 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ahmad",
"Bilal",
""
],
[
"Chen",
"Jinfu",
""
],
[
"Chen",
"Haibao",
""
]
] | TITLE: Feature selection strategies for optimized heart disease diagnosis using
ML and DL models
ABSTRACT: Heart disease remains one of the leading causes of morbidity and mortality
worldwide, necessitating the development of effective diagnostic tools to
enable early diagnosis and clinical decision-making. This study evaluates the
impact of feature selection techniques Mutual Information (MI), Analysis of
Variance (ANOVA), and Chi-Square on the predictive performance of various
machine learning (ML) and deep learning (DL) models using a dataset of clinical
indicators for heart disease. Eleven ML/DL models were assessed using metrics
such as precision, recall, AUC score, F1-score, and accuracy. Results indicate
that MI outperformed other methods, particularly for advanced models like
neural networks, achieving the highest accuracy of 82.3% and recall score of
0.94. Logistic regression (accuracy 82.1%) and random forest (accuracy 80.99%)
also demonstrated improved performance with MI. Simpler models such as Naive
Bayes and decision trees achieved comparable results with ANOVA and Chi-Square,
yielding accuracies of 76.45% and 75.99%, respectively, making them
computationally efficient alternatives. Conversely, k Nearest Neighbors (KNN)
and Support Vector Machines (SVM) exhibited lower performance, with accuracies
ranging between 51.52% and 54.43%, regardless of the feature selection method.
This study provides a comprehensive comparison of feature selection methods for
heart disease prediction, demonstrating the critical role of feature selection
in optimizing model performance. The results offer practical guidance for
selecting appropriate feature selection techniques based on the chosen
classification algorithm, contributing to the development of more accurate and
efficient diagnostic tools for enhanced clinical decision-making in cardiology.
|
2503.16578 | Yang Chen | Yang Chen and Hui Wang and Shiyao Wang and Junyang Chen and Jiabei He
and Jiaming Zhou and Xi Yang and Yequan Wang and Yonghua Lin and Yong Qin | SeniorTalk: A Chinese Conversation Dataset with Rich Annotations for
Super-Aged Seniors | null | null | null | null | cs.CL cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | While voice technologies increasingly serve aging populations, current
systems exhibit significant performance gaps due to inadequate training data
capturing elderly-specific vocal characteristics like presbyphonia and
dialectal variations. The limited data available on super-aged individuals in
existing elderly speech datasets, coupled with overly simple recording styles
and annotation dimensions, exacerbates this issue. To address the critical
scarcity of speech data from individuals aged 75 and above, we introduce
SeniorTalk, a carefully annotated Chinese spoken dialogue dataset. This dataset
contains 55.53 hours of speech from 101 natural conversations involving 202
participants, ensuring a strategic balance across gender, region, and age.
Through detailed annotation across multiple dimensions, it can support a wide
range of speech tasks. We perform extensive experiments on speaker
verification, speaker diarization, speech recognition, and speech editing
tasks, offering crucial insights for the development of speech technologies
targeting this age group.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 11:31:47 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Chen",
"Yang",
""
],
[
"Wang",
"Hui",
""
],
[
"Wang",
"Shiyao",
""
],
[
"Chen",
"Junyang",
""
],
[
"He",
"Jiabei",
""
],
[
"Zhou",
"Jiaming",
""
],
[
"Yang",
"Xi",
""
],
[
"Wang",
"Yequan",
""
],
[
"Lin",
"Yonghua",
""
],
[
"Qin",
"Yong",
""
]
] | TITLE: SeniorTalk: A Chinese Conversation Dataset with Rich Annotations for
Super-Aged Seniors
ABSTRACT: While voice technologies increasingly serve aging populations, current
systems exhibit significant performance gaps due to inadequate training data
capturing elderly-specific vocal characteristics like presbyphonia and
dialectal variations. The limited data available on super-aged individuals in
existing elderly speech datasets, coupled with overly simple recording styles
and annotation dimensions, exacerbates this issue. To address the critical
scarcity of speech data from individuals aged 75 and above, we introduce
SeniorTalk, a carefully annotated Chinese spoken dialogue dataset. This dataset
contains 55.53 hours of speech from 101 natural conversations involving 202
participants, ensuring a strategic balance across gender, region, and age.
Through detailed annotation across multiple dimensions, it can support a wide
range of speech tasks. We perform extensive experiments on speaker
verification, speaker diarization, speech recognition, and speech editing
tasks, offering crucial insights for the development of speech technologies
targeting this age group.
|
2503.16581 | Arbi Haza Nasution | Zahra Khalila, Arbi Haza Nasution, Winda Monika, Aytug Onan, Yohei
Murakami, Yasir Bin Ismail Radi, Noor Mohammad Osmani | Investigating Retrieval-Augmented Generation in Quranic Studies: A Study
of 13 Open-Source Large Language Models | 11 pages, keywords: Large-language-models; retrieval-augmented
generation; question answering; Quranic studies; Islamic teachings | International Journal of Advanced Computer Science and
Applications(IJACSA), 16(2), 2025 | 10.14569/IJACSA.2025.01602134 | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurate and contextually faithful responses are critical when applying large
language models (LLMs) to sensitive and domain-specific tasks, such as
answering queries related to quranic studies. General-purpose LLMs often
struggle with hallucinations, where generated responses deviate from
authoritative sources, raising concerns about their reliability in religious
contexts. This challenge highlights the need for systems that can integrate
domain-specific knowledge while maintaining response accuracy, relevance, and
faithfulness. In this study, we investigate 13 open-source LLMs categorized
into large (e.g., Llama3:70b, Gemma2:27b, QwQ:32b), medium (e.g., Gemma2:9b,
Llama3:8b), and small (e.g., Llama3.2:3b, Phi3:3.8b). A Retrieval-Augmented
Generation (RAG) is used to make up for the problems that come with using
separate models. This research utilizes a descriptive dataset of Quranic surahs
including the meanings, historical context, and qualities of the 114 surahs,
allowing the model to gather relevant knowledge before responding. The models
are evaluated using three key metrics set by human evaluators: context
relevance, answer faithfulness, and answer relevance. The findings reveal that
large models consistently outperform smaller models in capturing query
semantics and producing accurate, contextually grounded responses. The
Llama3.2:3b model, even though it is considered small, does very well on
faithfulness (4.619) and relevance (4.857), showing the promise of smaller
architectures that have been well optimized. This article examines the
trade-offs between model size, computational efficiency, and response quality
while using LLMs in domain-specific applications.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 13:26:30 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Khalila",
"Zahra",
""
],
[
"Nasution",
"Arbi Haza",
""
],
[
"Monika",
"Winda",
""
],
[
"Onan",
"Aytug",
""
],
[
"Murakami",
"Yohei",
""
],
[
"Radi",
"Yasir Bin Ismail",
""
],
[
"Osmani",
"Noor Mohammad",
""
]
] | TITLE: Investigating Retrieval-Augmented Generation in Quranic Studies: A Study
of 13 Open-Source Large Language Models
ABSTRACT: Accurate and contextually faithful responses are critical when applying large
language models (LLMs) to sensitive and domain-specific tasks, such as
answering queries related to quranic studies. General-purpose LLMs often
struggle with hallucinations, where generated responses deviate from
authoritative sources, raising concerns about their reliability in religious
contexts. This challenge highlights the need for systems that can integrate
domain-specific knowledge while maintaining response accuracy, relevance, and
faithfulness. In this study, we investigate 13 open-source LLMs categorized
into large (e.g., Llama3:70b, Gemma2:27b, QwQ:32b), medium (e.g., Gemma2:9b,
Llama3:8b), and small (e.g., Llama3.2:3b, Phi3:3.8b). A Retrieval-Augmented
Generation (RAG) is used to make up for the problems that come with using
separate models. This research utilizes a descriptive dataset of Quranic surahs
including the meanings, historical context, and qualities of the 114 surahs,
allowing the model to gather relevant knowledge before responding. The models
are evaluated using three key metrics set by human evaluators: context
relevance, answer faithfulness, and answer relevance. The findings reveal that
large models consistently outperform smaller models in capturing query
semantics and producing accurate, contextually grounded responses. The
Llama3.2:3b model, even though it is considered small, does very well on
faithfulness (4.619) and relevance (4.857), showing the promise of smaller
architectures that have been well optimized. This article examines the
trade-offs between model size, computational efficiency, and response quality
while using LLMs in domain-specific applications.
|
2503.16582 | Ruiqi Yang | Ruiqi Yang, Jianxu Wang, Wei Yuan, Xun Wang, Mei Li | Machine Learning-Based Genomic Linguistic Analysis (Gene Sequence
Feature Learning): A Case Study on Predicting Heavy Metal Response Genes in
Rice | null | null | null | null | cs.LG cs.AI q-bio.GN | http://creativecommons.org/licenses/by/4.0/ | This study explores the application of machine learning-based genetic
linguistics for identifying heavy metal response genes in rice (Oryza sativa).
