id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 3
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-08-22 00:00:00
| authors_parsed
listlengths 1
3.31k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2508.15216
|
Vipooshan Vipulananthan
|
Vipooshan Vipulananthan, Kumudu Mohottala, Kavindu Chinthana, Nimsara Paramulla, Charith D Chitraranjan
|
STAGNet: A Spatio-Temporal Graph and LSTM Framework for Accident Anticipation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Accident prediction and timely warnings play a key role in improving road safety by reducing the risk of injury to road users and minimizing property damage. Advanced Driver Assistance Systems (ADAS) are designed to support human drivers and are especially useful when they can anticipate potential accidents before they happen. While many existing systems depend on a range of sensors such as LiDAR, radar, and GPS, relying solely on dash-cam video input presents a more challenging but a more cost-effective and easily deployable solution. In this work, we incorporate better spatio-temporal features and aggregate them through a recurrent network to improve upon state-of-the-art graph neural networks for predicting accidents from dash-cam videos. Experiments using three publicly available datasets show that our proposed STAGNet model achieves higher average precision and mean time-to-collision values than previous methods, both when cross-validated on a given dataset and when trained and tested on different datasets.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:02:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Vipulananthan",
"Vipooshan",
""
],
[
"Mohottala",
"Kumudu",
""
],
[
"Chinthana",
"Kavindu",
""
],
[
"Paramulla",
"Nimsara",
""
],
[
"Chitraranjan",
"Charith D",
""
]
] |
2508.15217
|
Sishuo Chen
|
Sishuo Chen, Zhangming Chan, Xiang-Rong Sheng, Lei Zhang, Sheng Chen, Chenghuan Hou, Han Zhu, Jian Xu, Bo Zheng
|
See Beyond a Single View: Multi-Attribution Learning Leads to Better Conversion Rate Prediction
|
Accepted at CIKM 2025
| null | null | null |
cs.LG cs.IR
|
http://creativecommons.org/licenses/by/4.0/
|
Conversion rate (CVR) prediction is a core component of online advertising systems, where the attribution mechanisms-rules for allocating conversion credit across user touchpoints-fundamentally determine label generation and model optimization. While many industrial platforms support diverse attribution mechanisms (e.g., First-Click, Last-Click, Linear, and Data-Driven Multi-Touch Attribution), conventional approaches restrict model training to labels from a single production-critical attribution mechanism, discarding complementary signals in alternative attribution perspectives.
To address this limitation, we propose a novel Multi-Attribution Learning (MAL) framework for CVR prediction that integrates signals from multiple attribution perspectives to better capture the underlying patterns driving user conversions. Specifically, MAL is a joint learning framework consisting of two core components: the Attribution Knowledge Aggregator (AKA) and the Primary Target Predictor (PTP). AKA is implemented as a multi-task learner that integrates knowledge extracted from diverse attribution labels. PTP, in contrast, focuses on the task of generating well-calibrated conversion probabilities that align with the system-optimized attribution metric (e.g., CVR under the Last-Click attribution), ensuring direct compatibility with industrial deployment requirements. Additionally, we propose CAT, a novel training strategy that leverages the Cartesian product of all attribution label combinations to generate enriched supervision signals. This design substantially enhances the performance of the attribution knowledge aggregator. Empirical evaluations demonstrate the superiority of MAL over single-attribution learning baselines, achieving +0.51% GAUC improvement on offline metrics. Online experiments demonstrate that MAL achieved a +2.6% increase in ROI (Return on Investment).
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:05:31 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chen",
"Sishuo",
""
],
[
"Chan",
"Zhangming",
""
],
[
"Sheng",
"Xiang-Rong",
""
],
[
"Zhang",
"Lei",
""
],
[
"Chen",
"Sheng",
""
],
[
"Hou",
"Chenghuan",
""
],
[
"Zhu",
"Han",
""
],
[
"Xu",
"Jian",
""
],
[
"Zheng",
"Bo",
""
]
] |
2508.15218
|
Momoka Furuhashi
|
Momoka Furuhashi, Kouta Nakayama, Takashi Kodama, and Saku Sugawara
|
Are Checklists Really Useful for Automatic Evaluation of Generative Tasks?
|
Accepted to the EMNLP 2025 Main Conference
| null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Automatic evaluation of generative tasks using large language models faces challenges due to ambiguous criteria. Although automatic checklist generation is a potentially promising approach, its usefulness remains underexplored. We investigate whether checklists should be used for all questions or selectively, generate them using six methods, evaluate their effectiveness across eight model sizes, and identify checklist items that correlate with human evaluations. Through experiments on pairwise comparison and direct scoring tasks, we find that selective checklist use tends to improve evaluation performance in pairwise settings, while its benefits are less consistent in direct scoring. Our analysis also shows that even checklist items with low correlation to human scores often reflect human-written criteria, indicating potential inconsistencies in human evaluation. These findings highlight the need to more clearly define objective evaluation criteria to guide both human and automatic evaluations. \footnote{Our code is available at~https://github.com/momo0817/checklist-effectiveness-study
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:07:45 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Furuhashi",
"Momoka",
""
],
[
"Nakayama",
"Kouta",
""
],
[
"Kodama",
"Takashi",
""
],
[
"Sugawara",
"Saku",
""
]
] |
2508.15219
|
Muktish Acharyya
|
Ishita Tikader and Muktish Acharyya
|
Multiple crossover in the decay of metastable volume fraction of a Blume-Capel ferromagnetic needle
|
10 Pages Latex, 12 captioned figures, 5 Tables
| null | null |
PU-Physics-21-8-2025
|
cond-mat.stat-mech
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The transient behaviours of a Blume-Capel ferromagnetic needle have been studied extensively by Monte-Carlo simulation. The needle has an elongated length in one direction compared to its cross-section. In the context of transient behaviour, we have captured the decay of the metastable state and the magnetic relaxation behaviours in our study. The dependence of metastable behaviour with anisotropy (single site) has been studied. \emph{Interestingly, we have observed multiple (different values of $n$ in different time domains) crossover in the decay of metastable volume fraction (obeying Avrami's law $\beta \sim {\rm exp}(-Kt^n)$). We have identified the crossover time, and the values of $n$ are estimated precisely.} The mean (magnetic) reversal time has been studied as a function of anisotropy. It is observed to be almost independent of anisotropy for its negative value; however, it is found to decrease exponentially with positive anisotropy. The exponential relaxation behaviour (in the corresponding paramagnetic phase) is observed. The range of the relaxation time is observed to decrease as the temperature increases.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:08:33 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Tikader",
"Ishita",
""
],
[
"Acharyya",
"Muktish",
""
]
] |
2508.15220
|
Aniruddha Joshi
|
Aniruddha Joshi, Supratik Chakraborty, S Akshay, Shetal Shah, Hazem Torfah and Sanjit Seshia
|
Locally Pareto-Optimal Interpretations for Black-Box Machine Learning Models
|
This work has been accepted at ATVA'25
| null | null | null |
cs.LG cs.AI cs.LO
|
http://creativecommons.org/licenses/by/4.0/
|
Creating meaningful interpretations for black-box machine learning models involves balancing two often conflicting objectives: accuracy and explainability. Exploring the trade-off between these objectives is essential for developing trustworthy interpretations. While many techniques for multi-objective interpretation synthesis have been developed, they typically lack formal guarantees on the Pareto-optimality of the results. Methods that do provide such guarantees, on the other hand, often face severe scalability limitations when exploring the Pareto-optimal space. To address this, we develop a framework based on local optimality guarantees that enables more scalable synthesis of interpretations. Specifically, we consider the problem of synthesizing a set of Pareto-optimal interpretations with local optimality guarantees, within the immediate neighborhood of each solution. Our approach begins with a multi-objective learning or search technique, such as Multi-Objective Monte Carlo Tree Search, to generate a best-effort set of Pareto-optimal candidates with respect to accuracy and explainability. We then verify local optimality for each candidate as a Boolean satisfiability problem, which we solve using a SAT solver. We demonstrate the efficacy of our approach on a set of benchmarks, comparing it against previous methods for exploring the Pareto-optimal front of interpretations. In particular, we show that our approach yields interpretations that closely match those synthesized by methods offering global guarantees.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:11:20 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Joshi",
"Aniruddha",
""
],
[
"Chakraborty",
"Supratik",
""
],
[
"Akshay",
"S",
""
],
[
"Shah",
"Shetal",
""
],
[
"Torfah",
"Hazem",
""
],
[
"Seshia",
"Sanjit",
""
]
] |
2508.15221
|
Xiao-Ping Chen
|
Xiao-Ping Chen and Chun-Lei Tang
|
On the extremal functions of second order uncertainty principles: symmetry and symmetry breaking
| null | null | null | null |
math.AP
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
This paper focus on the symmetry and symmetry breaking about the second order Hydrogen Uncertainty Principle. \emph{Firstly}, by choosing a suitable test function, we give a negative answer to the conjecture presented by Cazacu, Flynn and Lam in [\emph{J. Funct. Anal.} \textbf{283} (2022), Paper No. 109659, 37 pp] for $N\in\{2,3\}$, and emphasizing the symmetry breaking phenomenon. \emph{Secondly}, we obtain a family of sharp weighted second order Hydrogen Uncertainty Principle, and prove the extremal functions are radial, which extends the work of Duong and Nguyen [The sharp second order Caffareli-Kohn-Nirenberg inequality and stability estimates for the sharp second order uncertainty principle, arXiv:2102.01425].
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:19:45 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chen",
"Xiao-Ping",
""
],
[
"Tang",
"Chun-Lei",
""
]
] |
2508.15222
|
T. Ed Li
|
Hantao Zhang, Jingyang Liu, Ed Li
|
See it. Say it. Sorted: Agentic System for Compositional Diagram Generation
| null | null | null | null |
cs.AI cs.CV cs.MA
|
http://creativecommons.org/licenses/by/4.0/
|
We study sketch-to-diagram generation: converting rough hand sketches into precise, compositional diagrams. Diffusion models excel at photorealism but struggle with the spatial precision, alignment, and symbolic structure required for flowcharts. We introduce See it. Say it. Sorted., a training-free agentic system that couples a Vision-Language Model (VLM) with Large Language Models (LLMs) to produce editable Scalable Vector Graphics (SVG) programs. The system runs an iterative loop in which a Critic VLM proposes a small set of qualitative, relational edits; multiple candidate LLMs synthesize SVG updates with diverse strategies (conservative->aggressive, alternative, focused); and a Judge VLM selects the best candidate, ensuring stable improvement. This design prioritizes qualitative reasoning over brittle numerical estimates, preserves global constraints (e.g., alignment, connectivity), and naturally supports human-in-the-loop corrections. On 10 sketches derived from flowcharts in published papers, our method more faithfully reconstructs layout and structure than two frontier closed-source image generation LLMs (GPT-5 and Gemini-2.5-Pro), accurately composing primitives (e.g., multi-headed arrows) without inserting unwanted text. Because outputs are programmatic SVGs, the approach is readily extensible to presentation tools (e.g., PowerPoint) via APIs and can be specialized with improved prompts and task-specific tools. The codebase is open-sourced at https://github.com/hantaoZhangrichard/see_it_say_it_sorted.git.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:20:36 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Hantao",
""
],
[
"Liu",
"Jingyang",
""
],
[
"Li",
"Ed",
""
]
] |
2508.15223
|
Yasuhito Kobayashi
|
Yasuhito Kobayashi, Hidefumi Takahashi, Shunsuke Kitou, Akitoshi Nakano, Hajime Sagayama, Yuichi Yamasaki, Shintaro Ishiwata
|
A rutile-based homologous series Na(PtO$_2$)$_{2\it{n}+1}$ discovered by computationally assisted high-pressure synthesis
| null | null |
10.1021/acs.inorgchem.5c02074
| null |
cond-mat.mtrl-sci cond-mat.str-el
|
http://creativecommons.org/licenses/by/4.0/
|
Layered transition metal oxides typified by the Ruddlesden-Popper phase have been extensively studied for its applications in high-temperature superconductivity, catalysis, and battery technologies. Despite the remarkable structural diversity and catalytic functionality of platinum oxides, the exploration of layered polymorphs has remained significantly constrained mainly due to the high inertness of platinum. Here, we discover a new homologous series of layered ternary oxides, Na(PtO$_2$)$_{2\it{n}+1}$, by a combination of highly oxidizing high-pressure methods and density functional theory (DFT) calculations. This series features unprecedented layered structural motifs, rutile-based PtO$_6$ octahedra and one-dimensional PtO$_4$ square-planar columns, which enables systematic control of dimensionality. Furthermore, we demonstrate a computationally-assisted identification of isomeric and putative members of this homologous series as confirmed by controlled synthesis and quantitative analysis of diffuse scattering data. This approach provides an effective platform for the exhaustive exploration of metastable transition metal oxides with rich structural variations.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:23:19 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kobayashi",
"Yasuhito",
""
],
[
"Takahashi",
"Hidefumi",
""
],
[
"Kitou",
"Shunsuke",
""
],
[
"Nakano",
"Akitoshi",
""
],
[
"Sagayama",
"Hajime",
""
],
[
"Yamasaki",
"Yuichi",
""
],
[
"Ishiwata",
"Shintaro",
""
]
] |
2508.15224
|
Georgii Shestakov
|
N. N. Achasov and G. N. Shestakov
|
Phenomenological description of the $D^*_{s0}(2317)\to D_s\pi^0$ decay
|
10 pages, 7 figures
| null | null | null |
hep-ph hep-ex hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
For coupled channels $D^0K^+$, $D^+K^0$, $D_s^+\eta$, and $D_s^+\pi^0$, the $S$-wave scattering amplitudes are constructed taking into account the mixing of the isoscalar resonance $D^*_{s0}(2317)^+$ with nonresonance amplitudes with isospin $I=1$. The phenomenological approach we use allows us to quite simply clear up the general structure of the $D^*_{s0}(2317)^+\to D_s^+\pi^0$ decay amplitude violating isospin. We show that the phase of this amplitude coincides with the phase of the nonresonanct $D_s^+\pi^0 $ scattering amplitude in agreement with the Watson theorem. Its modulus squared, as it should be, determines the width of the resonance peak in the $D_s^+\pi^0$ channel. Taking into account the $\pi^0-\eta$ mixing in internal lines up to the second order inclusively ensures that the unitarity condition is fulfilled. The presented analysis complements the description of the $D^*_{s0}(2317 )^+\to D_s^+\pi^0$ decay based on the coupled channel unitarized chiral perturbation theory. The numerical estimates obtained by us for the $D^*_{s0}(2317)^+\to D_s^+\pi^0$ decay width do not contradict those available in the literature.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:25:02 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Achasov",
"N. N.",
""
],
[
"Shestakov",
"G. N.",
""
]
] |
2508.15225
|
Alex Fedorov
|
Yi Yuan, Joseph Van Duyn, Runze Yan, Zhuoyi Huang, Sulaiman Vesal, Sergey Plis, Xiao Hu, Gloria Hyunjung Kwak, Ran Xiao, Alex Fedorov
|
Learning ECG Representations via Poly-Window Contrastive Learning
|
This work has been accepted for publication in IEEE-EMBS International Conference on Biomedical and Health Informatics 2025. The final published version will be available via IEEE Xplore
| null | null | null |
cs.LG eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Electrocardiogram (ECG) analysis is foundational for cardiovascular disease diagnosis, yet the performance of deep learning models is often constrained by limited access to annotated data. Self-supervised contrastive learning has emerged as a powerful approach for learning robust ECG representations from unlabeled signals. However, most existing methods generate only pairwise augmented views and fail to leverage the rich temporal structure of ECG recordings. In this work, we present a poly-window contrastive learning framework. We extract multiple temporal windows from each ECG instance to construct positive pairs and maximize their agreement via statistics. Inspired by the principle of slow feature analysis, our approach explicitly encourages the model to learn temporally invariant and physiologically meaningful features that persist across time. We validate our approach through extensive experiments and ablation studies on the PTB-XL dataset. Our results demonstrate that poly-window contrastive learning consistently outperforms conventional two-view methods in multi-label superclass classification, achieving higher AUROC (0.891 vs. 0.888) and F1 scores (0.680 vs. 0.679) while requiring up to four times fewer pre-training epochs (32 vs. 128) and 14.8% in total wall clock pre-training time reduction. Despite processing multiple windows per sample, we achieve a significant reduction in the number of training epochs and total computation time, making our method practical for training foundational models. Through extensive ablations, we identify optimal design choices and demonstrate robustness across various hyperparameters. These findings establish poly-window contrastive learning as a highly efficient and scalable paradigm for automated ECG analysis and provide a promising general framework for self-supervised representation learning in biomedical time-series data.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:26:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yuan",
"Yi",
""
],
[
"Van Duyn",
"Joseph",
""
],
[
"Yan",
"Runze",
""
],
[
"Huang",
"Zhuoyi",
""
],
[
"Vesal",
"Sulaiman",
""
],
[
"Plis",
"Sergey",
""
],
[
"Hu",
"Xiao",
""
],
[
"Kwak",
"Gloria Hyunjung",
""
],
[
"Xiao",
"Ran",
""
],
[
"Fedorov",
"Alex",
""
]
] |
2508.15226
|
Yong-Bo Tang
|
Yong-Bo Tang
|
Precision calculation of hyperfine-structure constants for extracting nuclear quadrupole moment of $^{43}$Ca
|
12
| null | null | null |
physics.atom-ph
|
http://creativecommons.org/licenses/by/4.0/
|
There have been several reported values for the nuclear quadrupole moment of $^{43}$Ca, but significant discrepancies exist among these reported values, ranging from \(-0.0408(8)\)~b to \(-0.065(20)\)~b. In this work, we performed an accurate calculation of the electric field gradients of the \(4s4p~^3\!P_{1}\), \(4s4p~^3\!P_{2}\) and \(4s3d~^1\!D_2\) states in the $^{43}$Ca atom using a hybrid method. This hybrid method integrates the advantages of the configuration interaction method and the coupled-cluster method, and can simultaneously account for core-core, core-valence, and valence-valence correlations. By combining our calculated results with the experimental values of the electric quadrupole hyperfine-structure constants of these three states, an accurate and reliable nuclear quadrupole moment of $^{43}$Ca was determined to be \(-0.0479(6)\)~b, which could be recommended as a reference for \(^{43}\text{Ca}\).
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:29:43 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Tang",
"Yong-Bo",
""
]
] |
2508.15227
|
WenFan Wang
|
Wen-Fan Wang, Ting-Ying Lee, Chien-Ting Lu, Che-Wei Hsu, Nil Ponsa Campany, Yu Chen, Mike Y. Chen, Bing-Yu Chen
|
GenTune: Toward Traceable Prompts to Improve Controllability of Image Refinement in Environment Design
|
Accepted ACM Symposium on User Interface Software and Technology (UIST '25)
| null |
10.1145/3746059.3747774
| null |
cs.HC cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Environment designers in the entertainment industry create imaginative 2D and 3D scenes for games, films, and television, requiring both fine-grained control of specific details and consistent global coherence. Designers have increasingly integrated generative AI into their workflows, often relying on large language models (LLMs) to expand user prompts for text-to-image generation, then iteratively refining those prompts and applying inpainting. However, our formative study with 10 designers surfaced two key challenges: (1) the lengthy LLM-generated prompts make it difficult to understand and isolate the keywords that must be revised for specific visual elements; and (2) while inpainting supports localized edits, it can struggle with global consistency and correctness. Based on these insights, we present GenTune, an approach that enhances human--AI collaboration by clarifying how AI-generated prompts map to image content. Our GenTune system lets designers select any element in a generated image, trace it back to the corresponding prompt labels, and revise those labels to guide precise yet globally consistent image refinement. In a summative study with 20 designers, GenTune significantly improved prompt--image comprehension, refinement quality, and efficiency, and overall satisfaction (all $p < .01$) compared to current practice. A follow-up field study with two studios further demonstrated its effectiveness in real-world settings.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:31:01 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Wen-Fan",
""
],
[
"Lee",
"Ting-Ying",
""
],
[
"Lu",
"Chien-Ting",
""
],
[
"Hsu",
"Che-Wei",
""
],
[
"Campany",
"Nil Ponsa",
""
],
[
"Chen",
"Yu",
""
],
[
"Chen",
"Mike Y.",
""
],
[
"Chen",
"Bing-Yu",
""
]
] |
2508.15228
|
Ziang Cao
|
Ziang Cao, Zhaoxi Chen, Liang Pan, Ziwei Liu
|
Collaborative Multi-Modal Coding for High-Quality 3D Generation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
3D content inherently encompasses multi-modal characteristics and can be projected into different modalities (e.g., RGB images, RGBD, and point clouds). Each modality exhibits distinct advantages in 3D asset modeling: RGB images contain vivid 3D textures, whereas point clouds define fine-grained 3D geometries. However, most existing 3D-native generative architectures either operate predominantly within single-modality paradigms-thus overlooking the complementary benefits of multi-modality data-or restrict themselves to 3D structures, thereby limiting the scope of available training datasets. To holistically harness multi-modalities for 3D modeling, we present TriMM, the first feed-forward 3D-native generative model that learns from basic multi-modalities (e.g., RGB, RGBD, and point cloud). Specifically, 1) TriMM first introduces collaborative multi-modal coding, which integrates modality-specific features while preserving their unique representational strengths. 2) Furthermore, auxiliary 2D and 3D supervision are introduced to raise the robustness and performance of multi-modal coding. 3) Based on the embedded multi-modal code, TriMM employs a triplane latent diffusion model to generate 3D assets of superior quality, enhancing both the texture and the geometric detail. Extensive experiments on multiple well-known datasets demonstrate that TriMM, by effectively leveraging multi-modality, achieves competitive performance with models trained on large-scale datasets, despite utilizing a small amount of training data. Furthermore, we conduct additional experiments on recent RGB-D datasets, verifying the feasibility of incorporating other multi-modal datasets into 3D generation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:31:14 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Cao",
"Ziang",
""
],
[
"Chen",
"Zhaoxi",
""
],
[
"Pan",
"Liang",
""
],
[
"Liu",
"Ziwei",
""
]
] |
2508.15229
|
Yayu Zhou
|
Hanling Zhang, Yayu Zhou, Tongcheng Fang, Zhihang Yuan, Guohao Dai, Yu Wang
|
VocabTailor: Dynamic Vocabulary Selection for Downstream Tasks in Small Language Models
| null | null | null | null |
cs.CL cs.AI cs.LG
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Small Language Models (SLMs) provide computational advantages in resource-constrained environments, yet memory limitations remain a critical bottleneck for edge device deployment. A substantial portion of SLMs' memory footprint stems from vocabulary-related components, particularly embeddings and language modeling (LM) heads, due to large vocabulary sizes. Existing static vocabulary pruning, while reducing memory usage, suffers from rigid, one-size-fits-all designs that cause information loss from the prefill stage and a lack of flexibility. In this work, we identify two key principles underlying the vocabulary reduction challenge: the lexical locality principle, the observation that only a small subset of tokens is required during any single inference, and the asymmetry in computational characteristics between vocabulary-related components of SLM. Based on these insights, we introduce VocabTailor, a novel decoupled dynamic vocabulary selection framework that addresses memory constraints through offloading embedding and implements a hybrid static-dynamic vocabulary selection strategy for LM Head, enabling on-demand loading of vocabulary components. Comprehensive experiments across diverse downstream tasks demonstrate that VocabTailor achieves a reduction of up to 99% in the memory usage of vocabulary-related components with minimal or no degradation in task performance, substantially outperforming existing static vocabulary pruning.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:32:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Hanling",
""
],
[
"Zhou",
"Yayu",
""
],
[
"Fang",
"Tongcheng",
""
],
[
"Yuan",
"Zhihang",
""
],
[
"Dai",
"Guohao",
""
],
[
"Wang",
"Yu",
""
]
] |
2508.15230
|
Zhang-Qi Yin Dr
|
Da Zhang, Xin Li, Yibin Guo, Haifeng Yu, Yirong Jin, and Zhang-Qi Yin
|
Robust and Efficient Quantum Reservoir Computing with Discrete Time Crystal
|
12 pages, 7 figures
| null | null | null |
quant-ph cs.AI cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
The rapid development of machine learning and quantum computing has placed quantum machine learning at the forefront of research. However, existing quantum machine learning algorithms based on quantum variational algorithms face challenges in trainability and noise robustness. In order to address these challenges, we introduce a gradient-free, noise-robust quantum reservoir computing algorithm that harnesses discrete time crystal dynamics as a reservoir. We first calibrate the memory, nonlinear, and information scrambling capacities of the quantum reservoir, revealing their correlation with dynamical phases and non-equilibrium phase transitions. We then apply the algorithm to the binary classification task and establish a comparative quantum kernel advantage. For ten-class classification, both noisy simulations and experimental results on superconducting quantum processors match ideal simulations, demonstrating the enhanced accuracy with increasing system size and confirming the topological noise robustness. Our work presents the first experimental demonstration of quantum reservoir computing for image classification based on digital quantum simulation. It establishes the correlation between quantum many-body non-equilibrium phase transitions and quantum machine learning performance, providing new design principles for quantum reservoir computing and broader quantum machine learning algorithms in the NISQ era.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:40:46 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Da",
""
],
[
"Li",
"Xin",
""
],
[
"Guo",
"Yibin",
""
],
[
"Yu",
"Haifeng",
""
],
[
"Jin",
"Yirong",
""
],
[
"Yin",
"Zhang-Qi",
""
]
] |
2508.15231
|
Shihao Dong
|
Shihao Dong, Xiaotong Zhou, Yuhui Zheng, Huiying Xu, Xinzhong Zhu
|
Center-Oriented Prototype Contrastive Clustering
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Contrastive learning is widely used in clustering tasks due to its discriminative representation. However, the conflict problem between classes is difficult to solve effectively. Existing methods try to solve this problem through prototype contrast, but there is a deviation between the calculation of hard prototypes and the true cluster center. To address this problem, we propose a center-oriented prototype contrastive clustering framework, which consists of a soft prototype contrastive module and a dual consistency learning module. In short, the soft prototype contrastive module uses the probability that the sample belongs to the cluster center as a weight to calculate the prototype of each category, while avoiding inter-class conflicts and reducing prototype drift. The dual consistency learning module aligns different transformations of the same sample and the neighborhoods of different samples respectively, ensuring that the features have transformation-invariant semantic information and compact intra-cluster distribution, while providing reliable guarantees for the calculation of prototypes. Extensive experiments on five datasets show that the proposed method is effective compared to the SOTA. Our code is published on https://github.com/LouisDong95/CPCC.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:41:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Dong",
"Shihao",
""
],
[
"Zhou",
"Xiaotong",
""
],
[
"Zheng",
"Yuhui",
""
],
[
"Xu",
"Huiying",
""
],
[
"Zhu",
"Xinzhong",
""
]
] |
2508.15232
|
Jinyu Chen
|
Ruipu Wu, Yige Zhang, Jinyu Chen, Linjiang Huang, Shifeng Zhang, Xu Zhou, Liang Wang, and Si Liu
|
AeroDuo: Aerial Duo for UAV-based Vision and Language Navigation
|
Accepted by ACM MM 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Aerial Vision-and-Language Navigation (VLN) is an emerging task that enables Unmanned Aerial Vehicles (UAVs) to navigate outdoor environments using natural language instructions and visual cues. However, due to the extended trajectories and complex maneuverability of UAVs, achieving reliable UAV-VLN performance is challenging and often requires human intervention or overly detailed instructions. To harness the advantages of UAVs' high mobility, which could provide multi-grained perspectives, while maintaining a manageable motion space for learning, we introduce a novel task called Dual-Altitude UAV Collaborative VLN (DuAl-VLN). In this task, two UAVs operate at distinct altitudes: a high-altitude UAV responsible for broad environmental reasoning, and a low-altitude UAV tasked with precise navigation. To support the training and evaluation of the DuAl-VLN, we construct the HaL-13k, a dataset comprising 13,838 collaborative high-low UAV demonstration trajectories, each paired with target-oriented language instructions. This dataset includes both unseen maps and an unseen object validation set to systematically evaluate the model's generalization capabilities across novel environments and unfamiliar targets. To consolidate their complementary strengths, we propose a dual-UAV collaborative VLN framework, AeroDuo, where the high-altitude UAV integrates a multimodal large language model (Pilot-LLM) for target reasoning, while the low-altitude UAV employs a lightweight multi-stage policy for navigation and target grounding. The two UAVs work collaboratively and only exchange minimal coordinate information to ensure efficiency.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:43:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wu",
"Ruipu",
""
],
[
"Zhang",
"Yige",
""
],
[
"Chen",
"Jinyu",
""
],
[
"Huang",
"Linjiang",
""
],
[
"Zhang",
"Shifeng",
""
],
[
"Zhou",
"Xu",
""
],
[
"Wang",
"Liang",
""
],
[
"Liu",
"Si",
""
]
] |
2508.15233
|
Wenju Xu
|
Wenju Xu
|
Pretrained Diffusion Models Are Inherently Skipped-Step Samplers
| null | null | null | null |
cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Diffusion models have been achieving state-of-the-art results across various generation tasks. However, a notable drawback is their sequential generation process, requiring long-sequence step-by-step generation. Existing methods, such as DDIM, attempt to reduce sampling steps by constructing a class of non-Markovian diffusion processes that maintain the same training objective. However, there remains a gap in understanding whether the original diffusion process can achieve the same efficiency without resorting to non-Markovian processes. In this paper, we provide a confirmative answer and introduce skipped-step sampling, a mechanism that bypasses multiple intermediate denoising steps in the iterative generation process, in contrast with the traditional step-by-step refinement of standard diffusion inference. Crucially, we demonstrate that this skipped-step sampling mechanism is derived from the same training objective as the standard diffusion model, indicating that accelerated sampling via skipped-step sampling via a Markovian way is an intrinsic property of pretrained diffusion models. Additionally, we propose an enhanced generation method by integrating our accelerated sampling technique with DDIM. Extensive experiments on popular pretrained diffusion models, including the OpenAI ADM, Stable Diffusion, and Open Sora models, show that our method achieves high-quality generation with significantly reduced sampling steps.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:45:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xu",
"Wenju",
""
]
] |
2508.15234
|
E M Hasantha Ekanayake
|
E.M.H.E.B. Ekanayake, Nikhat Khan, and Nikhil Shukla
|
Bridging the Analog and the Probabilistic Computing Divide: Configuring Oscillator Ising Machines as P-bit Engines
|
14 pages, 7 figures
| null | null | null |
physics.comp-ph
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Oscillator Ising Machines (OIMs) and probabilistic bit (p-bit)-based computing platforms have emerged as promising paradigms for tackling complex combinatorial optimization problems. Although traditionally viewed as distinct approaches, this work presents a theoretically grounded framework for configuring OIMs as p-bit engines. We demonstrate that this functionality can be enabled through a novel interplay between first- and second harmonic injection to the oscillators. Our work identifies new synergies between the two methods and broadens the scope of applications for OIMs. We further show that the proposed approach can be applied to other analog dynamical systems, such as the Dynamical Ising Machine.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:45:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ekanayake",
"E. M. H. E. B.",
""
],
[
"Khan",
"Nikhat",
""
],
[
"Shukla",
"Nikhil",
""
]
] |
2508.15235
|
Hou-Jun Lu
|
Suo-Ning Wang, Hou-Jun L\"u, Xiao-Xuan Liu, Jared Rice, Jia Ren, and En-Wei Liang
|
The color evolution of magnetar-powered kilonova emission in merging neutron star-neutron star systems
|
11 pages and 4 figure. Accepted for publication in MNRAS
| null | null | null |
astro-ph.HE
|
http://creativecommons.org/publicdomain/zero/1.0/
|
The first direct detection of the gravitational wave (GW) event GW170817 and its electromagnetic (EM) counterpart open a new window for studying of multi-messenger astronomy. However, how to identify the remnant of binary neutron star (NS) merger via EM radiation remain an open question. In this paper, we propose a method of color evolution of kilonova emission to identify its progenitors. We assume that the energy of the kilonova is contributed from radioactive decay, magnetar spin-down, and pulsar wind nebula (PWN). The color evolution of kilonova emission associated with short GRB is significant when the spectrum is thermal emission, while it tends towards a constant when the spectrum is non-thermal radiation. On the other hand, if the central engine is a black hole (BH) which is promptly generated by the NS-NS merger or NS-BH merger, then the kilonova is powered only by the radioactive decay. There is no color evolution at the beginning before the peak of kilonova emission, but is significantly and rapidly increasing after the peak. On the contrary, if the central engine is a magnetar or stable NS, the kilonova emission is contributed from radioactive decay, magnetar, and PWN. The color evolution after the peak of kilonova emission is complex behavior which depends on the rotational energy and spin-down time-scale of magnetar, and finally tend to a constant in the late state.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:45:58 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Suo-Ning",