By integrating convolutional neural networks and random forest algorithms, we
developed a hybrid model capable of extracting and learning meaningful features
from gene sequences, such as k-mer frequencies and physicochemical properties.
The model was trained and tested on datasets of genes, achieving high
predictive performance (precision: 0.89, F1-score: 0.82). RNA-seq and qRT-PCR
experiments conducted on rice leaves which exposed to Hg0, revealed
differential expression of genes associated with heavy metal responses, which
validated the model's predictions. Co-expression network analysis identified
103 related genes, and a literature review indicated that these genes are
highly likely to be involved in heavy metal-related biological processes. By
integrating and comparing the analysis results with those of differentially
expressed genes (DEGs), the validity of the new machine learning method was
further demonstrated. This study highlights the efficacy of combining machine
learning with genetic linguistics for large-scale gene prediction. It
demonstrates a cost-effective and efficient approach for uncovering molecular
mechanisms underlying heavy metal responses, with potential applications in
developing stress-tolerant crop varieties.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 13:41:31 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Yang",
"Ruiqi",
""
],
[
"Wang",
"Jianxu",
""
],
[
"Yuan",
"Wei",
""
],
[
"Wang",
"Xun",
""
],
[
"Li",
"Mei",
""
]
] | TITLE: Machine Learning-Based Genomic Linguistic Analysis (Gene Sequence
Feature Learning): A Case Study on Predicting Heavy Metal Response Genes in
Rice
ABSTRACT: This study explores the application of machine learning-based genetic
linguistics for identifying heavy metal response genes in rice (Oryza sativa).
By integrating convolutional neural networks and random forest algorithms, we
developed a hybrid model capable of extracting and learning meaningful features
from gene sequences, such as k-mer frequencies and physicochemical properties.
The model was trained and tested on datasets of genes, achieving high
predictive performance (precision: 0.89, F1-score: 0.82). RNA-seq and qRT-PCR
experiments conducted on rice leaves which exposed to Hg0, revealed
differential expression of genes associated with heavy metal responses, which
validated the model's predictions. Co-expression network analysis identified
103 related genes, and a literature review indicated that these genes are
highly likely to be involved in heavy metal-related biological processes. By
integrating and comparing the analysis results with those of differentially
expressed genes (DEGs), the validity of the new machine learning method was
further demonstrated. This study highlights the efficacy of combining machine
learning with genetic linguistics for large-scale gene prediction. It
demonstrates a cost-effective and efficient approach for uncovering molecular
mechanisms underlying heavy metal responses, with potential applications in
developing stress-tolerant crop varieties.
|
2503.16584 | Xin Huang | Xin Huang, Shiyao Zhu, Ziyu Wang, Yaping He, Hao Jin, Zhengkui Liu | EVA-MED: An Enhanced Valence-Arousal Multimodal Emotion Dataset for
Emotion Recognition | 10 pages, 6figures, 1table | null | null | null | cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We introduce a novel multimodal emotion recognition dataset that enhances the
precision of Valence-Arousal Model while accounting for individual differences.
This dataset includes electroencephalography (EEG), electrocardiography (ECG),
and pulse interval (PI) from 64 participants. Data collection employed two
emotion induction paradigms: video stimuli that targeted different valence
levels (positive, neutral, and negative) and the Mannheim Multicomponent Stress
Test (MMST), which induced high arousal through cognitive, emotional, and
social stressors. To enrich the dataset, participants' personality traits,
anxiety, depression, and emotional states were assessed using validated
questionnaires. By capturing a broad spectrum of affective responses while
accounting for individual differences, this dataset provides a robust resource
for precise emotion modeling. The integration of multimodal physiological data
with psychological assessments lays a strong foundation for personalized
emotion recognition. We anticipate this resource will support the development
of more accurate, adaptive, and individualized emotion recognition systems
across diverse applications.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 14:41:50 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Huang",
"Xin",
""
],
[
"Zhu",
"Shiyao",
""
],
[
"Wang",
"Ziyu",
""
],
[
"He",
"Yaping",
""
],
[
"Jin",
"Hao",
""
],
[
"Liu",
"Zhengkui",
""
]
] | TITLE: EVA-MED: An Enhanced Valence-Arousal Multimodal Emotion Dataset for
Emotion Recognition
ABSTRACT: We introduce a novel multimodal emotion recognition dataset that enhances the
precision of Valence-Arousal Model while accounting for individual differences.
This dataset includes electroencephalography (EEG), electrocardiography (ECG),
and pulse interval (PI) from 64 participants. Data collection employed two
emotion induction paradigms: video stimuli that targeted different valence
levels (positive, neutral, and negative) and the Mannheim Multicomponent Stress
Test (MMST), which induced high arousal through cognitive, emotional, and
social stressors. To enrich the dataset, participants' personality traits,
anxiety, depression, and emotional states were assessed using validated
questionnaires. By capturing a broad spectrum of affective responses while
accounting for individual differences, this dataset provides a robust resource
for precise emotion modeling. The integration of multimodal physiological data
with psychological assessments lays a strong foundation for personalized
emotion recognition. We anticipate this resource will support the development
of more accurate, adaptive, and individualized emotion recognition systems
across diverse applications.