""
],
[
"Lü",
"Hou-Jun",
""
],
[
"Liu",
"Xiao-Xuan",
""
],
[
"Rice",
"Jared",
""
],
[
"Ren",
"Jia",
""
],
[
"Liang",
"En-Wei",
""
]
] |
2508.15236
|
Jiamu Wang
|
Jiamu Wang, Keunho Byeon, Jinsol Song, Anh Nguyen, Sangjeong Ahn, Sung Hak Lee, Jin Tae Kwak
|
Pathology-Informed Latent Diffusion Model for Anomaly Detection in Lymph Node Metastasis
| null | null | null | null |
eess.IV cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Anomaly detection is an emerging approach in digital pathology for its ability to efficiently and effectively utilize data for disease diagnosis. While supervised learning approaches deliver high accuracy, they rely on extensively annotated datasets, suffering from data scarcity in digital pathology. Unsupervised anomaly detection, however, offers a viable alternative by identifying deviations from normal tissue distributions without requiring exhaustive annotations. Recently, denoising diffusion probabilistic models have gained popularity in unsupervised anomaly detection, achieving promising performance in both natural and medical imaging datasets. Building on this, we incorporate a vision-language model with a diffusion model for unsupervised anomaly detection in digital pathology, utilizing histopathology prompts during reconstruction. Our approach employs a set of pathology-related keywords associated with normal tissues to guide the reconstruction process, facilitating the differentiation between normal and abnormal tissues. To evaluate the effectiveness of the proposed method, we conduct experiments on a gastric lymph node dataset from a local hospital and assess its generalization ability under domain shift using a public breast lymph node dataset. The experimental results highlight the potential of the proposed method for unsupervised anomaly detection across various organs in digital pathology. Code: https://github.com/QuIIL/AnoPILaD.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:48:55 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Jiamu",
""
],
[
"Byeon",
"Keunho",
""
],
[
"Song",
"Jinsol",
""
],
[
"Nguyen",
"Anh",
""
],
[
"Ahn",
"Sangjeong",
""
],
[
"Lee",
"Sung Hak",
""
],
[
"Kwak",
"Jin Tae",
""
]
] |
2508.15237
|
Wenyuan Li
|
Jingtang Ma, Xianglin Wu, Wenyuan Li
|
Option pricing under non-Markovian stochastic volatility models: A deep signature approach
| null | null | null | null |
q-fin.MF
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper studies the pricing problem in which the underlying asset follows a non-Markovian stochastic volatility model. Classical partial differential equation methods face significant challenges in this context, as the option prices depend not only on the current state, but also on the entire historical path of the process. To overcome these difficulties, we reformulate the asset dynamics as a rough stochastic differential equation and then represent the rough paths via linear or non-linear combinations of time-extended Brownian motion signatures. This representation transforms a rough stochastic differential equation to a classical stochastic differential equation, allowing the application of standard analytical tools. We propose a deep signature approach for both linear and nonlinear representations and rigorously prove the convergence of the algorithm. Numerical examples demonstrate the effectiveness of our approach for both Markovian and non-Markovian volatility models, offering a theoretically grounded and computationally efficient framework for option pricing.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:49:36 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ma",
"Jingtang",
""
],
[
"Wu",
"Xianglin",
""
],
[
"Li",
"Wenyuan",
""
]
] |
2508.15238
|
Shengxin Liu
|
Yinyu Liu, Kaiqiang Yu, Shengxin Liu, Cheng Long, Zhaoquan Gu
|
Temporal $k$-Core Query, Revisited
| null | null | null | null |
cs.DB
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Querying cohesive subgraphs in temporal graphs is essential for understanding the dynamic structure of real-world networks, such as evolving communities in social platforms, shifting hyperlink structures on the Web, and transient communication patterns in call networks. Recently, research has focused on the temporal $k$-core query, which aims to identify all $k$-cores across all possible time sub-intervals within a given query interval. The state-of-the-art algorithm OTCD mitigates redundant computations over overlapping sub-intervals by exploiting inclusion relationships among $k$-cores in different time intervals. Nevertheless, OTCD remains limited in scalability due to the combinatorial growth in interval enumeration and repeated processing. In this paper, we revisit the temporal $k$-core query problem and introduce a novel algorithm CoreT, which dynamically records the earliest timestamp at which each vertex or edge enters a $k$-core. This strategy enables substantial pruning of redundant computations. As a result, CoreT requires only a single pass over the query interval and achieves improved time complexity, which is linear in both the number of temporal edges within the query interval and the duration of the interval, making it highly scalable for long-term temporal analysis. Experimental results on large real-world datasets show that CoreT achieves up to four orders of magnitude speedup compared to the existing state-of-the-art OTCD, demonstrating its effectiveness and scalability for temporal $k$-core analysis.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:53:00 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Yinyu",
""
],
[
"Yu",
"Kaiqiang",
""
],
[
"Liu",
"Shengxin",
""
],
[
"Long",
"Cheng",
""
],
[
"Gu",
"Zhaoquan",
""
]
] |
2508.15239
|
Peerat Limkonchotiwat
|
Peerat Limkonchotiwat, Pume Tuchinda, Lalita Lowphansirikul, Surapon Nonesung, Panuthep Tasawong, Alham Fikri Aji, Can Udomcharoenchaikit, Sarana Nutanong
|
WangchanThaiInstruct: An instruction-following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in Thai
|
Accepted to EMNLP 2025 (Main). Model and Dataset: https://huggingface.co/collections/airesearch/wangchan-thai-instruction-6835722a30b98e01598984fd
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large language models excel at instruction-following in English, but their performance in low-resource languages like Thai remains underexplored. Existing benchmarks often rely on translations, missing cultural and domain-specific nuances needed for real-world use. We present WangchanThaiInstruct, a human-authored Thai dataset for evaluation and instruction tuning, covering four professional domains and seven task types. Created through a multi-stage quality control process with annotators, domain experts, and AI researchers, WangchanThaiInstruct supports two studies: (1) a zero-shot evaluation showing performance gaps on culturally and professionally specific tasks, and (2) an instruction tuning study with ablations isolating the effect of native supervision. Models fine-tuned on WangchanThaiInstruct outperform those using translated data in both in-domain and out-of-domain benchmarks. These findings underscore the need for culturally and professionally grounded instruction data to improve LLM alignment in low-resource, linguistically diverse settings.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 04:54:05 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Limkonchotiwat",
"Peerat",
""
],
[
"Tuchinda",
"Pume",
""
],
[
"Lowphansirikul",
"Lalita",
""
],
[
"Nonesung",
"Surapon",
""
],
[
"Tasawong",
"Panuthep",
""
],
[
"Aji",
"Alham Fikri",
""
],
[
"Udomcharoenchaikit",
"Can",
""
],
[
"Nutanong",
"Sarana",
""
]
] |
2508.15240
|
Sabab Aosaf
|
Sabab Aosaf, Muhammad Ali Nayeem, Afsana Haque, M Sohel Rahmana
|
Computational Intelligence based Land-use Allocation Approaches for Mixed Use Areas
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Urban land-use allocation represents a complex multi-objective optimization problem critical for sustainable urban development policy. This paper presents novel computational intelligence approaches for optimizing land-use allocation in mixed-use areas, addressing inherent trade-offs between land-use compatibility and economic objectives. We develop multiple optimization algorithms, including custom variants integrating differential evolution with multi-objective genetic algorithms. Key contributions include: (1) CR+DES algorithm leveraging scaled difference vectors for enhanced exploration, (2) systematic constraint relaxation strategy improving solution quality while maintaining feasibility, and (3) statistical validation using Kruskal-Wallis tests with compact letter displays. Applied to a real-world case study with 1,290 plots, CR+DES achieves 3.16\% improvement in land-use compatibility compared to state-of-the-art methods, while MSBX+MO excels in price optimization with 3.3\% improvement. Statistical analysis confirms algorithms incorporating difference vectors significantly outperform traditional approaches across multiple metrics. The constraint relaxation technique enables broader solution space exploration while maintaining practical constraints. These findings provide urban planners and policymakers with evidence-based computational tools for balancing competing objectives in land-use allocation, supporting more effective urban development policies in rapidly urbanizing regions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:00:44 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Aosaf",
"Sabab",
""
],
[
"Nayeem",
"Muhammad Ali",
""
],
[
"Haque",
"Afsana",
""
],
[
"Rahmana",
"M Sohel",
""
]
] |
2508.15241
|
Jian Guo
|
Xiaojun Chen, Jian Guo and Guan Wang
|
Differential Stochastic Variational Inequalities with Parametric Optimization
|
34 pages, 5 figures
| null | null | null |
math.OC math.DS
|
http://creativecommons.org/licenses/by/4.0/
|
The differential stochastic variational inequality with parametric convex optimization (DSVI-O) is an ordinary differential equation whose right-hand side involves a stochastic variational inequality and solutions of several dynamic and random parametric convex optimization problems. We consider that the distribution of the random variable is time-dependent and assume that the involved functions are continuous and the expectation is well-defined. We show that the DSVI-O has a weak solution with integrable and measurable solutions of the parametric optimization problems. Moreover, we propose a discrete scheme of DSVI-O by using a time-stepping approximation and the sample average approximation and prove the convergence of the discrete scheme. We illustrate our theoretical results of DSVI-O with applications in an embodied intelligence system for the elderly health by synthetic health care data generated by Multimodal Large Language Models.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:03:47 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chen",
"Xiaojun",
""
],
[
"Guo",
"Jian",
""
],
[
"Wang",
"Guan",
""
]
] |
2508.15242
|
Takenori Kataoka
|
Cornelius Greither, Takenori Kataoka
|
Determining class groups as Galois modules up to equivalence for some nonabelian extensions
|
33 pages
| null | null | null |
math.NT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In previous papers, the Galois module structure of minus class groups was studied for abelian CM extensions. In this paper, we discuss some nonabelian cases, focusing on metacyclic extensions. For a certain class of these, we obtain a complete description of the Galois modules that occur as minus class groups, modulo a certain equivalence relation on modules, which was introduced earlier by the same authors.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:07:41 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Greither",
"Cornelius",
""
],
[
"Kataoka",
"Takenori",
""
]
] |
2508.15243
|
Yixin Gao
|
Yixin Gao, Xin Li, Xiaohan Pan, Runsen Feng, Bingchen Li, Yunpeng Qi, Yiting Lu, Zhengxue Cheng, Zhibo Chen, J\"orn Ostermann
|
Comp-X: On Defining an Interactive Learned Image Compression Paradigm With Expert-driven LLM Agent
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
We present Comp-X, the first intelligently interactive image compression paradigm empowered by the impressive reasoning capability of large language model (LLM) agent. Notably, commonly used image codecs usually suffer from limited coding modes and rely on manual mode selection by engineers, making them unfriendly for unprofessional users. To overcome this, we advance the evolution of image coding paradigm by introducing three key innovations: (i) multi-functional coding framework, which unifies different coding modes of various objective/requirements, including human-machine perception, variable coding, and spatial bit allocation, into one framework. (ii) interactive coding agent, where we propose an augmented in-context learning method with coding expert feedback to teach the LLM agent how to understand the coding request, mode selection, and the use of the coding tools. (iii) IIC-bench, the first dedicated benchmark comprising diverse user requests and the corresponding annotations from coding experts, which is systematically designed for intelligently interactive image compression evaluation. Extensive experimental results demonstrate that our proposed Comp-X can understand the coding requests efficiently and achieve impressive textual interaction capability. Meanwhile, it can maintain comparable compression performance even with a single coding framework, providing a promising avenue for artificial general intelligence (AGI) in image compression.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:09:30 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gao",
"Yixin",
""
],
[
"Li",
"Xin",
""
],
[
"Pan",
"Xiaohan",
""
],
[
"Feng",
"Runsen",
""
],
[
"Li",
"Bingchen",
""
],
[
"Qi",
"Yunpeng",
""
],
[
"Lu",
"Yiting",
""
],
[
"Cheng",
"Zhengxue",
""
],
[
"Chen",
"Zhibo",
""
],
[
"Ostermann",
"Jörn",
""
]
] |
2508.15244
|
Sangmin Lee
|
Sangmin Lee, Woojin Chung, Seyun Um, and Hong-Goo Kang
|
UniCoM: A Universal Code-Switching Speech Generator
|
Accepted to EMNLP 2025 Findings
| null | null | null |
cs.CL cs.SD eess.AS
|
http://creativecommons.org/licenses/by/4.0/
|
Code-switching (CS), the alternation between two or more languages within a single speaker's utterances, is common in real-world conversations and poses significant challenges for multilingual speech technology. However, systems capable of handling this phenomenon remain underexplored, primarily due to the scarcity of suitable datasets. To resolve this issue, we propose Universal Code-Mixer (UniCoM), a novel pipeline for generating high-quality, natural CS samples without altering sentence semantics. Our approach utilizes an algorithm we call Substituting WORDs with Synonyms (SWORDS), which generates CS speech by replacing selected words with their translations while considering their parts of speech. Using UniCoM, we construct Code-Switching FLEURS (CS-FLEURS), a multilingual CS corpus designed for automatic speech recognition (ASR) and speech-to-text translation (S2TT). Experimental results show that CS-FLEURS achieves high intelligibility and naturalness, performing comparably to existing datasets on both objective and subjective metrics. We expect our approach to advance CS speech technology and enable more inclusive multilingual systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:11:21 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lee",
"Sangmin",
""
],
[
"Chung",
"Woojin",
""
],
[
"Um",
"Seyun",
""
],
[
"Kang",
"Hong-Goo",
""
]
] |
2508.15245
|
Dongyan Fu
|
Dongyan Fu, Yubing Dong and S. Kumano
|
Transversity Generalized Parton Distributions of $\Delta$ with the Diquark Spectator Model
| null | null | null | null |
hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We show quark transversity generalized parton pistributions (GPDs) of $\Delta^+$ isobar by using the diquark spectator model for the first time. First, this model is tested by electric charge, magnetic-dipole and axial charge form factors, and it is used for calculating the transversity GPDs $H^{qT}_{1,3,5,7}$ of $\Delta^+$. The quark transversity distribution $h_1$ is then obtained from the transversity GPDs in the forward limit. Then, helicity-flip amplitudes are shown numerically by using relations between the helicity amplitudes and the GPDs. Finally, by taking first moments of the GPDs, tensor form factors are obtained and we predict the tensor charge. Experimentally, $N$-$\Delta$ transition GPDs are investigated in deeply virtual Compton scattering and virtual meson-production processes, and generalized distribution amplitudes, which correspond to the $s$-channel GPDs, could be investigated by the two-photon processes $\gamma^* \gamma \to \Delta \bar\Delta$ at the electron-positron colliders. Therefore, the spin-3/2 $\Delta$ GPDs could become interesting quantities experimentally in future.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:13:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Fu",
"Dongyan",
""
],
[
"Dong",
"Yubing",
""
],
[
"Kumano",
"S.",
""
]
] |
2508.15246
|
Gerg\H{o} Nemes
|
Gerg\H{o} Nemes
|
Hyperasymptotics for linear difference equations with an irregular singularity of rank one: Polynomial coefficients
|
35 pages, 1 figure
| null | null | null |
math.CA
|
http://creativecommons.org/licenses/by/4.0/
|
Hyperasymptotics is an analytical method that incorporates exponentially small contributions into asymptotic approximations, thereby expanding their domain of validity, improving accuracy, and providing deeper insight into the underlying singularity structures. It also allows for the computation of problem-specific invariants, such as Stokes multipliers, whose values are often assumed or remain unknown in other approaches. For differential equations, unlike standard asymptotic expansions, hyperasymptotic expansions determine solutions uniquely. In this paper, we extend the hyperasymptotic method to inverse factorial series solutions of certain higher-order linear difference equations and demonstrate that the resulting expansions also determine the solutions uniquely. We further indicate how the connection coefficients appearing in these expansions can be computed numerically using hyperasymptotic techniques. In addition, we give explicit remainder bounds for the inverse factorial series solutions. Our main tool is the Mellin--Borel transform. The expansions are expressed via universal hyperterminant functions, closely related to the hyperterminants familiar from integral and differential equation contexts. The results are illustrated by the Gauss hypergeometric function with a large third parameter and a third-order difference equation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:14:29 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Nemes",
"Gergő",
""
]
] |
2508.15247
|
James Melbourne
|
Andreas Malliaris, James Melbourne, Cyril Roberto, and Michael Roysdon
|
Functional Liftings of Restricted Geometric Inequalities
| null | null | null | null |
math.FA math.MG math.PR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We investigate what we term ``generalized sup-convolutions". We show that functional inequalities that enjoy an interpretation as sup-convolution inequalities can be deduced from the special case of indicator functions corresponding to a geometric inequality. As consequences we derive a Borell-Brascamp Lieb inequality for the Gaussian Brunn-Minkowski inequality and give a functional analog of the log-Brunn-Minkowski conjecture. We also show that Barthe's reverse Brascamp-Lieb inequality is equivalent to the geometric inequality it implies and give a functional formulation of a conjecture of Schneider. Though we focus on Euclidean applications, our results are general and can be directly applied in more abstract settings, like groups or even topological measure spaces without algebraic structure, we instantiate this claim with a Borell-Brascamp-Lieb type inequality for nilpotent Lie groups.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:16:59 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Malliaris",
"Andreas",
""
],
[
"Melbourne",
"James",
""
],
[
"Roberto",
"Cyril",
""
],
[
"Roysdon",
"Michael",
""
]
] |
2508.15248
|
Shunnosuke Ikeda
|
Masato Inokuma, Shunnosuke Ikeda and Yuichi Takano
|
Estimating profitable price bounds for prescriptive price optimization
| null | null | null | null |
math.OC cs.CE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Pricing of products and services, which has a significant impact on consumer demand, is one of the most important factors in maximizing business profits. Prescriptive price optimization is a prominent data-driven pricing methodology consisting of two phases: demand forecasting and price optimization. In the practice of prescriptive price optimization, the price of each item is typically set within a predetermined range defined by lower and upper bounds. Narrow price ranges can lead to missed opportunities, while wide price ranges run the risk of proposing unrealistic prices; therefore, determining profitable price bounds while maintaining the reliability of the suggested prices is a critical challenge that directly affects the effectiveness of prescriptive price optimization. We propose two methods for estimating price bounds in prescriptive price optimization so that future total revenue derived from the optimized prices will be maximized. Our first method for price bounds estimation uses the bootstrap procedure to estimate confidence intervals for optimal prices. Our second method uses the Nelder--Mead simplex method for black-box price bounds optimization that maximizes total revenue estimated through $K$-fold cross-validation. Experimental results with synthetic price--demand datasets demonstrate that our methods successfully narrowed down the price range while maintaining high revenues, particularly when the number of items was small or the demand noise level was low. Moreover, as more data accumulated, the comparative advantage of our methods further increased.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:19:26 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Inokuma",
"Masato",
""
],
[
"Ikeda",
"Shunnosuke",
""
],
[
"Takano",
"Yuichi",
""
]
] |
2508.15249
|
Alaul Islam
|
Alaul Islam, Fairouz Grioui, Raimund Dachselt, Petra Isenberg
|
Visualization on Smart Wristbands: Results from an In-situ Design Workshop with Four Scenarios
|
12 pages, 8 figures
| null | null | null |
cs.HC
|
http://creativecommons.org/licenses/by/4.0/
|
We present the results of an in-situ ideation workshop for designing data visualizations on smart wristbands that can show data around the entire wrist of a wearer. Wristbands pose interesting challenges because the visibility of different areas of the band depends on the wearer's arm posture. We focused on four usage scenarios that lead to different postures: office work, leisurely walks, cycling, and driving. As the technology for smart wristbands is not yet commercially available, we conducted a paper-based ideation exercise that showed how spatial layout and visualization design on smart wristbands may need to vary depending on the types of data items of interest and arm postures. Participants expressed a strong preference for responsive visualization designs that could adapt to the movement of wearers' arms. Supplemental material from the study is available here: https://osf.io/4hrca/.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:19:31 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Islam",
"Alaul",
""
],
[
"Grioui",
"Fairouz",
""
],
[
"Dachselt",
"Raimund",
""
],
[
"Isenberg",
"Petra",
""
]
] |
2508.15250
|
Yilin Jiang
|
Yilin Jiang, Mingzi Zhang, Sheng Jin, Zengyi Yu, Xiangjie Kong, Binghao Tu
|
EMNLP: Educator-role Moral and Normative Large Language Models Profiling
|
24pages, 12 figures, Accepted by EMNLP Main Confrence
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Simulating Professions (SP) enables Large Language Models (LLMs) to emulate professional roles. However, comprehensive psychological and ethical evaluation in these contexts remains lacking. This paper introduces EMNLP, an Educator-role Moral and Normative LLMs Profiling framework for personality profiling, moral development stage measurement, and ethical risk under soft prompt injection. EMNLP extends existing scales and constructs 88 teacher-specific moral dilemmas, enabling profession-oriented comparison with human teachers. A targeted soft prompt injection set evaluates compliance and vulnerability in teacher SP. Experiments on 12 LLMs show teacher-role LLMs exhibit more idealized and polarized personalities than human teachers, excel in abstract moral reasoning, but struggle with emotionally complex situations. Models with stronger reasoning are more vulnerable to harmful prompt injection, revealing a paradox between capability and safety. The model temperature and other hyperparameters have limited influence except in some risk behaviors. This paper presents the first benchmark to assess ethical and psychological alignment of teacher-role LLMs for educational AI. Resources are available at https://e-m-n-l-p.github.io/.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:21:37 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Jiang",
"Yilin",
""
],
[
"Zhang",
"Mingzi",
""
],
[
"Jin",
"Sheng",
""
],
[
"Yu",
"Zengyi",
""
],
[
"Kong",
"Xiangjie",
""
],
[
"Tu",
"Binghao",
""
]
] |
2508.15251
|
Aqib Mir
|
Aqib Nazir Mir, Danish Raza Rizvi
|
Explainable Knowledge Distillation for Efficient Medical Image Classification
| null | null | null | null |
eess.IV cs.AI cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
This study comprehensively explores knowledge distillation frameworks for COVID-19 and lung cancer classification using chest X-ray (CXR) images. We employ high-capacity teacher models, including VGG19 and lightweight Vision Transformers (Visformer-S and AutoFormer-V2-T), to guide the training of a compact, hardware-aware student model derived from the OFA-595 supernet. Our approach leverages hybrid supervision, combining ground-truth labels with teacher models' soft targets to balance accuracy and computational efficiency. We validate our models on two benchmark datasets: COVID-QU-Ex and LCS25000, covering multiple classes, including COVID-19, healthy, non-COVID pneumonia, lung, and colon cancer. To interpret the spatial focus of the models, we employ Score-CAM-based visualizations, which provide insight into the reasoning process of both teacher and student networks. The results demonstrate that the distilled student model maintains high classification performance with significantly reduced parameters and inference time, making it an optimal choice in resource-constrained clinical environments. Our work underscores the importance of combining model efficiency with explainability for practical, trustworthy medical AI solutions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:22:47 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mir",
"Aqib Nazir",
""
],
[
"Rizvi",
"Danish Raza",
""
]
] |
2508.15252
|
Shiyi Yang
|
Shiyi Yang, Xinshu Li, Guanglin Zhou, Chen Wang, Xiwei Xu, Liming Zhu, Lina Yao
|
Retrieval-Augmented Review Generation for Poisoning Recommender Systems
| null | null | null | null |
cs.CR cs.CL cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent studies have shown that recommender systems (RSs) are highly vulnerable to data poisoning attacks, where malicious actors inject fake user profiles, including a group of well-designed fake ratings, to manipulate recommendations. Due to security and privacy constraints in practice, attackers typically possess limited knowledge of the victim system and thus need to craft profiles that have transferability across black-box RSs. To maximize the attack impact, the profiles often remains imperceptible. However, generating such high-quality profiles with the restricted resources is challenging. Some works suggest incorporating fake textual reviews to strengthen the profiles; yet, the poor quality of the reviews largely undermines the attack effectiveness and imperceptibility under the practical setting.
To tackle the above challenges, in this paper, we propose to enhance the quality of the review text by harnessing in-context learning (ICL) capabilities of multimodal foundation models. To this end, we introduce a demonstration retrieval algorithm and a text style transfer strategy to augment the navie ICL. Specifically, we propose a novel practical attack framework named RAGAN to generate high-quality fake user profiles, which can gain insights into the robustness of RSs. The profiles are generated by a jailbreaker and collaboratively optimized on an instructional agent and a guardian to improve the attack transferability and imperceptibility. Comprehensive experiments on various real-world datasets demonstrate that RAGAN achieves the state-of-the-art poisoning attack performance.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:25:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yang",
"Shiyi",
""
],
[
"Li",
"Xinshu",
""
],
[
"Zhou",
"Guanglin",
""
],
[
"Wang",
"Chen",
""
],
[
"Xu",
"Xiwei",
""
],
[
"Zhu",
"Liming",
""
],
[
"Yao",
"Lina",
""
]
] |
2508.15253
|
Eunseong Choi
|
Eunseong Choi, June Park, Hyeri Lee, Jongwuk Lee
|
Conflict-Aware Soft Prompting for Retrieval-Augmented Generation
|
Accepted to EMNLP 2025; 14 pages; 5 figures, 11 tables
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Retrieval-augmented generation (RAG) enhances the capabilities of large language models (LLMs) by incorporating external knowledge into their input prompts. However, when the retrieved context contradicts the LLM's parametric knowledge, it often fails to resolve the conflict between incorrect external context and correct parametric knowledge, known as context-memory conflict. To tackle this problem, we introduce Conflict-Aware REtrieval-Augmented Generation (CARE), consisting of a context assessor and a base LLM. The context assessor encodes compact memory token embeddings from raw context tokens. Through grounded/adversarial soft prompting, the context assessor is trained to discern unreliable context and capture a guidance signal that directs reasoning toward the more reliable knowledge source. Extensive experiments show that CARE effectively mitigates context-memory conflicts, leading to an average performance gain of 5.0\% on QA and fact-checking benchmarks, establishing a promising direction for trustworthy and adaptive RAG systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:36:29 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Choi",
"Eunseong",
""
],
[
"Park",
"June",
""
],
[
"Lee",
"Hyeri",
""
],
[
"Lee",
"Jongwuk",
""
]
] |
2508.15254
|
Nathan Henry PhD
|
Nathan I. N. Henry, Christopher M. Thompson, Steven Marsh, Jack D. Aylward
|
Dosimetric Evaluation of MapCHECK in MapPHAN for Conformal Arc SABR Quality Assurance
| null | null | null | null |
physics.med-ph
|
http://creativecommons.org/licenses/by/4.0/
|
This study investigates the sensitivity and specificity of the MapCHECK phantom housed in MapPHAN to errors in conformal arc Stereotactic Ablative Body Radiotherapy (SABR). A specific focus is applied to Organ At Risk (OAR) dosimetry. Multi-Leaf Collimator (MLC) class shift errors up to 2 mm, and isocentre shift errors of up to 1 mm, were introduced to 23 simulated 6 MV X-ray conformal arc lung SABR plans in Raystation. Overall, 198 plans were delivered on a Varian Clinac iX linear accelerator to an in-house modified MapCHECK in MapPHAN. Gamma analysis was used to compare the MapCHECK measurements to the simulated Raystation plans. Receiver Operating Characteristic (ROC) curves were generated from these results to determine the sensitivity and specificity of the measurement technique to the introduced errors. For MapCHECK in MapPHAN, a combination of 5%/1 mm with 95% gamma tolerance and 2%/1 mm with 90% tolerance provides good sensitivity and specificity for Quality Assurance (QA) of conformal arc SABR plans.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:36:46 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Henry",
"Nathan I. N.",
""
],
[
"Thompson",
"Christopher M.",
""
],
[
"Marsh",
"Steven",
""
],
[
"Aylward",
"Jack D.",
""
]
] |
2508.15255
|
Chun-Hung Liu
|
Rishi Balaji, Victoria Khazhinsky, Chun-Hung Liu, Kevin Qin
|
Odd list-coloring of graphs of small Euler genus with no short cycles of specific types
| null | null | null | null |
math.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Odd coloring is a variant of proper coloring and has received wide attention. We study the list-coloring version of this notion in this paper. We prove that if $G$ is a graph embeddable in the torus or the Klein bottle with no cycle of length 3, 4, and 6 such that no 5-cycles share an edge, then for every function $L$ that assigns each vertex of $G$ a set $L(v)$ of size 5, there exists a proper coloring that assigns each vertex $v$ of $G$ an element of $L(v)$ such that for every non-isolated vertex, some color appears an odd number of times on its neighborhood. In particular, every graph embeddable in the torus or the Klein bottle with no cycle of length 3, 4, 6, and 8 is odd 5-choosable. The number of colors in these results are optimal, and there exist graphs embeddable in those surfaces of girth 6 requiring seven colors.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:37:36 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Balaji",
"Rishi",
""
],
[
"Khazhinsky",
"Victoria",
""
],
[
"Liu",
"Chun-Hung",
""
],
[
"Qin",