|
2503.16585 | M. Hadi Amini | Hadi Amini, Md Jueal Mia, Yasaman Saadati, Ahmed Imteaj, Seyedsina
Nabavirazavi, Urmish Thakker, Md Zarif Hossain, Awal Ahmed Fime, S.S. Iyengar | Distributed LLMs and Multimodal Large Language Models: A Survey on
Advances, Challenges, and Future Directions | null | null | null | null | cs.CL cs.CV cs.DC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Language models (LMs) are machine learning models designed to predict
linguistic patterns by estimating the probability of word sequences based on
large-scale datasets, such as text. LMs have a wide range of applications in
natural language processing (NLP) tasks, including autocomplete and machine
translation. Although larger datasets typically enhance LM performance,
scalability remains a challenge due to constraints in computational power and
resources. Distributed computing strategies offer essential solutions for
improving scalability and managing the growing computational demand. Further,
the use of sensitive datasets in training and deployment raises significant
privacy concerns. Recent research has focused on developing decentralized
techniques to enable distributed training and inference while utilizing diverse
computational resources and enabling edge AI. This paper presents a survey on
distributed solutions for various LMs, including large language models (LLMs),
vision language models (VLMs), multimodal LLMs (MLLMs), and small language
models (SLMs). While LLMs focus on processing and generating text, MLLMs are
designed to handle multiple modalities of data (e.g., text, images, and audio)
and to integrate them for broader applications. To this end, this paper reviews
key advancements across the MLLM pipeline, including distributed training,
inference, fine-tuning, and deployment, while also identifying the
contributions, limitations, and future areas of improvement. Further, it
categorizes the literature based on six primary focus areas of
decentralization. Our analysis describes gaps in current methodologies for
enabling distributed solutions for LMs and outline future research directions,
emphasizing the need for novel solutions to enhance the robustness and
applicability of distributed LMs.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 15:18:25 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Amini",
"Hadi",
""
],
[
"Mia",
"Md Jueal",
""
],
[
"Saadati",
"Yasaman",
""
],
[
"Imteaj",
"Ahmed",
""
],
[
"Nabavirazavi",
"Seyedsina",
""
],
[
"Thakker",
"Urmish",
""
],
[
"Hossain",
"Md Zarif",
""
],
[
"Fime",
"Awal Ahmed",
""
],
[
"Iyengar",
"S. S.",
""
]
] | TITLE: Distributed LLMs and Multimodal Large Language Models: A Survey on
Advances, Challenges, and Future Directions
ABSTRACT: Language models (LMs) are machine learning models designed to predict
linguistic patterns by estimating the probability of word sequences based on
large-scale datasets, such as text. LMs have a wide range of applications in
natural language processing (NLP) tasks, including autocomplete and machine
translation. Although larger datasets typically enhance LM performance,
scalability remains a challenge due to constraints in computational power and
resources. Distributed computing strategies offer essential solutions for
improving scalability and managing the growing computational demand. Further,
the use of sensitive datasets in training and deployment raises significant
privacy concerns. Recent research has focused on developing decentralized
techniques to enable distributed training and inference while utilizing diverse
computational resources and enabling edge AI. This paper presents a survey on
distributed solutions for various LMs, including large language models (LLMs),
vision language models (VLMs), multimodal LLMs (MLLMs), and small language
models (SLMs). While LLMs focus on processing and generating text, MLLMs are
designed to handle multiple modalities of data (e.g., text, images, and audio)
and to integrate them for broader applications. To this end, this paper reviews
key advancements across the MLLM pipeline, including distributed training,
inference, fine-tuning, and deployment, while also identifying the
contributions, limitations, and future areas of improvement. Further, it
categorizes the literature based on six primary focus areas of
decentralization. Our analysis describes gaps in current methodologies for
enabling distributed solutions for LMs and outline future research directions,
emphasizing the need for novel solutions to enhance the robustness and
applicability of distributed LMs.
|
2503.16591 | Luigi Piccinelli | Luigi Piccinelli, Christos Sakaridis, Mattia Segu, Yung-Hsu Yang,
Siyuan Li, Wim Abbeloos, Luc Van Gool | UniK3D: Universal Camera Monocular 3D Estimation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Monocular 3D estimation is crucial for visual perception. However, current
methods fall short by relying on oversimplified assumptions, such as pinhole
camera models or rectified images. These limitations severely restrict their
general applicability, causing poor performance in real-world scenarios with
fisheye or panoramic images and resulting in substantial context loss. To
address this, we present UniK3D, the first generalizable method for monocular
3D estimation able to model any camera. Our method introduces a spherical 3D
representation which allows for better disentanglement of camera and scene
geometry and enables accurate metric 3D reconstruction for unconstrained camera
models. Our camera component features a novel, model-independent representation
of the pencil of rays, achieved through a learned superposition of spherical
harmonics. We also introduce an angular loss, which, together with the camera
module design, prevents the contraction of the 3D outputs for wide-view
cameras. A comprehensive zero-shot evaluation on 13 diverse datasets
demonstrates the state-of-the-art performance of UniK3D across 3D, depth, and
camera metrics, with substantial gains in challenging large-field-of-view and
panoramic settings, while maintaining top accuracy in conventional pinhole
small-field-of-view domains. Code and models are available at
github.com/lpiccinelli-eth/unik3d .
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 17:49:23 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Piccinelli",
"Luigi",
""
],
[
"Sakaridis",
"Christos",
""
],
[
"Segu",
"Mattia",
""
],
[
"Yang",
"Yung-Hsu",
""
],
[
"Li",
"Siyuan",
""
],
[
"Abbeloos",
"Wim",
""
],
[
"Van Gool",
"Luc",
""
]
] | TITLE: UniK3D: Universal Camera Monocular 3D Estimation
ABSTRACT: Monocular 3D estimation is crucial for visual perception. However, current
methods fall short by relying on oversimplified assumptions, such as pinhole
camera models or rectified images. These limitations severely restrict their
general applicability, causing poor performance in real-world scenarios with
fisheye or panoramic images and resulting in substantial context loss. To
address this, we present UniK3D, the first generalizable method for monocular
3D estimation able to model any camera. Our method introduces a spherical 3D
representation which allows for better disentanglement of camera and scene
geometry and enables accurate metric 3D reconstruction for unconstrained camera
models. Our camera component features a novel, model-independent representation
of the pencil of rays, achieved through a learned superposition of spherical
harmonics. We also introduce an angular loss, which, together with the camera
module design, prevents the contraction of the 3D outputs for wide-view
cameras. A comprehensive zero-shot evaluation on 13 diverse datasets
demonstrates the state-of-the-art performance of UniK3D across 3D, depth, and
camera metrics, with substantial gains in challenging large-field-of-view and
panoramic settings, while maintaining top accuracy in conventional pinhole
small-field-of-view domains. Code and models are available at
github.com/lpiccinelli-eth/unik3d .
|
2503.16614 | Victor Aguiar Evangelista De Farias | Maria de Lourdes M. Silva, Andr\'e L. C. Mendon\c{c}a, Eduardo R. D.
Neto, Iago C. Chaves, Felipe T. Brito, Victor A. E. Farias, Javam C. Machado | Classification of User Reports for Detection of Faulty Computer
Components using NLP Models: A Case Study | 9 pages, 2 figures | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Computer manufacturers typically offer platforms for users to report faults.
However, there remains a significant gap in these platforms' ability to
effectively utilize textual reports, which impedes users from describing their
issues in their own words. In this context, Natural Language Processing (NLP)
offers a promising solution, by enabling the analysis of user-generated text.
This paper presents an innovative approach that employs NLP models to classify
user reports for detecting faulty computer components, such as CPU, memory,
motherboard, video card, and more. In this work, we build a dataset of 341 user
reports obtained from many sources. Additionally, through extensive
experimental evaluation, our approach achieved an accuracy of 79% with our
dataset.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:11:26 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Silva",
"Maria de Lourdes M.",
""
],
[
"Mendonça",
"André L. C.",
""
],
[
"Neto",
"Eduardo R. D.",
""
],
[
"Chaves",
"Iago C.",
""
],
[
"Brito",
"Felipe T.",
""
],
[
"Farias",
"Victor A. E.",
""
],
[
"Machado",
"Javam C.",
""
]
] | TITLE: Classification of User Reports for Detection of Faulty Computer
Components using NLP Models: A Case Study
ABSTRACT: Computer manufacturers typically offer platforms for users to report faults.
However, there remains a significant gap in these platforms' ability to
effectively utilize textual reports, which impedes users from describing their
issues in their own words. In this context, Natural Language Processing (NLP)
offers a promising solution, by enabling the analysis of user-generated text.
This paper presents an innovative approach that employs NLP models to classify
user reports for detecting faulty computer components, such as CPU, memory,
motherboard, video card, and more. In this work, we build a dataset of 341 user
reports obtained from many sources. Additionally, through extensive
experimental evaluation, our approach achieved an accuracy of 79% with our
dataset.
|
2503.16616 | Xiaoran Zhang | Xiaoran Zhang, Byung-Woo Hong, Hyoungseob Park, Daniel H. Pak,
Anne-Marie Rickmann, Lawrence H. Staib, James S. Duncan, Alex Wong | Progressive Test Time Energy Adaptation for Medical Image Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We propose a model-agnostic, progressive test-time energy adaptation approach
for medical image segmentation. Maintaining model performance across diverse
medical datasets is challenging, as distribution shifts arise from inconsistent
imaging protocols and patient variations. Unlike domain adaptation methods that
require multiple passes through target data - impractical in clinical settings
- our approach adapts pretrained models progressively as they process test
data. Our method leverages a shape energy model trained on source data, which
assigns an energy score at the patch level to segmentation maps: low energy
represents in-distribution (accurate) shapes, while high energy signals
out-of-distribution (erroneous) predictions. By minimizing this energy score at
test time, we refine the segmentation model to align with the target
distribution. To validate the effectiveness and adaptability, we evaluated our
framework on eight public MRI (bSSFP, T1- and T2-weighted) and X-ray datasets
spanning cardiac, spinal cord, and lung segmentation. We consistently
outperform baselines both quantitatively and qualitatively.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:15:50 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zhang",
"Xiaoran",
""
],
[
"Hong",
"Byung-Woo",
""
],
[
"Park",
"Hyoungseob",
""
],
[
"Pak",
"Daniel H.",
""
],
[
"Rickmann",
"Anne-Marie",
""
],
[
"Staib",
"Lawrence H.",
""
],
[
"Duncan",
"James S.",
""
],
[
"Wong",
"Alex",
""
]
] | TITLE: Progressive Test Time Energy Adaptation for Medical Image Segmentation
ABSTRACT: We propose a model-agnostic, progressive test-time energy adaptation approach
for medical image segmentation. Maintaining model performance across diverse
medical datasets is challenging, as distribution shifts arise from inconsistent
imaging protocols and patient variations. Unlike domain adaptation methods that
require multiple passes through target data - impractical in clinical settings
- our approach adapts pretrained models progressively as they process test
data. Our method leverages a shape energy model trained on source data, which
assigns an energy score at the patch level to segmentation maps: low energy
represents in-distribution (accurate) shapes, while high energy signals
out-of-distribution (erroneous) predictions. By minimizing this energy score at
test time, we refine the segmentation model to align with the target
distribution. To validate the effectiveness and adaptability, we evaluated our
framework on eight public MRI (bSSFP, T1- and T2-weighted) and X-ray datasets
spanning cardiac, spinal cord, and lung segmentation. We consistently
outperform baselines both quantitatively and qualitatively.
|
2503.16628 | Moshiur Rahman Tonmoy | Moshiur Rahman Tonmoy, Md. Mithun Hossain, Nilanjan Dey, M. F. Mridha | MobilePlantViT: A Mobile-friendly Hybrid ViT for Generalized Plant
Disease Image Classification | Submitted to a journal for peer-review under IEEE Transactions series | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Plant diseases significantly threaten global food security by reducing crop
yields and undermining agricultural sustainability. AI-driven automated
classification has emerged as a promising solution, with deep learning models
demonstrating impressive performance in plant disease identification. However,
deploying these models on mobile and edge devices remains challenging due to
high computational demands and resource constraints, highlighting the need for
lightweight, accurate solutions for accessible smart agriculture systems. To
address this, we propose MobilePlantViT, a novel hybrid Vision Transformer
(ViT) architecture designed for generalized plant disease classification, which
optimizes resource efficiency while maintaining high performance. Extensive
experiments across diverse plant disease datasets of varying scales show our
model's effectiveness and strong generalizability, achieving test accuracies
ranging from 80% to over 99%. Notably, with only 0.69 million parameters, our
architecture outperforms the smallest versions of MobileViTv1 and MobileViTv2,
despite their higher parameter counts. These results underscore the potential
of our approach for real-world, AI-powered automated plant disease
classification in sustainable and resource-efficient smart agriculture systems.
All codes will be available in the GitHub repository:
https://github.com/moshiurtonmoy/MobilePlantViT
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:34:02 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Tonmoy",
"Moshiur Rahman",
""
],
[
"Hossain",
"Md. Mithun",
""
],
[
"Dey",
"Nilanjan",
""
],
[
"Mridha",
"M. F.",
""
]
] | TITLE: MobilePlantViT: A Mobile-friendly Hybrid ViT for Generalized Plant
Disease Image Classification
ABSTRACT: Plant diseases significantly threaten global food security by reducing crop
yields and undermining agricultural sustainability. AI-driven automated
classification has emerged as a promising solution, with deep learning models
demonstrating impressive performance in plant disease identification. However,
deploying these models on mobile and edge devices remains challenging due to
high computational demands and resource constraints, highlighting the need for
lightweight, accurate solutions for accessible smart agriculture systems. To
address this, we propose MobilePlantViT, a novel hybrid Vision Transformer
(ViT) architecture designed for generalized plant disease classification, which
optimizes resource efficiency while maintaining high performance. Extensive
experiments across diverse plant disease datasets of varying scales show our
model's effectiveness and strong generalizability, achieving test accuracies
ranging from 80% to over 99%. Notably, with only 0.69 million parameters, our
architecture outperforms the smallest versions of MobileViTv1 and MobileViTv2,
despite their higher parameter counts. These results underscore the potential
of our approach for real-world, AI-powered automated plant disease
classification in sustainable and resource-efficient smart agriculture systems.
All codes will be available in the GitHub repository:
https://github.com/moshiurtonmoy/MobilePlantViT
|
2503.16630 | Dana Cohen-Bar | Dana Cohen-Bar, Daniel Cohen-Or, Gal Chechik, Yoni Kasten | TriTex: Learning Texture from a Single Mesh via Triplane Semantic
Features | Project page: https://danacohen95.github.io/TriTex/ | null | null | null | cs.GR cs.CV | http://creativecommons.org/licenses/by/4.0/ | As 3D content creation continues to grow, transferring semantic textures
between 3D meshes remains a significant challenge in computer graphics. While
recent methods leverage text-to-image diffusion models for texturing, they
often struggle to preserve the appearance of the source texture during texture
transfer. We present \ourmethod, a novel approach that learns a volumetric
texture field from a single textured mesh by mapping semantic features to
surface colors. Using an efficient triplane-based architecture, our method
enables semantic-aware texture transfer to a novel target mesh. Despite
training on just one example, it generalizes effectively to diverse shapes
within the same category. Extensive evaluation on our newly created benchmark
dataset shows that \ourmethod{} achieves superior texture transfer quality and
fast inference times compared to existing methods. Our approach advances
single-example texture transfer, providing a practical solution for maintaining
visual coherence across related 3D models in applications like game development
and simulation.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:35:03 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Cohen-Bar",
"Dana",
""
],
[
"Cohen-Or",
"Daniel",
""
],
[
"Chechik",
"Gal",
""
],
[
"Kasten",
"Yoni",
""
]
] | TITLE: TriTex: Learning Texture from a Single Mesh via Triplane Semantic
Features
ABSTRACT: As 3D content creation continues to grow, transferring semantic textures
between 3D meshes remains a significant challenge in computer graphics. While
recent methods leverage text-to-image diffusion models for texturing, they
often struggle to preserve the appearance of the source texture during texture
transfer. We present \ourmethod, a novel approach that learns a volumetric
texture field from a single textured mesh by mapping semantic features to
surface colors. Using an efficient triplane-based architecture, our method
enables semantic-aware texture transfer to a novel target mesh. Despite
training on just one example, it generalizes effectively to diverse shapes
within the same category. Extensive evaluation on our newly created benchmark
dataset shows that \ourmethod{} achieves superior texture transfer quality and
fast inference times compared to existing methods. Our approach advances
single-example texture transfer, providing a practical solution for maintaining
visual coherence across related 3D models in applications like game development
and simulation.