"Kevin",
""
]
] |
2508.15256
|
Jinsol Song
|
Jinsol Song, Jiamu Wang, Anh Tien Nguyen, Keunho Byeon, Sangjeong Ahn, Sung Hak Lee, Jin Tae Kwak
|
Normal and Abnormal Pathology Knowledge-Augmented Vision-Language Model for Anomaly Detection in Pathology Images
|
Accepted at ICCV 2025. \c{opyright} IEEE 2025. This is the author's accepted version (camera-ready) of the paper. The definitive version is published in the Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV 2025). DOI will be updated when available
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Anomaly detection in computational pathology aims to identify rare and scarce anomalies where disease-related data are often limited or missing. Existing anomaly detection methods, primarily designed for industrial settings, face limitations in pathology due to computational constraints, diverse tissue structures, and lack of interpretability. To address these challenges, we propose Ano-NAViLa, a Normal and Abnormal pathology knowledge-augmented Vision-Language model for Anomaly detection in pathology images. Ano-NAViLa is built on a pre-trained vision-language model with a lightweight trainable MLP. By incorporating both normal and abnormal pathology knowledge, Ano-NAViLa enhances accuracy and robustness to variability in pathology images and provides interpretability through image-text associations. Evaluated on two lymph node datasets from different organs, Ano-NAViLa achieves the state-of-the-art performance in anomaly detection and localization, outperforming competing models.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:40:23 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Song",
"Jinsol",
""
],
[
"Wang",
"Jiamu",
""
],
[
"Nguyen",
"Anh Tien",
""
],
[
"Byeon",
"Keunho",
""
],
[
"Ahn",
"Sangjeong",
""
],
[
"Lee",
"Sung Hak",
""
],
[
"Kwak",
"Jin Tae",
""
]
] |
2508.15257
|
Eduard Elias Bahingayi
|
Eduard E. Bahingayi, Shuying Lin, Murat Uysal, Marco Di Renzo, Le-Nam Tran
|
A Refined Alternating Optimization for Sum Rate Maximization in SIM-Aided Multiuser MISO Systems
| null | null | null | null |
eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Stacked intelligent metasurfaces (SIMs) have emerged as a disruptive technology for future wireless networks. To investigate their capabilities, we study the sum rate maximization problem in an SIM-based multiuser (MU) multiple-input single-output (MISO) downlink system. A vast majority of pioneer studies, if not all, address this fundamental problem using the prevailing alternating optimization (AO) framework, where the digital beamforming (DB) and SIM phase shifts are optimized alternately. However, many of these approaches suffer from suboptimal performance, quickly leading to performance saturation, when the number of SIM layers increases assuming the \emph{fixed SIM thickness}. In this letter, we demonstrate that significant performance gains can still be achieved, and such saturation does not occur with the proposed method in the considered setting. To this end, we provide practical design guidelines to improve AO-based optimization of digital precoders and SIM phase shifts. Specifically, we show that (i) optimizing the SIM phase shifts first yields significant performance improvements, compared to optimizing the DB first; and (ii) when applying projected gradient (PG) methods, which are gradually becoming more popular to optimize the phase shifts thanks to their scalability, we find that using an iterative PG method achieves better performance than the single PG step, which is commonly used in existing solutions. Based on these customizations, the proposed method achieves a higher achievable sum rate (ASR) of up to $\ensuremath{115.53\%}$, compared to benchmark schemes for the scenarios under consideration.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:40:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bahingayi",
"Eduard E.",
""
],
[
"Lin",
"Shuying",
""
],
[
"Uysal",
"Murat",
""
],
[
"Di Renzo",
"Marco",
""
],
[
"Tran",
"Le-Nam",
""
]
] |
2508.15258
|
Dooyoung Kim
|
Dooyoung Kim, Woontack Woo
|
Spatio-Temporal Mixed and Augmented Reality Experience Description for Interactive Playback
|
4 pages, 2 figures, Accepted in the IEEE ISMAR 2025 XRStand Workshop
| null | null | null |
cs.HC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose the Spatio-Temporal Mixed and Augmented Reality Experience Description (MAR-ED), a novel framework to standardize the representation of past events for interactive and adaptive playback in a user's present physical space. While current spatial media technologies have primarily focused on capturing or replaying content as static assets, often disconnected from the viewer's environment or offering limited interactivity, the means to describe an experience's underlying semantic and interactive structure remains underexplored. We propose a descriptive framework called MAR-ED based on three core primitives: 1) Event Primitives for semantic scene graph representation, 2) Keyframe Primitives for efficient and meaningful data access, and 3) Playback Primitives for user-driven adaptive interactive playback of recorded MAR experience. The proposed flowchart of the three-stage process of the proposed MAR-ED framework transforms a recorded experience into a unique adaptive MAR experience during playback, where its spatio-temporal structure dynamically conforms to a new environment and its narrative can be altered by live user input. Drawing on this framework, personal digital memories and recorded events can evolve beyond passive 2D/3D videos into immersive, spatially-integrated group experiences, opening new paradigms for training, cultural heritage, and interactive storytelling without requiring complex, per-user adaptive rendering.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:43:28 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kim",
"Dooyoung",
""
],
[
"Woo",
"Woontack",
""
]
] |
2508.15259
|
Gauri Devi
|
Gauri Devi, Satya Ranjan Nayak and B. K. Singh
|
Study of Particle Production in Au+Au Collisions at $\sqrt{s_{NN}} =$ 11.5-200 GeV Using HYDJET++ Framework
|
11 pages, regular article
| null | null | null |
hep-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Using the HYDJET++ model, we study the production of multi-strange particles($\phi$ and $\Omega$) in Au+Au collisions at $\sqrt{s_{NN}} =$ 11.5, 19.6, 27, 39, and 200 GeV as functions of transverse momentum and centrality. The model calculations employ earlier freeze-out hypersurfaces for multi-strange particle production and are compared with STAR experimental data. Results indicate that the HYDJET++ model demonstrates a better agreement with experimental data for multi-strange particles at higher energies, particularly in relation to early thermal freeze-out. Meanwhile, the default version performs more accurately for lighter, non-strange particles at the same energy scales. We also analyze identified particle and mixed particle ratios, providing insights into strangeness enhancement. The results provide additional constraints on parton energy loss models, contributing to a more precise determination of the transport properties of hot and dense QCD matter.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:48:10 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Devi",
"Gauri",
""
],
[
"Nayak",
"Satya Ranjan",
""
],
[
"Singh",
"B. K.",
""
]
] |
2508.15260
|
Jiawei Zhao
|
Yichao Fu, Xuewei Wang, Yuandong Tian, Jiawei Zhao
|
Deep Think with Confidence
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Large Language Models (LLMs) have shown great potential in reasoning tasks through test-time scaling methods like self-consistency with majority voting. However, this approach often leads to diminishing returns in accuracy and high computational overhead. To address these challenges, we introduce Deep Think with Confidence (DeepConf), a simple yet powerful method that enhances both reasoning efficiency and performance at test time. DeepConf leverages model-internal confidence signals to dynamically filter out low-quality reasoning traces during or after generation. It requires no additional model training or hyperparameter tuning and can be seamlessly integrated into existing serving frameworks. We evaluate DeepConf across a variety of reasoning tasks and the latest open-source models, including Qwen 3 and GPT-OSS series. Notably, on challenging benchmarks such as AIME 2025, DeepConf@512 achieves up to 99.9% accuracy and reduces generated tokens by up to 84.7% compared to full parallel thinking.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:48:38 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Fu",
"Yichao",
""
],
[
"Wang",
"Xuewei",
""
],
[
"Tian",
"Yuandong",
""
],
[
"Zhao",
"Jiawei",
""
]
] |
2508.15261
|
Minas Pafis
|
Minas Pafis, Natalia Tziotziou
|
On the deterministic interior body of random polytopes
| null | null | null | null |
math.MG math.PR
|
http://creativecommons.org/licenses/by/4.0/
|
Let $\{X_i\}_{i=1}^{\infty}$ be a sequence of independent copies of a random vector $X$ in $\mathbb{R}^n$. We revisit the question to determine the asymptotic shape of the random polytope $K_N={\rm conv}\{X_1,\ldots ,X_N\}$ where $N>n$. We show that for any $\beta\in (0,1)$ there exists a constant $c(\beta)>0$ such that the following holds true: If $\mu $ is a Borel probability measure on ${\mathbb R}^n$ then, for all $N\geq c(\beta)n$ we have that $K_N\supseteq T_{\beta\ln(\frac{N}{n})}(\mu)$ with probability greater than $1-\exp(-\tfrac{1}{2}N^{1-\beta}n^{\beta})$, where $T_p(\mu)$ is the convex set of all points $x\in\mathbb{R}^n$ with half-space depth greater than or equal to $e^{-p}$. Our approach does not require any additional assumptions about the measure $\mu$ and hence it generalizes and/or improves a sequence of previous results. Moreover, for the class of strongly regular measures we compare the family $\{T_p(\mu)\}_{p>0}$ to other natural families of convex bodies associated with $\mu$, such as the $L_p$-centroid bodies of $\mu$ or the level sets of the Cram\'{e}r transform of $\mu$, and use this information in order to estimate the size of a random $K_N$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:49:11 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Pafis",
"Minas",
""
],
[
"Tziotziou",
"Natalia",
""
]
] |
2508.15262
|
Lining Chen
|
Lining Chen, Qingwen Zeng, Huaming Chen
|
M-$LLM^3$REC: A Motivation-Aware User-Item Interaction Framework for Enhancing Recommendation Accuracy with LLMs
|
10pages
| null |
10.1145/3746252.3761214
| null |
cs.IR cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recommendation systems have been essential for both user experience and platform efficiency by alleviating information overload and supporting decision-making. Traditional methods, i.e., content-based filtering, collaborative filtering, and deep learning, have achieved impressive results in recommendation systems. However, the cold-start and sparse-data scenarios are still challenging to deal with. Existing solutions either generate pseudo-interaction sequence, which often introduces redundant or noisy signals, or rely heavily on semantic similarity, overlooking dynamic shifts in user motivation. To address these limitations, this paper proposes a novel recommendation framework, termed M-$LLM^3$REC, which leverages large language models for deep motivational signal extraction from limited user interactions. M-$LLM^3$REC comprises three integrated modules: the Motivation-Oriented Profile Extractor (MOPE), Motivation-Oriented Trait Encoder (MOTE), and Motivational Alignment Recommender (MAR). By emphasizing motivation-driven semantic modeling, M-$LLM^3$REC demonstrates robust, personalized, and generalizable recommendations, particularly boosting performance in cold-start situations in comparison with the state-of-the-art frameworks.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:50:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chen",
"Lining",
""
],
[
"Zeng",
"Qingwen",
""
],
[
"Chen",
"Huaming",
""
]
] |
2508.15263
|
Liu Yang
|
Liu Yang, Zhaochun Ren, Ziqi Zhao, Pengjie Ren, Zhumin Chen, Xinyi Li, Shuaiqiang Wang, Dawei Yin and Xin Xin
|
Curriculum Approximate Unlearning for Session-based Recommendation
| null | null | null | null |
cs.IR
|
http://creativecommons.org/licenses/by/4.0/
|
Approximate unlearning for session-based recommendation refers to eliminating the influence of specific training samples from the recommender without retraining of (sub-)models. Gradient ascent (GA) is a representative method to conduct approximate unlearning. However, there still exist dual challenges to apply GA for session-based recommendation. On the one hand, naive applying of GA could lead to degradation of recommendation performance. On the other hand, existing studies fail to consider the ordering of unlearning samples when simultaneously processing multiple unlearning requests, leading to sub-optimal recommendation performance and unlearning effect. To address the above challenges, we introduce CAU, a curriculum approximate unlearning framework tailored to session-based recommendation. CAU handles the unlearning task with a GA term on unlearning samples. Specifically, to address the first challenge, CAU formulates the overall optimization task as a multi-objective optimization problem, where the GA term for unlearning samples is combined with retaining terms for preserving performance. The multi-objective optimization problem is solved through seeking the Pareto-Optimal solution, which achieves effective unlearning with trivial sacrifice on recommendation performance. To tackle the second challenge, CAU adopts a curriculum-based sequence to conduct unlearning on batches of unlearning samples. The key motivation is to perform unlearning from easy samples to harder ones. To this end, CAU first introduces two metrics to measure the unlearning difficulty, including gradient unlearning difficulty and embedding unlearning difficulty. Then, two strategies, hard-sampling and soft-sampling, are proposed to select unlearning samples according to difficulty scores.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:52:28 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yang",
"Liu",
""
],
[
"Ren",
"Zhaochun",
""
],
[
"Zhao",
"Ziqi",
""
],
[
"Ren",
"Pengjie",
""
],
[
"Chen",
"Zhumin",
""
],
[
"Li",
"Xinyi",
""
],
[
"Wang",
"Shuaiqiang",
""
],
[
"Yin",
"Dawei",
""
],
[
"Xin",
"Xin",
""
]
] |
2508.15264
|
Lindsey Kuper
|
Patrick Redmond, Jonathan Castello, Jos\'e Manuel Calder\'on Trilla, Lindsey Kuper
|
Exploring the Theory and Practice of Concurrency in the Entity-Component-System Pattern
|
This is an extended version (with appendices) of the OOPSLA 2025 paper
| null | null | null |
cs.PL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Entity-Component-System (ECS) software design pattern, long used in game development, encourages a clean separation of identity (entities), data properties (components), and computational behaviors (systems). Programs written using the ECS pattern are naturally concurrent, and the pattern offers modularity, flexibility, and performance benefits that have led to a proliferation of ECS frameworks. Nevertheless, the ECS pattern is little-known and not well understood outside of a few domains. Existing explanations of the ECS pattern tend to be mired in the concrete details of particular ECS frameworks, or they explain the pattern in terms of imperfect metaphors or in terms of what it is not. We seek a rigorous understanding of the ECS pattern via the design of a formal model, Core ECS, that abstracts away the details of specific implementations to reveal the essence of software using the ECS pattern. We identify a class of Core ECS programs that behave deterministically regardless of scheduling, enabling use of the ECS pattern as a deterministic-by-construction concurrent programming model. With Core ECS as a point of comparison, we then survey several real-world ECS frameworks and find that they all leave opportunities for deterministic concurrency unexploited. Our findings point out a space for new ECS implementation techniques that better leverage such opportunities.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:55:07 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Redmond",
"Patrick",
""
],
[
"Castello",
"Jonathan",
""
],
[
"Trilla",
"José Manuel Calderón",
""
],
[
"Kuper",
"Lindsey",
""
]
] |
2508.15265
|
Yi Zhou
|
Yi Zhou, Yuhao Deng, Yu-Shi Tian, Peng Wu, Wenjie Hu, Haoxiang Wang, Ewout Steyerberg, Xiao-Hua Zhou
|
CSTEapp: An interactive R-Shiny application of the covariate-specific treatment effect curve for visualizing individualized treatment rule
| null | null | null | null |
stat.CO stat.AP
|
http://creativecommons.org/licenses/by/4.0/
|