|
2503.16635 | Yinchi Zhou | Yinchi Zhou, Huidong Xie, Menghua Xia, Qiong Liu, Bo Zhou, Tianqi
Chen, Jun Hou, Liang Guo, Xinyuan Zheng, Hanzhong Wang, Biao Li, Axel
Rominger, Kuangyu Shi, Nicha C. Dvorneka, and Chi Liu | Fed-NDIF: A Noise-Embedded Federated Diffusion Model For Low-Count
Whole-Body PET Denoising | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Low-count positron emission tomography (LCPET) imaging can reduce patients'
exposure to radiation but often suffers from increased image noise and reduced
lesion detectability, necessitating effective denoising techniques. Diffusion
models have shown promise in LCPET denoising for recovering degraded image
quality. However, training such models requires large and diverse datasets,
which are challenging to obtain in the medical domain. To address data scarcity
and privacy concerns, we combine diffusion models with federated learning -- a
decentralized training approach where models are trained individually at
different sites, and their parameters are aggregated on a central server over
multiple iterations. The variation in scanner types and image noise levels
within and across institutions poses additional challenges for federated
learning in LCPET denoising. In this study, we propose a novel noise-embedded
federated learning diffusion model (Fed-NDIF) to address these challenges,
leveraging a multicenter dataset and varying count levels. Our approach
incorporates liver normalized standard deviation (NSTD) noise embedding into a
2.5D diffusion model and utilizes the Federated Averaging (FedAvg) algorithm to
aggregate locally trained models into a global model, which is subsequently
fine-tuned on local datasets to optimize performance and obtain personalized
models. Extensive validation on datasets from the University of Bern, Ruijin
Hospital in Shanghai, and Yale-New Haven Hospital demonstrates the superior
performance of our method in enhancing image quality and improving lesion
quantification. The Fed-NDIF model shows significant improvements in PSNR,
SSIM, and NMSE of the entire 3D volume, as well as enhanced lesion
detectability and quantification, compared to local diffusion models and
federated UNet-based models.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:37:46 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Zhou",
"Yinchi",
""
],
[
"Xie",
"Huidong",
""
],
[
"Xia",
"Menghua",
""
],
[
"Liu",
"Qiong",
""
],
[
"Zhou",
"Bo",
""
],
[
"Chen",
"Tianqi",
""
],
[
"Hou",
"Jun",
""
],
[
"Guo",
"Liang",
""
],
[
"Zheng",
"Xinyuan",
""
],
[
"Wang",
"Hanzhong",
""
],
[
"Li",
"Biao",
""
],
[
"Rominger",
"Axel",
""
],
[
"Shi",
"Kuangyu",
""
],
[
"Dvorneka",
"Nicha C.",
""
],
[
"Liu",
"Chi",
""
]
] | TITLE: Fed-NDIF: A Noise-Embedded Federated Diffusion Model For Low-Count
Whole-Body PET Denoising
ABSTRACT: Low-count positron emission tomography (LCPET) imaging can reduce patients'
exposure to radiation but often suffers from increased image noise and reduced
lesion detectability, necessitating effective denoising techniques. Diffusion
models have shown promise in LCPET denoising for recovering degraded image
quality. However, training such models requires large and diverse datasets,
which are challenging to obtain in the medical domain. To address data scarcity
and privacy concerns, we combine diffusion models with federated learning -- a
decentralized training approach where models are trained individually at
different sites, and their parameters are aggregated on a central server over
multiple iterations. The variation in scanner types and image noise levels
within and across institutions poses additional challenges for federated
learning in LCPET denoising. In this study, we propose a novel noise-embedded
federated learning diffusion model (Fed-NDIF) to address these challenges,
leveraging a multicenter dataset and varying count levels. Our approach
incorporates liver normalized standard deviation (NSTD) noise embedding into a
2.5D diffusion model and utilizes the Federated Averaging (FedAvg) algorithm to
aggregate locally trained models into a global model, which is subsequently
fine-tuned on local datasets to optimize performance and obtain personalized
models. Extensive validation on datasets from the University of Bern, Ruijin
Hospital in Shanghai, and Yale-New Haven Hospital demonstrates the superior
performance of our method in enhancing image quality and improving lesion
quantification. The Fed-NDIF model shows significant improvements in PSNR,
SSIM, and NMSE of the entire 3D volume, as well as enhanced lesion
detectability and quantification, compared to local diffusion models and
federated UNet-based models.
|
2503.16639 | Thomas Kreutz | Thomas Kreutz, Max M\"uhlh\"auser, Alejandro Sanchez Guinea | Whenever, Wherever: Towards Orchestrating Crowd Simulations with
Spatio-Temporal Spawn Dynamics | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Realistic crowd simulations are essential for immersive virtual environments,
relying on both individual behaviors (microscopic dynamics) and overall crowd
patterns (macroscopic characteristics). While recent data-driven methods like
deep reinforcement learning improve microscopic realism, they often overlook
critical macroscopic features such as crowd density and flow, which are
governed by spatio-temporal spawn dynamics, namely, when and where agents enter
a scene. Traditional methods, like random spawn rates, stochastic processes, or
fixed schedules, are not guaranteed to capture the underlying complexity or
lack diversity and realism. To address this issue, we propose a novel approach
called nTPP-GMM that models spatio-temporal spawn dynamics using Neural
Temporal Point Processes (nTPPs) that are coupled with a spawn-conditional
Gaussian Mixture Model (GMM) for agent spawn and goal positions. We evaluate
our approach by orchestrating crowd simulations of three diverse real-world
datasets with nTPP-GMM. Our experiments demonstrate the orchestration with
nTPP-GMM leads to realistic simulations that reflect real-world crowd scenarios
and allow crowd analysis.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 18:46:41 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kreutz",
"Thomas",
""
],
[
"Mühlhäuser",
"Max",
""
],
[
"Guinea",
"Alejandro Sanchez",
""
]
] | TITLE: Whenever, Wherever: Towards Orchestrating Crowd Simulations with
Spatio-Temporal Spawn Dynamics
ABSTRACT: Realistic crowd simulations are essential for immersive virtual environments,
relying on both individual behaviors (microscopic dynamics) and overall crowd
patterns (macroscopic characteristics). While recent data-driven methods like
deep reinforcement learning improve microscopic realism, they often overlook
critical macroscopic features such as crowd density and flow, which are
governed by spatio-temporal spawn dynamics, namely, when and where agents enter
a scene. Traditional methods, like random spawn rates, stochastic processes, or
fixed schedules, are not guaranteed to capture the underlying complexity or
lack diversity and realism. To address this issue, we propose a novel approach
called nTPP-GMM that models spatio-temporal spawn dynamics using Neural
Temporal Point Processes (nTPPs) that are coupled with a spawn-conditional
Gaussian Mixture Model (GMM) for agent spawn and goal positions. We evaluate
our approach by orchestrating crowd simulations of three diverse real-world
datasets with nTPP-GMM. Our experiments demonstrate the orchestration with
nTPP-GMM leads to realistic simulations that reflect real-world crowd scenarios
and allow crowd analysis.
|
2503.16661 | Nikos Kanakaris | Alejandro Ariza-Casabona, Nikos Kanakaris, Daniele Malitesta | ContextGNN goes to Elliot: Towards Benchmarking Relational Deep Learning
for Static Link Prediction (aka Personalized Item Recommendation) | null | null | null | null | cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Relational deep learning (RDL) settles among the most exciting advances in
machine learning for relational databases, leveraging the representational
power of message passing graph neural networks (GNNs) to derive useful
knowledge and run predicting tasks on tables connected through
primary-to-foreign key links. The RDL paradigm has been successfully applied to
recommendation lately, through its most recent representative deep learning
architecture namely, ContextGNN. While acknowledging ContextGNN's improved
performance on real-world recommendation datasets and tasks, preliminary tests
for the more traditional static link prediction task (aka personalized item
recommendation) on the popular Amazon Book dataset have demonstrated how
ContextGNN has still room for improvement compared to other state-of-the-art
GNN-based recommender systems. To this end, with this paper, we integrate
ContextGNN within Elliot, a popular framework for reproducibility and
benchmarking analyses, counting around 50 state-of-the-art recommendation
models from the literature to date. On such basis, we run preliminary
experiments on three standard recommendation datasets and against six
state-of-the-art GNN-based recommender systems, confirming similar trends to
those observed by the authors in their original paper. The code is publicly
available on GitHub:
https://github.com/danielemalitesta/Rel-DeepLearning-RecSys.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:17:09 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Ariza-Casabona",
"Alejandro",
""
],
[
"Kanakaris",
"Nikos",
""
],
[
"Malitesta",
"Daniele",
""
]
] | TITLE: ContextGNN goes to Elliot: Towards Benchmarking Relational Deep Learning
for Static Link Prediction (aka Personalized Item Recommendation)
ABSTRACT: Relational deep learning (RDL) settles among the most exciting advances in
machine learning for relational databases, leveraging the representational
power of message passing graph neural networks (GNNs) to derive useful
knowledge and run predicting tasks on tables connected through
primary-to-foreign key links. The RDL paradigm has been successfully applied to
recommendation lately, through its most recent representative deep learning
architecture namely, ContextGNN. While acknowledging ContextGNN's improved
performance on real-world recommendation datasets and tasks, preliminary tests
for the more traditional static link prediction task (aka personalized item
recommendation) on the popular Amazon Book dataset have demonstrated how
ContextGNN has still room for improvement compared to other state-of-the-art
GNN-based recommender systems. To this end, with this paper, we integrate
ContextGNN within Elliot, a popular framework for reproducibility and
benchmarking analyses, counting around 50 state-of-the-art recommendation
models from the literature to date. On such basis, we run preliminary
experiments on three standard recommendation datasets and against six
state-of-the-art GNN-based recommender systems, confirming similar trends to
those observed by the authors in their original paper. The code is publicly
available on GitHub:
https://github.com/danielemalitesta/Rel-DeepLearning-RecSys.
|
2503.16664 | Martin Kosteln\'ik | Martin Kosteln\'ik, Karel Bene\v{s}, Michal Hradi\v{s} | TextBite: A Historical Czech Document Dataset for Logical Page
Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Logical page segmentation is an important step in document analysis, enabling
better semantic representations, information retrieval, and text understanding.
Previous approaches define logical segmentation either through text or
geometric objects, relying on OCR or precise geometry. To avoid the need for
OCR, we define the task purely as segmentation in the image domain.
Furthermore, to ensure the evaluation remains unaffected by geometrical
variations that do not impact text segmentation, we propose to use only
foreground text pixels in the evaluation metric and disregard all background
pixels. To support research in logical document segmentation, we introduce
TextBite, a dataset of historical Czech documents spanning the 18th to 20th
centuries, featuring diverse layouts from newspapers, dictionaries, and
handwritten records. The dataset comprises 8,449 page images with 78,863
annotated segments of logically and thematically coherent text. We propose a
set of baseline methods combining text region detection and relation
prediction. The dataset, baselines and evaluation framework can be accessed at
https://github.com/DCGM/textbite-dataset.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:19:12 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kostelník",
"Martin",
""
],
[
"Beneš",
"Karel",
""
],
[
"Hradiš",
"Michal",
""
]
] | TITLE: TextBite: A Historical Czech Document Dataset for Logical Page
Segmentation
ABSTRACT: Logical page segmentation is an important step in document analysis, enabling
better semantic representations, information retrieval, and text understanding.
Previous approaches define logical segmentation either through text or
geometric objects, relying on OCR or precise geometry. To avoid the need for
OCR, we define the task purely as segmentation in the image domain.
Furthermore, to ensure the evaluation remains unaffected by geometrical
variations that do not impact text segmentation, we propose to use only
foreground text pixels in the evaluation metric and disregard all background
pixels. To support research in logical document segmentation, we introduce
TextBite, a dataset of historical Czech documents spanning the 18th to 20th
centuries, featuring diverse layouts from newspapers, dictionaries, and
handwritten records. The dataset comprises 8,449 page images with 78,863
annotated segments of logically and thematically coherent text. We propose a
set of baseline methods combining text region detection and relation
prediction. The dataset, baselines and evaluation framework can be accessed at
https://github.com/DCGM/textbite-dataset.
|
2503.16667 | Oriol Vendrell-Gallart | Oriol Vendrell-Gallart, Nima Negarandeh, Zahra Zanjani Foumani, Mahsa
Amiri, Lorenzo Valdevit, Ramin Bostanabad | A preliminary data fusion study to assess the feasibility of Foundation
Process-Property Models in Laser Powder Bed Fusion | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Foundation models are at the forefront of an increasing number of critical
applications. In regards to technologies such as additive manufacturing (AM),
these models have the potential to dramatically accelerate process optimization
and, in turn, design of next generation materials. A major challenge that
impedes the construction of foundation process-property models is data
scarcity. To understand the impact of this challenge, and since foundation
models rely on data fusion, in this work we conduct controlled experiments
where we focus on the transferability of information across different material
systems and properties. More specifically, we generate experimental datasets
from 17-4 PH and 316L stainless steels (SSs) in Laser Powder Bed Fusion (LPBF)
where we measure the effect of five process parameters on porosity and
hardness. We then leverage Gaussian processes (GPs) for process-property
modeling in various configurations to test if knowledge about one material
system or property can be leveraged to build more accurate machine learning
models for other material systems or properties. Through extensive
cross-validation studies and probing the GPs' interpretable hyperparameters, we
study the intricate relation among data size and dimensionality, complexity of
the process-property relations, noise, and characteristics of machine learning
models. Our findings highlight the need for structured learning approaches that
incorporate domain knowledge in building foundation process-property models
rather than relying on uninformed data fusion in data-limited applications.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:29:38 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Vendrell-Gallart",
"Oriol",
""
],
[
"Negarandeh",
"Nima",
""
],
[
"Foumani",
"Zahra Zanjani",
""
],
[
"Amiri",
"Mahsa",
""
],
[
"Valdevit",
"Lorenzo",
""
],
[
"Bostanabad",
"Ramin",
""
]
] | TITLE: A preliminary data fusion study to assess the feasibility of Foundation
Process-Property Models in Laser Powder Bed Fusion
ABSTRACT: Foundation models are at the forefront of an increasing number of critical
applications. In regards to technologies such as additive manufacturing (AM),
these models have the potential to dramatically accelerate process optimization
and, in turn, design of next generation materials. A major challenge that
impedes the construction of foundation process-property models is data
scarcity. To understand the impact of this challenge, and since foundation
models rely on data fusion, in this work we conduct controlled experiments
where we focus on the transferability of information across different material
systems and properties. More specifically, we generate experimental datasets
from 17-4 PH and 316L stainless steels (SSs) in Laser Powder Bed Fusion (LPBF)
where we measure the effect of five process parameters on porosity and
hardness. We then leverage Gaussian processes (GPs) for process-property
modeling in various configurations to test if knowledge about one material
system or property can be leveraged to build more accurate machine learning
models for other material systems or properties. Through extensive
cross-validation studies and probing the GPs' interpretable hyperparameters, we
study the intricate relation among data size and dimensionality, complexity of
the process-property relations, noise, and characteristics of machine learning
models. Our findings highlight the need for structured learning approaches that
incorporate domain knowledge in building foundation process-property models
rather than relying on uninformed data fusion in data-limited applications.
|
2503.16669 | Yichen Huang | Yichen Huang, Zachary Novack, Koichi Saito, Jiatong Shi, Shinji
Watanabe, Yuki Mitsufuji, John Thickstun, Chris Donahue | Aligning Text-to-Music Evaluation with Human Preferences | null | null | null | null | cs.SD cs.AI eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite significant recent advances in generative acoustic text-to-music
(TTM) modeling, robust evaluation of these models lags behind, relying in
particular on the popular Fr\'echet Audio Distance (FAD). In this work, we
rigorously study the design space of reference-based divergence metrics for
evaluating TTM models through (1) designing four synthetic meta-evaluations to
measure sensitivity to particular musical desiderata, and (2) collecting and
evaluating on MusicPrefs, the first open-source dataset of human preferences
for TTM systems. We find that not only is the standard FAD setup inconsistent
on both synthetic and human preference data, but that nearly all existing
metrics fail to effectively capture desiderata, and are only weakly correlated
with human perception. We propose a new metric, the MAUVE Audio Divergence
(MAD), computed on representations from a self-supervised audio embedding
model. We find that this metric effectively captures diverse musical desiderata
(average rank correlation 0.84 for MAD vs. 0.49 for FAD and also correlates
more strongly with MusicPrefs (0.62 vs. 0.14).