In precision medicine, deriving the individualized treatment rule (ITR) is crucial for recommending the optimal treatment based on patients' baseline covariates. The covariate-specific treatment effect (CSTE) curve presents a graphical method to visualize an ITR within a causal inference framework. Recent advancements have enhanced the causal interpretation of the CSTE curves and provided methods for deriving simultaneous confidence bands for various study types. To facilitate the implementation of these methods and make ITR estimation more accessible, we developed CSTEapp, a web-based application built on the R Shiny framework. CSTEapp allows users to upload data and create CSTE curves through simple point and click operations, making it the first application for estimating the ITRs. CSTEapp simplifies the analytical process by providing interactive graphical user interfaces with dynamic results, enabling users to easily report optimal treatments for individual patients based on their covariates information. Currently, CSTEapp is applicable to studies with binary and time-to-event outcomes, and we continually expand its capabilities to accommodate other outcome types as new methods emerge. We demonstrate the utility of CSTEapp using real-world examples and simulation datasets. By making advanced statistical methods more accessible, CSTEapp empowers researchers and practitioners across various fields to advance precision medicine and improve patient outcomes.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 05:58:41 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhou",
"Yi",
""
],
[
"Deng",
"Yuhao",
""
],
[
"Tian",
"Yu-Shi",
""
],
[
"Wu",
"Peng",
""
],
[
"Hu",
"Wenjie",
""
],
[
"Wang",
"Haoxiang",
""
],
[
"Steyerberg",
"Ewout",
""
],
[
"Zhou",
"Xiao-Hua",
""
]
] |
2508.15266
|
Hanxin Lou
|
Han-Xin Lou, Jing-Jing Chen, Xing-Guo Ye, Zhen-Bing Tan, An-Qi Wang, Qing Yin, Xin Liao, Jing-Zhi Fang, Xing-Yu Liu, Yi-Lin He, Zhen-Tao Zhang, Chuan Li, Zhong-Ming Wei, Xiu-Mei Ma, Dapeng Yu, Zhi-Min Liao
|
Radio-Frequency Quantum Rectification in Kagome Superconductor CsV3Sb5
|
39 pages, 14 figures
| null | null | null |
cond-mat.supr-con
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Rectification of electromagnetic fields into direct current (DC) is pivotal for energy harvesting, wireless charging, and next-generation communication technologies. The superconducting diode effect, which exploits the nonreciprocal transport of dissipationless superconducting currents, offers ultra-low power consumption and high rectification ratios. Combining the superconducting diode effect with the AC Josephson effect holds promise for converting radio-frequency (rf) irradiation into a quantized DC output. However, experimental realization has been hindered by challenges in achieving the necessary symmetry breaking and fabricating high-performance Josephson junctions. Here we demonstrate the quantum rectification in kagome superconductor CsV3Sb5, which hosts emergent Josephson effects and a zero-field Josephson diode. Under rf irradiation, a DC voltage emerges without applied bias, scaling linearly with frequency as V = hf/2e, where h is Planck's constant, f is the microwave frequency, and e is the electron charge. Furthermore, the rectified voltage exhibits quantized steps with increasing rf power, consistent with Shapiro step quantization. Our work establishes CsV3Sb5 as a versatile platform for wireless quantum power supplies and charging, and underscores the intertwined order parameters as a promising pathway for precise quantum matter control.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:00:11 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lou",
"Han-Xin",
""
],
[
"Chen",
"Jing-Jing",
""
],
[
"Ye",
"Xing-Guo",
""
],
[
"Tan",
"Zhen-Bing",
""
],
[
"Wang",
"An-Qi",
""
],
[
"Yin",
"Qing",
""
],
[
"Liao",
"Xin",
""
],
[
"Fang",
"Jing-Zhi",
""
],
[
"Liu",
"Xing-Yu",
""
],
[
"He",
"Yi-Lin",
""
],
[
"Zhang",
"Zhen-Tao",
""
],
[
"Li",
"Chuan",
""
],
[
"Wei",
"Zhong-Ming",
""
],
[
"Ma",
"Xiu-Mei",
""
],
[
"Yu",
"Dapeng",
""
],
[
"Liao",
"Zhi-Min",
""
]
] |
2508.15267
|
Ruilin Zhou
|
Ruilin Zhou, Jinglei Cheng, Yuhang Gan, Junyu Liu, Chen Qian
|
Optimizing Compilation for Distributed Quantum Computing via Clustering and Annealing
| null | null | null | null |
quant-ph cs.DC
|
http://creativecommons.org/licenses/by/4.0/
|
Efficiently mapping quantum programs onto Distributed quantum computing (DQC) are challenging, particularly when considering the heterogeneous quantum processing units (QPUs) with different structures. In this paper, we present a comprehensive compilation framework that addresses these challenges with three key insights: exploiting structural patterns within quantum circuits, using clustering for initial qubit placement, and adjusting qubit mapping with annealing algorithms. Experimental results demonstrate the effectiveness of our methods and the capability to handle complex heterogeneous distributed quantum systems. Our evaluation shows that our method reduces the objective value at most 88.40\% compared to the baseline.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:00:24 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhou",
"Ruilin",
""
],
[
"Cheng",
"Jinglei",
""
],
[
"Gan",
"Yuhang",
""
],
[
"Liu",
"Junyu",
""
],
[
"Qian",
"Chen",
""
]
] |
2508.15268
|
Gaosheng Zhao
|
Gaosheng Zhao and Dong In Kim
|
Toward Autonomous Digital Populations for Communication-Sensing-Computation Ecosystem
| null | null | null | null |
cs.NI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Future communication networks are expected to achieve deep integration of communication, sensing, and computation, forming a tightly coupled and autonomously operating infrastructure system. However, current reliance on centralized control, static design, and human intervention continues to constrain the multidimensional evolution of network functions and applications, limiting adaptability and resilience in large-scale, layered, and complex environments. To address these challenges, this paper proposes a nature-inspired architectural framework that leverages digital twin technology to organize connected devices at the edge into functional digital populations, while enabling the emergence of an evolvable digital ecosystem through multi-population integration at the cloud. We believe that this framework, which combines engineering methodologies with sociotechnical insights, lays the theoretical foundation for building next-generation communication networks with dynamic coordination, distributed decision-making, continuous adaptation, and evolutionary capabilities.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:02:44 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhao",
"Gaosheng",
""
],
[
"Kim",
"Dong In",
""
]
] |
2508.15269
|
Sarveshwar Sharma
|
Deepak Gautam, Sarveshwar Sharma, Igor Kaganovich, and Bhooshan Paradkar
|
Simulation Studies of Resonant Excitation of Electron Bernstein Waves in Capacitive Discharges
| null | null | null | null |
physics.plasm-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The behavior of capacitive coupled plasma (CCP) discharges is investigated in a mildly magnetized regime, defined by the condition 1 $\leq$ $f_{ce}/f_{rf}$ $\lt$ 2, where $f_{ce}$ and $f_{rf}$ are the cyclotron and radio-frequencies (RF), respectively. This regime exhibits complex and distinctive plasma dynamics due to the interplay between RF fields and the externally applied magnetic field. Two prominent phenomena are observed in this regime. First, the plasma density profile becomes asymmetric across the discharge, deviating from the typical symmetric distribution seen in unmagnetized CCPs. Second, electron Bernstein waves (EBWs), high-frequency electrostatic waves, are excited and propagate within the bulk plasma, particularly along steep electron density gradients. As the strength of the magnetic field increases within this regime, the CCP discharge undergoes a transition from a symmetric configuration to an asymmetric one, and then returns to a symmetric profile at higher field strengths. Notably, the excitation and propagation of EBWs are strongly correlated with the presence of discharge asymmetry and localized density gradients. These waves play a significant role in energy transport and electron heating under mildly magnetized conditions. To gain deeper insight into the underlying physics, detailed numerical simulations are carried out using the particle-in-cell Monte Carlo collision (PIC-MCC) technique. These simulations capture the kinetic behavior of electrons and ions, including the collisionless effects and sheath dynamics essential to understanding the excitation of EBWs and the evolution of discharge symmetry. The study thus sheds light on the role of weak magnetic fields in shaping plasma behavior and highlights the importance of wave-particle interactions in magnetized CCPs.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:04:56 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gautam",
"Deepak",
""
],
[
"Sharma",
"Sarveshwar",
""
],
[
"Kaganovich",
"Igor",
""
],
[
"Paradkar",
"Bhooshan",
""
]
] |
2508.15270
|
Daniel Baker
|
Daniel Baker and Ue-Li Pen
|
Single Epoch Measurements of Dispersion Measure Gradients Towards PSR B0834+06
| null | null | null | null |
astro-ph.HE astro-ph.IM
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding the evolution of pulsar dispersion measures is vital to high precision timing experiments, as well as astrometric experiments to determine pulsar positions and proper motions. In this work, we present a novel approach to measuring dispersion measure gradients using pulsar scintillometry. This approach makes use of the multipath propagation through the interstellar medium to simultaneously probe dispersion measures along multiple sight lines from a single observation. Using existing data of PSR B0834+06, we are able to measure gradients of $9.7\pm0.3\times10^{-6}~\rm{pc}~\rm{cm}^{-3}~\rm{mas}^{-1}$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:05:43 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Baker",
"Daniel",
""
],
[
"Pen",
"Ue-Li",
""
]
] |
2508.15271
|
Yongtao Li
|
Yongtao Li, Hong Liu, Shengtong Zhang
|
An edge-spectral Erd\H{o}s-Stone-Simonovits theorem and its stability
|
30 pages. Any suggestions are welcome
| null | null | null |
math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
We study the extremal problem that relates the spectral radius $\lambda (G)$ of an $F$-free graph $G$ with its number of edges. Firstly, we prove that for any graph $F$ with chromatic number $\chi (F)=r+1\ge 3$, if $G$ is an $F$-free graph on $m$ edges, then $\lambda^2(G)\le {(1-\frac{1}{r} + o(1))2m}$. This provides a unified extension of both the Erd\H{o}s--Stone--Simonovits theorem and its vertex-spectral version due to Nikiforov, and confirms a conjecture proposed by Li, Liu and Feng.
We also establish the corresponding edge-spectral stability, showing that if $G$ is an $F$-free graph on $m$ edges with $\lambda^2(G)=(1- \frac{1}{r} - o(1))2m$, then $G$ differs from a complete bipartite graph by $o(m)$ edges when $r=2$, and $G$ differs from an $r$-partite Tur\'{a}n graph by $o(m)$ edges when $r\ge 3$. This extends the classical Erd\H{o}s--Simonovits stability theorem.
As an application of our method, we improve a result of Zhai, Lin and Shu by showing that if $\lambda (G)>\sqrt{m}$, then there exist two vertices in $G$ that have at least $\frac{1}{2}\sqrt{m} - O(1)$ common neighbors. This bound is the best possible as witnessed by a random construction.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:05:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Li",
"Yongtao",
""
],
[
"Liu",
"Hong",
""
],
[
"Zhang",
"Shengtong",
""
]
] |
2508.15272
|
Han Li
|
Han Li, Shaofei Huang, Longfei Xu, Yulu Gao, Beipeng Mu, Si Liu
|
RATopo: Improving Lane Topology Reasoning via Redundancy Assignment
|
Accepted by ACM MM 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Lane topology reasoning plays a critical role in autonomous driving by modeling the connections among lanes and the topological relationships between lanes and traffic elements. Most existing methods adopt a first-detect-then-reason paradigm, where topological relationships are supervised based on the one-to-one assignment results obtained during the detection stage. This supervision strategy results in suboptimal topology reasoning performance due to the limited range of valid supervision. In this paper, we propose RATopo, a Redundancy Assignment strategy for lane Topology reasoning that enables quantity-rich and geometry-diverse topology supervision. Specifically, we restructure the Transformer decoder by swapping the cross-attention and self-attention layers. This allows redundant lane predictions to be retained before suppression, enabling effective one-to-many assignment. We also instantiate multiple parallel cross-attention blocks with independent parameters, which further enhances the diversity of detected lanes. Extensive experiments on OpenLane-V2 demonstrate that our RATopo strategy is model-agnostic and can be seamlessly integrated into existing topology reasoning frameworks, consistently improving both lane-lane and lane-traffic topology performance.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:06:02 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Li",
"Han",
""
],
[
"Huang",
"Shaofei",
""
],
[
"Xu",
"Longfei",
""
],
[
"Gao",
"Yulu",
""
],
[
"Mu",
"Beipeng",
""
],
[
"Liu",
"Si",
""
]
] |
2508.15273
|
Nicola Vassena
|
Alexander Blokhuis, Peter F. Stadler, Nicola Vassena
|
Stoichiometric recipes for periodic oscillations in reaction networks
|
20 pages (Main) + 63 pages (Supplementary Information)
| null | null | null |
q-bio.MN math.DS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Oscillatory chemical reactions are functional components in a variety of biological contexts. In chemistry, the construction and identification of even rudimentary oscillators remain elusive and lack a general framework. Using parameter-rich kinetics - a methodology enabling the disentanglement of parametric dependencies from structural analysis - we investigate the stoichiometry of chemical oscillators. We introduce the concept of oscillatory cores - minimal subnetworks that guarantee the potential for oscillations in any reaction network containing them. These cores fall into two classes, depending on whether they involve positive or negative feedback. In particular, the latter class unveils a family of oscillators - yet to be synthesized - that require a minimum number of reaction steps to exhibit oscillations, a phenomenon we refer to as the principle of length. We identify several mechanisms through which catalysis promotes oscillations: (I) furnishing instability (e.g. autocatalysis), (II) lifting dependencies, (III) lowering length thresholds. Notwithstanding this mechanistic ubiquity, we show that oscillators can also be realized without employing any catalysis. Our results highlight branches of chemistry where oscillators are likely to arise by chance, suggest new strategies for their design, and point to novel classes of oscillators yet to be realized experimentally.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:06:16 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Blokhuis",
"Alexander",
""
],
[
"Stadler",
"Peter F.",
""
],
[
"Vassena",
"Nicola",
""
]
] |
2508.15274
|
Koninika Pal
|
Lekshmi R Nair and Arun Sankar and Koninika Pal
|
TComQA: Extracting Temporal Commonsense from Text
| null |
IRRAG@SIGIR 2025
| null | null |
cs.CL cs.IR
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding events necessitates grasping their temporal context, which is often not explicitly stated in natural language. For example, it is not a trivial task for a machine to infer that a museum tour may last for a few hours, but can not take months. Recent studies indicate that even advanced large language models (LLMs) struggle in generating text that require reasoning with temporal commonsense due to its infrequent explicit mention in text. Therefore, automatically mining temporal commonsense for events enables the creation of robust language models. In this work, we investigate the capacity of LLMs to extract temporal commonsense from text and evaluate multiple experimental setups to assess their effectiveness. Here, we propose a temporal commonsense extraction pipeline that leverages LLMs to automatically mine temporal commonsense and use it to construct TComQA, a dataset derived from SAMSum and RealNews corpora. TComQA has been validated through crowdsourcing and achieves over 80\% precision in extracting temporal commonsense. The model trained with TComQA also outperforms an LLM fine-tuned on existing dataset of temporal question answering task.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:07:40 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Nair",
"Lekshmi R",
""
],
[
"Sankar",
"Arun",
""
],
[
"Pal",
"Koninika",
""
]
] |
2508.15275
|
Antonio Barletta Prof.