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:31:04 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Huang",
"Yichen",
""
],
[
"Novack",
"Zachary",
""
],
[
"Saito",
"Koichi",
""
],
[
"Shi",
"Jiatong",
""
],
[
"Watanabe",
"Shinji",
""
],
[
"Mitsufuji",
"Yuki",
""
],
[
"Thickstun",
"John",
""
],
[
"Donahue",
"Chris",
""
]
] | TITLE: Aligning Text-to-Music Evaluation with Human Preferences
ABSTRACT: Despite significant recent advances in generative acoustic text-to-music
(TTM) modeling, robust evaluation of these models lags behind, relying in
particular on the popular Fr\'echet Audio Distance (FAD). In this work, we
rigorously study the design space of reference-based divergence metrics for
evaluating TTM models through (1) designing four synthetic meta-evaluations to
measure sensitivity to particular musical desiderata, and (2) collecting and
evaluating on MusicPrefs, the first open-source dataset of human preferences
for TTM systems. We find that not only is the standard FAD setup inconsistent
on both synthetic and human preference data, but that nearly all existing
metrics fail to effectively capture desiderata, and are only weakly correlated
with human perception. We propose a new metric, the MAUVE Audio Divergence
(MAD), computed on representations from a self-supervised audio embedding
model. We find that this metric effectively captures diverse musical desiderata
(average rank correlation 0.84 for MAD vs. 0.49 for FAD and also correlates
more strongly with MusicPrefs (0.62 vs. 0.14).
|
2503.16674 | Molly Kennedy | Molly Kennedy, Ayyoob Imani, Timo Spinde, Hinrich Sch\"utze | Through the LLM Looking Glass: A Socratic Self-Assessment of Donkeys,
Elephants, and Markets | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While detecting and avoiding bias in LLM-generated text is becoming
increasingly important, media bias often remains subtle and subjective, making
it particularly difficult to identify and mitigate. In this study, we assess
media bias in LLM-generated content and LLMs' ability to detect subtle
ideological bias. We conduct this evaluation using two datasets, PoliGen and
EconoLex, covering political and economic discourse, respectively. We evaluate
eight widely used LLMs by prompting them to generate articles and analyze their
ideological preferences via self-assessment. By using self-assessment, the
study aims to directly measure the models' biases rather than relying on
external interpretations, thereby minimizing subjective judgments about media
bias. Our results reveal a consistent preference of Democratic over Republican
positions across all models. Conversely, in economic topics, biases vary among
Western LLMs, while those developed in China lean more strongly toward
socialism.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:40:40 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Kennedy",
"Molly",
""
],
[
"Imani",
"Ayyoob",
""
],
[
"Spinde",
"Timo",
""
],
[
"Schütze",
"Hinrich",
""
]
] | TITLE: Through the LLM Looking Glass: A Socratic Self-Assessment of Donkeys,
Elephants, and Markets
ABSTRACT: While detecting and avoiding bias in LLM-generated text is becoming
increasingly important, media bias often remains subtle and subjective, making
it particularly difficult to identify and mitigate. In this study, we assess
media bias in LLM-generated content and LLMs' ability to detect subtle
ideological bias. We conduct this evaluation using two datasets, PoliGen and
EconoLex, covering political and economic discourse, respectively. We evaluate
eight widely used LLMs by prompting them to generate articles and analyze their
ideological preferences via self-assessment. By using self-assessment, the
study aims to directly measure the models' biases rather than relying on
external interpretations, thereby minimizing subjective judgments about media
bias. Our results reveal a consistent preference of Democratic over Republican
positions across all models. Conversely, in economic topics, biases vary among
Western LLMs, while those developed in China lean more strongly toward
socialism.
|
2503.16693 | Zhan Cheng | Zhan Cheng, Bolin Shen, Tianming Sha, Yuan Gao, Shibo Li and Yushun
Dong | ATOM: A Framework of Detecting Query-Based Model Extraction Attacks for
Graph Neural Networks | null | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by/4.0/ | Graph Neural Networks (GNNs) have gained traction in Graph-based Machine
Learning as a Service (GMLaaS) platforms, yet they remain vulnerable to
graph-based model extraction attacks (MEAs), where adversaries reconstruct
surrogate models by querying the victim model. Existing defense mechanisms,
such as watermarking and fingerprinting, suffer from poor real-time
performance, susceptibility to evasion, or reliance on post-attack
verification, making them inadequate for handling the dynamic characteristics
of graph-based MEA variants. To address these limitations, we propose ATOM, a
novel real-time MEA detection framework tailored for GNNs. ATOM integrates
sequential modeling and reinforcement learning to dynamically detect evolving
attack patterns, while leveraging $k$-core embedding to capture the structural
properties, enhancing detection precision. Furthermore, we provide theoretical
analysis to characterize query behaviors and optimize detection strategies.
Extensive experiments on multiple real-world datasets demonstrate that ATOM
outperforms existing approaches in detection performance, maintaining stable
across different time steps, thereby offering a more effective defense
mechanism for GMLaaS environments.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 20:25:32 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Cheng",
"Zhan",
""
],
[
"Shen",
"Bolin",
""
],
[
"Sha",
"Tianming",
""
],
[
"Gao",
"Yuan",
""
],
[
"Li",
"Shibo",
""
],
[
"Dong",
"Yushun",
""
]
] | TITLE: ATOM: A Framework of Detecting Query-Based Model Extraction Attacks for
Graph Neural Networks
ABSTRACT: Graph Neural Networks (GNNs) have gained traction in Graph-based Machine
Learning as a Service (GMLaaS) platforms, yet they remain vulnerable to
graph-based model extraction attacks (MEAs), where adversaries reconstruct
surrogate models by querying the victim model. Existing defense mechanisms,
such as watermarking and fingerprinting, suffer from poor real-time
performance, susceptibility to evasion, or reliance on post-attack
verification, making them inadequate for handling the dynamic characteristics
of graph-based MEA variants. To address these limitations, we propose ATOM, a
novel real-time MEA detection framework tailored for GNNs. ATOM integrates
sequential modeling and reinforcement learning to dynamically detect evolving
attack patterns, while leveraging $k$-core embedding to capture the structural
properties, enhancing detection precision. Furthermore, we provide theoretical
analysis to characterize query behaviors and optimize detection strategies.
Extensive experiments on multiple real-world datasets demonstrate that ATOM
outperforms existing approaches in detection performance, maintaining stable
across different time steps, thereby offering a more effective defense
mechanism for GMLaaS environments.
|
2503.16718 | Massa Baali | Massa Baali, Xiang Li, Hao Chen, Rita Singh, Bhiksha Raj | CAARMA: Class Augmentation with Adversarial Mixup Regularization | null | null | null | null | cs.SD cs.CL cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Speaker verification is a typical zero-shot learning task, where inference of
unseen classes is performed by comparing embeddings of test instances to known
examples. The models performing inference must hence naturally generate
embeddings that cluster same-class instances compactly, while maintaining
separation across classes. In order to learn to do so, they are typically
trained on a large number of classes (speakers), often using specialized
losses. However real-world speaker datasets often lack the class diversity
needed to effectively learn this in a generalizable manner. We introduce
CAARMA, a class augmentation framework that addresses this problem by
generating synthetic classes through data mixing in the embedding space,
expanding the number of training classes. To ensure the authenticity of the
synthetic classes we adopt a novel adversarial refinement mechanism that
minimizes categorical distinctions between synthetic and real classes. We
evaluate CAARMA on multiple speaker verification tasks, as well as other
representative zero-shot comparison-based speech analysis tasks and obtain
consistent improvements: our framework demonstrates a significant improvement
of 8\% over all baseline models. Code for CAARMA will be released.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 21:41:16 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Baali",
"Massa",
""
],
[
"Li",
"Xiang",
""
],
[
"Chen",
"Hao",
""
],
[
"Singh",
"Rita",
""
],
[
"Raj",
"Bhiksha",
""
]
] | TITLE: CAARMA: Class Augmentation with Adversarial Mixup Regularization
ABSTRACT: Speaker verification is a typical zero-shot learning task, where inference of
unseen classes is performed by comparing embeddings of test instances to known
examples. The models performing inference must hence naturally generate
embeddings that cluster same-class instances compactly, while maintaining
separation across classes. In order to learn to do so, they are typically
trained on a large number of classes (speakers), often using specialized
losses. However real-world speaker datasets often lack the class diversity
needed to effectively learn this in a generalizable manner. We introduce
CAARMA, a class augmentation framework that addresses this problem by
generating synthetic classes through data mixing in the embedding space,
expanding the number of training classes. To ensure the authenticity of the
synthetic classes we adopt a novel adversarial refinement mechanism that
minimizes categorical distinctions between synthetic and real classes. We
evaluate CAARMA on multiple speaker verification tasks, as well as other
representative zero-shot comparison-based speech analysis tasks and obtain
consistent improvements: our framework demonstrates a significant improvement
of 8\% over all baseline models. Code for CAARMA will be released.
|
2503.16724 | Zhaoxin Li | Zhaoxin Li, Zhang Xi-Jia, Batuhan Altundas, Letian Chen, Rohan Paleja,
Matthew Gombolay | Towards Automated Semantic Interpretability in Reinforcement Learning
via Vision-Language Models | null | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semantic Interpretability in Reinforcement Learning (RL) enables
transparency, accountability, and safer deployment by making the agent's
decisions understandable and verifiable. Achieving this, however, requires a
feature space composed of human-understandable concepts, which traditionally
rely on human specification and fail to generalize to unseen environments. In
this work, we introduce Semantically Interpretable Reinforcement Learning with
Vision-Language Models Empowered Automation (SILVA), an automated framework
that leverages pre-trained vision-language models (VLM) for semantic feature
extraction and interpretable tree-based models for policy optimization. SILVA
first queries a VLM to identify relevant semantic features for an unseen
environment, then extracts these features from the environment. Finally, it
trains an Interpretable Control Tree via RL, mapping the extracted features to
actions in a transparent and interpretable manner. To address the computational
inefficiency of extracting features directly with VLMs, we develop a feature
extraction pipeline that generates a dataset for training a lightweight
convolutional network, which is subsequently used during RL. By leveraging VLMs
to automate tree-based RL, SILVA removes the reliance on human annotation
previously required by interpretable models while also overcoming the inability
of VLMs alone to generate valid robot policies, enabling semantically
interpretable reinforcement learning without human-in-the-loop.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 21:53:19 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Li",
"Zhaoxin",
""
],
[
"Xi-Jia",
"Zhang",
""
],
[
"Altundas",
"Batuhan",
""
],
[
"Chen",
"Letian",
""
],
[
"Paleja",
"Rohan",
""
],
[
"Gombolay",
"Matthew",
""
]
] | TITLE: Towards Automated Semantic Interpretability in Reinforcement Learning
via Vision-Language Models
ABSTRACT: Semantic Interpretability in Reinforcement Learning (RL) enables
transparency, accountability, and safer deployment by making the agent's
decisions understandable and verifiable. Achieving this, however, requires a
feature space composed of human-understandable concepts, which traditionally
rely on human specification and fail to generalize to unseen environments. In
this work, we introduce Semantically Interpretable Reinforcement Learning with
Vision-Language Models Empowered Automation (SILVA), an automated framework
that leverages pre-trained vision-language models (VLM) for semantic feature
extraction and interpretable tree-based models for policy optimization. SILVA
first queries a VLM to identify relevant semantic features for an unseen
environment, then extracts these features from the environment. Finally, it
trains an Interpretable Control Tree via RL, mapping the extracted features to
actions in a transparent and interpretable manner. To address the computational
inefficiency of extracting features directly with VLMs, we develop a feature
extraction pipeline that generates a dataset for training a lightweight
convolutional network, which is subsequently used during RL. By leveraging VLMs
to automate tree-based RL, SILVA removes the reliance on human annotation
previously required by interpretable models while also overcoming the inability
of VLMs alone to generate valid robot policies, enabling semantically
interpretable reinforcement learning without human-in-the-loop.
|
2503.16730 | Srijan Sengupta | Subhankar Bhadra and Marianna Pensky and Srijan Sengupta | Scalable community detection in massive networks via predictive
assignment | null | null | null | null | stat.ME cs.SI physics.soc-ph | http://creativecommons.org/licenses/by/4.0/ | Massive network datasets are becoming increasingly common in scientific
applications. Existing community detection methods encounter significant
computational challenges for such massive networks due to two reasons. First,
the full network needs to be stored and analyzed on a single server, leading to
high memory costs. Second, existing methods typically use matrix factorization
or iterative optimization using the full network, resulting in high runtimes.
We propose a strategy called \textit{predictive assignment} to enable
computationally efficient community detection while ensuring statistical
accuracy. The core idea is to avoid large-scale matrix computations by breaking
up the task into a smaller matrix computation plus a large number of vector
computations that can be carried out in parallel. Under the proposed method,
community detection is carried out on a small subgraph to estimate the relevant
model parameters. Next, each remaining node is assigned to a community based on
these estimates. We prove that predictive assignment achieves strong
consistency under the stochastic blockmodel and its degree-corrected version.
We also demonstrate the empirical performance of predictive assignment on
simulated networks and two large real-world datasets: DBLP (Digital
Bibliography \& Library Project), a computer science bibliographical database,
and the Twitch Gamers Social Network.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 22:14:06 GMT"
}
] | 2025-03-24T00:00:00 | [
[
"Bhadra",
"Subhankar",
""
],
[
"Pensky",
"Marianna",
""
],
[
"Sengupta",
"Srijan",
""
]
] | TITLE: Scalable community detection in massive networks via predictive
assignment
ABSTRACT: Massive network datasets are becoming increasingly common in scientific
applications. Existing community detection methods encounter significant
computational challenges for such massive networks due to two reasons. First,
the full network needs to be stored and analyzed on a single server, leading to
high memory costs. Second, existing methods typically use matrix factorization
or iterative optimization using the full network, resulting in high runtimes.
We propose a strategy called \textit{predictive assignment} to enable
computationally efficient community detection while ensuring statistical
accuracy. The core idea is to avoid large-scale matrix computations by breaking
up the task into a smaller matrix computation plus a large number of vector
computations that can be carried out in parallel. Under the proposed method,
community detection is carried out on a small subgraph to estimate the relevant
model parameters. Next, each remaining node is assigned to a community based on
these estimates. We prove that predictive assignment achieves strong
consistency under the stochastic blockmodel and its degree-corrected version.
We also demonstrate the empirical performance of predictive assignment on
simulated networks and two large real-world datasets: DBLP (Digital
Bibliography \& Library Project), a computer science bibliographical database,
and the Twitch Gamers Social Network.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.