|
A. Barletta, M. Celli, P. V. Brand\~ao
|
Mixed convection and onset of instability due to asymmetric wall heat fluxes in a porous channel
|
22 pages, 14 figures, 1 table
| null | null | null |
physics.flu-dyn math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The problem of convective instability onset in a horizontal porous channel is explored. The channel's impermeable walls are heated with asymmetric thermal conditions modelled through unequal, but uniform, wall heat fluxes. A stationary solution describing the mixed convection flow is obtained from the governing local balance equations. Then, the linear instability of this flow is analysed by formulating an eigenvalue problem with normal modes. The research specifically highlights the role of the flow rate regime, parametrised through the P\'eclet number, where the Rayleigh number and the heat flux asymmetry ratio are key to defining when instability occurs. The numerical solution of the stability eigenvalue problem is achieved by employing the shooting method. Analytical results are also obtained by employing large-wavelength asymptotic expansions. A numerical analysis is performed to discuss the neutral stability curves and the critical values of the Rayleigh number under different flow and asymmetry conditions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:08:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Barletta",
"A.",
""
],
[
"Celli",
"M.",
""
],
[
"Brandão",
"P. V.",
""
]
] |
2508.15276
|
Zhongjun Ding
|
Zhongjun Ding, Yin Lin, Tianjing Zeng
|
AmbiSQL: Interactive Ambiguity Detection and Resolution for Text-to-SQL
| null | null | null | null |
cs.DB cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Text-to-SQL systems translate natural language questions into SQL queries, providing substantial value for non-expert users. While large language models (LLMs) show promising results for this task, they remain error-prone. Query ambiguity has been recognized as a major obstacle for LLM-based Text-to-SQL systems, leading to misinterpretation of user intent and inaccurate SQL generation. We demonstrate AmbiSQL, an interactive system that automatically detects query ambiguities and guides users through intuitive multiple-choice questions to clarify their intent. Our approach introduces a fine-grained ambiguity taxonomy for identifying ambiguities that affect database element mapping and LLM reasoning, then incorporates user feedback to rewrite ambiguous questions. Evaluation on an ambiguous query dataset shows that AmbiSQL achieves 87.2% precision in ambiguity detection and improves SQL exact match accuracy by 50% when integrated with Text-to-SQL systems. Our demonstration showcases the significant performance gains and highlights the system's practical usability. Code repo and demonstration are available at: https://github.com/JustinzjDing/AmbiSQL.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:10:28 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ding",
"Zhongjun",
""
],
[
"Lin",
"Yin",
""
],
[
"Zeng",
"Tianjing",
""
]
] |
2508.15277
|
Zijian Liang
|
Ping Zhang, Kai Niu, Yiming Liu, Zijian Liang, Nan Ma, Xiaodong Xu, Wenjun Xu, Mengying Sun, Yinqiu Liu, Xiaoyun Wang, and Ruichen Zhang
|
Way to Build Native AI-driven 6G Air Interface: Principles, Roadmap, and Outlook
|
14 pages, 7 figures
| null | null | null |
cs.IT cs.AI math.IT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Artificial intelligence (AI) is expected to serve as a foundational capability across the entire lifecycle of 6G networks, spanning design, deployment, and operation. This article proposes a native AI-driven air interface architecture built around two core characteristics: compression and adaptation. On one hand, compression enables the system to understand and extract essential semantic information from the source data, focusing on task relevance rather than symbol-level accuracy. On the other hand, adaptation allows the air interface to dynamically transmit semantic information across diverse tasks, data types, and channel conditions, ensuring scalability and robustness. This article first introduces the native AI-driven air interface architecture, then discusses representative enabling methodologies, followed by a case study on semantic communication in 6G non-terrestrial networks. Finally, it presents a forward-looking discussion on the future of native AI in 6G, outlining key challenges and research opportunities.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:11:04 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Ping",
""
],
[
"Niu",
"Kai",
""
],
[
"Liu",
"Yiming",
""
],
[
"Liang",
"Zijian",
""
],
[
"Ma",
"Nan",
""
],
[
"Xu",
"Xiaodong",
""
],
[
"Xu",
"Wenjun",
""
],
[
"Sun",
"Mengying",
""
],
[
"Liu",
"Yinqiu",
""
],
[
"Wang",
"Xiaoyun",
""
],
[
"Zhang",
"Ruichen",
""
]
] |
2508.15278
|
Chang Liu
|
Chang Liu, Adam A. Miller
|
$\texttt{HostSub_GP}$: Precise Galaxy Background Subtraction in Transient Spectroscopy with Gaussian Processes
|
19 pages, 9 figures, to be submitted to a journal
| null | null | null |
astro-ph.IM
|
http://creativecommons.org/licenses/by/4.0/
|
We present a novel host galaxy subtraction technique in longslit spectroscopy for extragalactic transients. Unlike classic methods which generally estimate the background using a simple linear interpolation of local galaxy flux in the 2D spectrum, our approach leverages multi-band archival images of the host galaxies to model the background emission from the galaxy in the 2D spectrum. Such imaging is readily accessible through wide-field imaging surveys. These images encode the wavelength-dependent galaxy profile along the slit. We construct a smooth prior for the 2D galaxy profile with a Gaussian process (GP) based on these reference images, and use another GP to model the correlated deviations from the prior in the observed spectrum. This enables accurate inference of the galaxy flux blended with the transient. On synthetic longslit data of a spiral galaxy extracted from a Multi Unit Spectroscopic Explorer (MUSE) hyper-spectral cube, the GP method remains robust as long as the host galaxy is spatially resolved and consistently outperforms the classic method. We apply the method to archival Keck spectra of two real transients, SN 2019eix and AT 2019qiz, to further demonstrate how the method uniquely recovers weak spectral features amid strong galaxy contamination, enabling refined constraints on the properties of both transients. We have released the software implementation, $\texttt{HostSub_GP}$, a scalable toolkit that leverages $\texttt{JAX}$, with an MIT license.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:11:08 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Chang",
""
],
[
"Miller",
"Adam A.",
""
]
] |
2508.15279
|
Liuyang Zhang
|
Shu-Cheng Chang, Chin-Tung Wu, Liuyang Zhang, Qiuxia Zhang
|
Rigidity and Classification of Legendrian Self-Shrinkers
| null | null | null | null |
math.DG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this article, we first classify Legendrian self-shrinkers in $\mathbb{R}% ^{3}$ and $\mathbb{R}^{5}$. We then proved a Legendrian rigidity theorem, which can be regarded as an analogue of the result of Li-Wang \cite{lw}. More precisely, let $F(\Sigma)\subset\mathbb{R}^{5}$ be an orientable Legendrian self-shrinker, if $\Vert A\Vert_{g}^{2}\leq2$ and the associated Legendrian immersion $\bar{F}\subset\mathbb{R}^{4}\times\mathbb{S}^{1}$ is compact, then $\bar{F}$ must be a flat minimal generalized Legendrian Clifford torus in $\mathbb{S}^{5}$, whose cone $\mathcal{C}(\bar{F}(\Sigma))$ is the Harvey-Lawson special Lagrangian cone in $\mathbb{C}^{3}$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:11:53 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chang",
"Shu-Cheng",
""
],
[
"Wu",
"Chin-Tung",
""
],
[
"Zhang",
"Liuyang",
""
],
[
"Zhang",
"Qiuxia",
""
]
] |
2508.15280
|
Dong Liu
|
Yuanchen Zhao, Li Rao, Dong E. Liu
|
Non-Commutative weak measurements: Entanglement, Symmetry Breaking, and the Role of Readout
|
38 pages, 4 figures
| null | null | null |
quant-ph cond-mat.dis-nn cond-mat.stat-mech
|
http://creativecommons.org/licenses/by/4.0/
|
The preparation of long-range entangled (LRE) states via quantum measurements is a promising strategy, yet its stability against realistic, non-commuting measurement noise remains a critical open question. Here, we systematically investigate the rich phase structure emerging from a minimal model of competing, non-commuting weak measurements: nearest-neighbor Ising ($Z_iZ_j$) and single-qubit transverse ($X_i$) operators. We analyze three experimentally relevant scenarios based on which measurement outcomes are read out: complete readout, no readout, and partial readout. Using a replica mean-field theory for higher dimensions, complemented by numerical simulations in one dimension, we derive the complete finite-time and stationary phase diagrams. Our analysis reveals a striking dependence on the readout protocol. Complete readout yields a direct transition between a short-range entangled (SRE) phase and a pure LRE phase. No readout (pure decoherence) precludes entanglement but exhibits a strong-to-weak spontaneous symmetry breaking (SWSSB) transition into a classically ordered mixed state. Most intriguingly, partial readout interpolates between these limits, featuring a mixed-state phase transition where the system can become trapped in the SWSSB phase or, for weaker non-commutativity, undergo successive symmetry breaking to reach a mixed LRE phase. A novel technical contribution is the use of a channel-fidelity-based partition function that allows us to simultaneously characterize both entanglement and SWSSB order, revealing a deep interplay between them in the replica limit. These results provide a cohesive picture for understanding measurement phase transitions, SWSSB, and mixed-state phase transitions, offering crucial insights for designing robust state preparation protocols on noisy quantum devices.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:12:15 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhao",
"Yuanchen",
""
],
[
"Rao",
"Li",
""
],
[
"Liu",
"Dong E.",
""
]
] |
2508.15281
|
Yi Xu
|
Yi Xu, Moyu Zhang, Chenxuan Li, Zhihao Liao, Haibo Xing, Hao Deng, Jinxin Hu, Yu Zhang, Xiaoyi Zeng, Jing Zhang
|
MMQ: Multimodal Mixture-of-Quantization Tokenization for Semantic ID Generation and User Behavioral Adaptation
| null | null | null | null |
cs.IR cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Recommender systems traditionally represent items using unique identifiers (ItemIDs), but this approach struggles with large, dynamic item corpora and sparse long-tail data, limiting scalability and generalization. Semantic IDs, derived from multimodal content such as text and images, offer a promising alternative by mapping items into a shared semantic space, enabling knowledge transfer and improving recommendations for new or rare items. However, existing methods face two key challenges: (1) balancing cross-modal synergy with modality-specific uniqueness, and (2) bridging the semantic-behavioral gap, where semantic representations may misalign with actual user preferences. To address these challenges, we propose Multimodal Mixture-of-Quantization (MMQ), a two-stage framework that trains a novel multimodal tokenizer. First, a shared-specific tokenizer leverages a multi-expert architecture with modality-specific and modality-shared experts, using orthogonal regularization to capture comprehensive multimodal information. Second, behavior-aware fine-tuning dynamically adapts semantic IDs to downstream recommendation objectives while preserving modality information through a multimodal reconstruction loss. Extensive offline experiments and online A/B tests demonstrate that MMQ effectively unifies multimodal synergy, specificity, and behavioral adaptation, providing a scalable and versatile solution for both generative retrieval and discriminative ranking tasks.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:15:49 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xu",
"Yi",
""
],
[
"Zhang",
"Moyu",
""
],
[
"Li",
"Chenxuan",
""
],
[
"Liao",
"Zhihao",
""
],
[
"Xing",
"Haibo",
""
],
[
"Deng",
"Hao",
""
],
[
"Hu",
"Jinxin",
""
],
[
"Zhang",
"Yu",
""
],
[
"Zeng",
"Xiaoyi",
""
],
[
"Zhang",
"Jing",
""
]
] |
2508.15282
|
Shivam Dubey
|
Saurabh Verma, Ekta Agrawal, and Shivam Dubey
|
Some results on Lower Assouad and quantization dimensions
| null | null | null | null |
math.DS
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we first show that the collection of all subsets of \( \mathbb{R} \) having lower dimension \( \gamma \in [0,1] \) is dense in \( \Pi(\mathbb{R}) \), the space of compact subsets of \( \mathbb{R} \). Furthermore, we show that the set of Borel probability measures with lower dimension \( \beta \in [0, m] \) is dense in \( \Omega(\mathbb{R}^m) \), the space of Borel probability measures on \( \mathbb{R}^m \). We also prove that the quantization and the lower dimension of a measure \( \vartheta \) coincide with those of the convolution of \( \vartheta \) with a finite combination of Dirac measures. In the end, we compute the lower dimension of the invariant measure associated with the product IFS.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:15:58 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Verma",
"Saurabh",
""
],
[
"Agrawal",
"Ekta",
""
],
[
"Dubey",
"Shivam",
""
]
] |
2508.15283
|
Amin Bigdeli
|
Amin Bigdeli, Negar Arabzadeh, Ebrahim Bagheri, Charles L. A. Clarke
|
Adversarial Attacks against Neural Ranking Models via In-Context Learning
| null | null | null | null |
cs.IR cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
While neural ranking models (NRMs) have shown high effectiveness, they remain susceptible to adversarial manipulation. In this work, we introduce Few-Shot Adversarial Prompting (FSAP), a novel black-box attack framework that leverages the in-context learning capabilities of Large Language Models (LLMs) to generate high-ranking adversarial documents. Unlike previous approaches that rely on token-level perturbations or manual rewriting of existing documents, FSAP formulates adversarial attacks entirely through few-shot prompting, requiring no gradient access or internal model instrumentation. By conditioning the LLM on a small support set of previously observed harmful examples, FSAP synthesizes grammatically fluent and topically coherent documents that subtly embed false or misleading information and rank competitively against authentic content. We instantiate FSAP in two modes: FSAP-IntraQ, which leverages harmful examples from the same query to enhance topic fidelity, and FSAP-InterQ, which enables broader generalization by transferring adversarial patterns across unrelated queries. Our experiments on the TREC 2020 and 2021 Health Misinformation Tracks, using four diverse neural ranking models, reveal that FSAP-generated documents consistently outrank credible, factually accurate documents. Furthermore, our analysis demonstrates that these adversarial outputs exhibit strong stance alignment and low detectability, posing a realistic and scalable threat to neural retrieval systems. FSAP also effectively generalizes across both proprietary and open-source LLMs.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:19:00 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bigdeli",
"Amin",
""
],
[
"Arabzadeh",
"Negar",
""
],
[
"Bagheri",
"Ebrahim",
""
],
[
"Clarke",
"Charles L. A.",
""
]
] |
2508.15284
|
Yuefeng Nie
|
Maosen Wang, Bo Hao, Wenjie Sun, Shengjun Yan, Shengwang Sun, Hongyi Zhang, Zhengbin Gu, Yuefeng Nie
|
Electron-Hole Crossover in $\mathrm{La}_{3-x}\mathrm{Sr}_{x}\mathrm{Ni}_{2}\mathrm{O}_{7-\delta}$ Thin Films
|
32 pages, 15 figures
| null | null | null |
cond-mat.supr-con cond-mat.mtrl-sci cond-mat.str-el
|
http://creativecommons.org/licenses/by/4.0/
|
The ambient-pressure superconductivity in $\mathrm{La}_{3}\mathrm{Ni}_{2}\mathrm{O}_{7}$ thin films via compressive epitaxial strain provides a highly accessible platform for diverse characterization techniques, facilitating the studies of high-temperature superconductivity. Here, we systematically map the superconducting dome in compressively strained $\mathrm{La}_{3-x}\mathrm{Sr}_{x}\mathrm{Ni}_{2}\mathrm{O}_{7-\delta}$ thin films by simultaneously tuning Sr doping and oxygen content and reveal an electron-hole crossover associated with the maximum transition temperature (${T}_{c}$). This electron-hole crossover, marked by an anomalous sign change in the Hall coefficient (${R}_{H}$), is reminiscent of electron-doped cuprates, which may signal a Fermi surface reconstruction. Beyond the superconducting dome, a $\ln1/T$ insulating regime and a $T$-linear resistivity regime are also resolved, resembling behaviors observed in cuprates and infinite-layer nickelates. This work reveals a dome-shaped relationship between ${T}_{c}$ and ${R}_{H}$ and establishes a key framework for understanding unconventional superconductivity in nickelate systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:19:23 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Maosen",
""
],
[
"Hao",
"Bo",
""
],
[
"Sun",
"Wenjie",
""
],
[
"Yan",
"Shengjun",
""
],
[
"Sun",
"Shengwang",
""
],
[
"Zhang",
"Hongyi",
""
],
[
"Gu",
"Zhengbin",
""
],
[
"Nie",
"Yuefeng",
""
]
] |
2508.15285
|
Chunyu Zhao
|
Chunyu Zhao, Hongzhi Wang, Kaixin Zhang, Hongliang Li, Yihan Zhang, Jiawei Zhang, Kunkai Gu, Yuan Tian, Xiangdong Huang, and Jingyi Xu
|
Efficient Cloud-Edge-Device Query Execution Based on Collaborative Scan Operator
|
12 pages, 23 figures. Submitted to IEEE Transactions on ICDE
| null | null | null |
cs.DB
|
http://creativecommons.org/licenses/by/4.0/
|
In cloud-edge-device (CED) collaborative query (CQ) processing, by leveraging CED collaboration, the advantages of both cloud computing and edge resources can be fully integrated. However, it is difficult to implement collaborative operators that can flexibly switch between the cloud and the edge during query execution. Thus, in this paper, we aim to improve the query performance when the edge resources reach a bottleneck. To achieve seamless switching of query execution between the cloud and edge, we propose a CQ processing method by establishing a CED collaborative framework based on the collaborative scan operator, so that query execution can be transferred to the cloud at any time when the edge resources are saturated. Extensive experiments show that, under sufficient network download bandwidth, the CED collaborative scan operator can effectively alleviate the performance degradation of scan operators caused by high I/O load and CPU wait time at the edge. It also achieves balanced resource scheduling between the cloud and edge.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:21:04 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhao",
"Chunyu",
""
],
[
"Wang",
"Hongzhi",
""
],
[
"Zhang",
"Kaixin",
""
],
[
"Li",
"Hongliang",
""
],
[
"Zhang",
"Yihan",
""
],
[
"Zhang",
"Jiawei",
""
],
[
"Gu",
"Kunkai",
""
],
[
"Tian",
"Yuan",
""
],
[
"Huang",
"Xiangdong",
""
],
[
"Xu",
"Jingyi",
""
]
] |
2508.15286
|
Yuegang Li
|
Yuegang Li, Junjie Wang, Tailong Xiao, Ze Zheng, Jingzheng Huang, Ming He, Jianping Fan, Guihua Zeng
|
Real-time imaging through dynamic scattering media enabled by fixed optical modulations
|
5figures
| null | null | null |
physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Dynamic scattering remains a significant challenge to the practical deployment of anti-scattering imaging. Existing methods, such as transmission matrix measurements, iterative wavefront shaping, and optical phase conjugation, depend on a quasi-static assumption, requiring the object and scattering medium to remain stable during a single imaging process to enable one-to-one compensation. However, image reconstruction becomes unattainable when this assumption is violated. Here, we propose a novel imaging strategy that counteracts time-dependent scattering perturbations through a fixed modulation module. This one-to-many compensation mechanism is realized via optical diffraction neural networks (ODNNs) trained on simulated datasets. For the first time, we reveal that its feasibility stems from the optical shower-curtain effect, and its effectiveness typically within 1-2 transport mean free paths. Our approach is not only immune to speckle decorrelation but also leverages it to enhance image quality. ODNNs generalize effectively to real-world scattering medium scenarios via multi-simulated scattering media learning and reconstruct images in real time with light-speed processing. We achieve 80 Hz imaging of moving objects in dynamic scattering media with decorrelation times < 1 ms, demonstrating applicability in large-field-of-view imaging and incoherent illumination scenarios. This work lays the foundation for high-speed, intelligent optical imaging, advancing practical anti-dynamic scattering techniques.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:22:53 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Li",
"Yuegang",
""
],
[
"Wang",
"Junjie",
""
],
[
"Xiao",
"Tailong",
""
],
[
"Zheng",
"Ze",
""
],
[
"Huang",
"Jingzheng",
""
],
[
"He",
"Ming",
""
],
[
"Fan",
"Jianping",
""
],
[
"Zeng",
"Guihua",
""
]
] |
2508.15287
|
Puja Porel
|
Puja Porel
|
Statistical Investigation of B-fields in Cores and Filaments Using JCMT/SCUPOL Legacy Survey Archival Data
|
20 pages, 12 figures
| null | null | null |
astro-ph.GA astro-ph.IM astro-ph.SR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
SCUPOL, the polarimeter for SCUBA on the James Clerk Maxwell Telescope (JCMT), was used for polarization observations of 104 regions at 850 micron wavelength and 15 arcsec resolution in the mapping mode by Matthews et al. (2009). They presented the polarization values and magnetic field morphologies in these regions. In this work, we took the opportunity to use this big legacy survey data to investigate further the collective statistical properties of the measured polarization in different star-forming regions containing cores and filaments. We did not reproduce the polarization maps but used the polarization value catalogs to investigate the statistics of distributions. In some of these regions, the data from the Combined Array for Research in Millimeter-wave Astronomy (CARMA) polarization observation at 1.3 mm wavelength and 2.5 arcsec resolution were also available from the TADPOL survey (Hull et al., 2014). We used that data and compared it with JCMT/SCUPOL values. We also study how the direction of outflows appears to relate to the mean B-field direction from large scale (JCMT observation at 15 arcsec) to small scale (CARMA observation at 2.5 arcsec) for the nine core regions common in both.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:23:48 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Porel",
"Puja",
""
]
] |
2508.15288
|
Li-Ting Ma
|
Li-Ting Ma, Kuo-Chuan Pan, Meng-Ru Wu and Rodrigo Fern\'andez
|
$r$-process Heating Feedback on Disk Outflows from Neutron Star Mergers
|
17 pages, 11 figures. Submitted
| null | null | null |
astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
Neutron star mergers produce $r$-process elements, with yields that are sensitive to the kinematic and thermodynamic properties of the ejecta. These ejecta properties are potentially affected by dynamically-important feedback from $r$-process heating, which is usually not coupled to the hydrodynamics in post-merger simulations modeling the ejecta launching and expansion. The multi-messenger detection of GW170817 showed the importance of producing reliable ejecta predictions, to maximize the diagnostic potential of future events. In this paper, we develop a prescription for including $r$-process heating as a source term in the hydrodynamic equations. This prescription depends on local fluid properties and on the $Y_{e}$ history as recorded by dedicated tracer particles, which exchange information with the grid using the Cloud-in-Cell method. The method is implemented in long-term viscous hydrodynamic simulations of accretion disk outflows to investigate its feedback on ejecta properties. We find that $r$-process heating can increase the unbound disk ejecta mass by $\sim 10\%$ relative to a baseline case that only considers alpha particle recombination. Nuclear heating also enhances the radial velocity of the ejecta with $Y_e < 0.25$ by up to a factor of two, while concurrently suppressing marginally-bound convective ejecta.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:25:54 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ma",
"Li-Ting",
""
],
[
"Pan",
"Kuo-Chuan",
""
],
[
"Wu",
"Meng-Ru",
""
],
[
"Fernández",
"Rodrigo",
""
]
] |
2508.15289
|
Steven D. Bass
|
Steven D. Bass
|
Cosmology with an emergent Standard Model
|
Brief Review, to appear in Modern Physics Letters A, 15 pages
| null | null | null |
hep-ph astro-ph.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We discuss new ideas that the Standard Model might be emergent with connection to electroweak vacuum stability and related consequences for cosmology. In this scenario, the gauge symmetries and particles of the Standard Model would be ``born'' in some phase transition at a large scale about $10^{16}$ GeV with the Standard Model parameters constrained by the requirement of vacuum stability.Emergent gauge symmetries are well known in condensed matter physics. Perhaps the Standard Model might also be emergent. In this case the particles would be the stable long range excitations of degrees of freedom that operate above the scale of emergence. The dark energy scale comes out with similar size to the tiny masses of light Majorana neutrinos. The emergence scenario comes with interesting constraints on possible dark matter structure. New physics at energy scales around $10^{16}$ GeV might be explored through its effects in the neutrino mass matrix plus using high frequency gravitational waves and polarisation observables in the cosmic microwave background.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:25:59 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bass",
"Steven D.",
""
]
] |
2508.15290
|
Peiqi Yin
|
Peiqi Yin, Xiao Yan, Qihui Zhou, Hui Li, Xiaolu Li, Lin Zhang, Meiling Wang, Xin Yao, James Cheng
|
Gorgeous: Revisiting the Data Layout for Disk-Resident High-Dimensional Vector Search
|
12 pages, 19 figures
| null | null | null |
cs.DB
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Similarity-based vector search underpins many important applications, but a key challenge is processing massive vector datasets (e.g., in TBs). To reduce costs, some systems utilize SSDs as the primary data storage. They employ a proximity graph, which connects similar vectors to form a graph and is the state-of-the-art index for vector search. However, these systems are hindered by sub-optimal data layouts that fail to effectively utilize valuable memory space to reduce disk access and suffer from poor locality for accessing disk-resident data. Through extensive profiling and analysis, we found that the structure of the proximity graph index is accessed more frequently than the vectors themselves, yet existing systems do not distinguish between the two. To address this problem, we design the Gorgeous system with the principle of prioritizing graph structure over vectors. Specifically, Gorgeous features a memory cache that keeps the adjacency lists of graph nodes to improve cache hits and a disk block format that explicitly stores neighbors' adjacency lists along with a vector to enhance data locality. Experimental results show that Gorgeous consistently outperforms two state-of-the-art disk-based systems for vector search, boosting average query throughput by over 60% and reducing query latency by over 35%.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:26:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yin",
"Peiqi",
""
],
[
"Yan",
"Xiao",
""
],
[
"Zhou",
"Qihui",
""
],
[
"Li",
"Hui",
""
],
[
"Li",
"Xiaolu",
""
],
[
"Zhang",
"Lin",
""
],
[
"Wang",
"Meiling",
""
],
[
"Yao",
"Xin",
""
],
[
"Cheng",
"James",
""
]
] |
2508.15291
|
Haji Gul
|
Haji Gul, Abul Ghani Naim, Ajaz Ahmad Bhat
|
Evaluating Knowledge Graph Complexity via Semantic, Spectral, and Structural Metrics for Link Prediction
| null | null | null | null |
cs.LG cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding dataset complexity is fundamental to evaluating and comparing link prediction models on knowledge graphs (KGs). While the Cumulative Spectral Gradient (CSG) metric, derived from probabilistic divergence between classes within a spectral clustering framework, has been proposed as a classifier agnostic complexity metric purportedly scaling with class cardinality and correlating with downstream performance, it has not been evaluated in KG settings so far. In this work, we critically examine CSG in the context of multi relational link prediction, incorporating semantic representations via transformer derived embeddings. Contrary to prior claims, we find that CSG is highly sensitive to parametrisation and does not robustly scale with the number of classes. Moreover, it exhibits weak or inconsistent correlation with standard performance metrics such as Mean Reciprocal Rank (MRR) and Hit@1. To deepen the analysis, we introduce and benchmark a set of structural and semantic KG complexity metrics. Our findings reveal that global and local relational ambiguity captured via Relation Entropy, node level Maximum Relation Diversity, and Relation Type Cardinality exhibit strong inverse correlations with MRR and Hit@1, suggesting these as more faithful indicators of task difficulty. Conversely, graph connectivity measures such as Average Degree, Degree Entropy, PageRank, and Eigenvector Centrality correlate positively with Hit@10. Our results demonstrate that CSGs purported stability and generalization predictive power fail to hold in link prediction settings and underscore the need for more stable, interpretable, and task-aligned measures of dataset complexity in knowledge driven learning.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:27:20 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gul",
"Haji",
""
],
[
"Naim",
"Abul Ghani",
""
],
[
"Bhat",
"Ajaz Ahmad",
""
]
] |
2508.15292
|
Matthew Adams
|
Matthew P. Adams, Gloria M. Monsalve-Bravo, Lucy G. Dowdell, Scott A. Sisson, Christopher Drovandi
|
Efficient sampling from a multivariate normal distribution subject to linear equality and inequality constraints
|
7 pages, 3 figures
| null | null | null |
stat.CO stat.AP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Sampling from multivariate normal distributions, subjected to a variety of restrictions, is a problem that is recurrent in statistics and computing. In the present work, we demonstrate a general framework to efficiently sample a multivariate normal distribution subject to any set of linear inequality constraints and/or linear equality constraints simultaneously. In the approach we detail, sampling a multivariate random variable from the domain formed by the intersection of linear constraints proceeds via a combination of elliptical slice sampling to address the inequality constraints, and linear mapping to address the equality constraints. We also detail a linear programming method for finding an initial sample on the linearly constrained domain; such a method is critical for sampling problems where the domain has small probability.
We demonstrate the validity of our methods on an arbitrarily chosen four-dimensional multivariate normal distribution subject to five inequality constraints and/or two equality constraints. Our approach compares favourably to direct sampling and/or accept-reject sampling methods; the latter methods vary widely in their efficiency, whereas the methods in the present work are rejection-free. Where practical we compare predictions of probability density functions between our sampling methods and analytical computation. For all simulations we demonstrate that our methods yield accurate computation of the mean and covariance of the multivariate normal distributions restricted by the imposed linear constraints. MATLAB codes to implement our methods are readily available at https://dx.doi.org/10.6084/m9.figshare.29956304 .
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:27:41 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Adams",
"Matthew P.",
""
],
[
"Monsalve-Bravo",
"Gloria M.",
""
],
[
"Dowdell",
"Lucy G.",
""
],
[
"Sisson",
"Scott A.",
""
],
[
"Drovandi",
"Christopher",
""
]
] |
2508.15293
|
Masayuki Hara
|
Masayuki Hara
|
Geometry on curves which go around a Whitney umbrella
|
12 pages, 3 figures
| null | null | null |
math.DG
|
http://creativecommons.org/licenses/by/4.0/
|
We consider curves which go around Whitney umbrella. Then we consider the geodesic and the normal curvatures, ruled surfaces generated by the normal vector and normal developable surfaces with respect to the tangent and bi-tangent vectors of the curve. Then we obtain several functions which represents geometry on Whitney umbrella. Looking at the top terms of them with respect to the radius, we study geometry on Whitney umbrella.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:29:33 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Hara",
"Masayuki",
""
]
] |
2508.15294
|
Gaoke Zhang
|
Gaoke Zhang, Bo Wang, Yunlong Ma, Dongming Zhao, Zifei Yu
|
Multiple Memory Systems for Enhancing the Long-term Memory of Agent
| null | null | null | null |
cs.AI cs.CL cs.MA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
An agent powered by large language models have achieved impressive results, but effectively handling the vast amounts of historical data generated during interactions remains a challenge. The current approach is to design a memory module for the agent to process these data. However, existing methods, such as MemoryBank and A-MEM, have poor quality of stored memory content, which affects recall performance and response quality. In order to better construct high-quality long-term memory content, we have designed a multiple memory system (MMS) inspired by cognitive psychology theory. The system processes short-term memory to multiple long-term memory fragments, and constructs retrieval memory units and contextual memory units based on these fragments, with a one-to-one correspondence between the two. During the retrieval phase, MMS will match the most relevant retrieval memory units based on the user's query. Then, the corresponding contextual memory units is obtained as the context for the response stage to enhance knowledge, thereby effectively utilizing historical data. Experiments on LoCoMo dataset compared our method with three others, proving its effectiveness. Ablation studies confirmed the rationality of our memory units. We also analyzed the robustness regarding the number of selected memory segments and the storage overhead, demonstrating its practical value.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:29:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Gaoke",
""
],
[
"Wang",
"Bo",
""
],
[
"Ma",
"Yunlong",
""
],
[
"Zhao",
"Dongming",
""
],
[
"Yu",
"Zifei",
""
]
] |
2508.15295
|
Houdun Zeng
|
Jiayin He, Houdun Zeng, Yi Zhang, Xiaoyuan Huang and Qiang Yuan
|
Contribution of Globular Clusters to Diffuse Gamma-ray Emission from Galactic Plane
|
11 pages, 6 figures, accepted for publication in ApJ
| null | null | null |
astro-ph.HE hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The diffuse Galactic $\gamma$-ray emission (DGE) provides a valuable probe for investigating the cosmic ray propagation and interactions within our Galactic environment. Recent observations have demonstrated systematic excesses of DGE compared with the conventional cosmic-ray propagation model predictions. While $\gamma$-ray emissions have been detected in a subset of globular clusters, their undetected populations may significantly contribute to the DGE. Motivated by this possibility, we present a comprehensive assessment of potential contributions from unresolved globular clusters to the DGE. In our analysis, a nonparametric method is employed to estimate the luminosity function and spatial distribution function of globular clusters using the Fermi-LAT fourth source catalog (4FGL) combined with a reference globular cluster catalog. Based on these distributions, we calculate the cumulative contribution of unresolved globular cluster populations to the DGE observed by Fermi-LAT and the Large High Altitude Air Shower Observatory (LHAASO). Our results reveal that globular clusters account for only $\sim$2\% of the DGE at the TeV range, and smaller than $1\%$ in the GeV regime, which is effectively negligible.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:31:55 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"He",
"Jiayin",
""
],
[
"Zeng",
"Houdun",
""
],
[
"Zhang",
"Yi",
""
],
[
"Huang",
"Xiaoyuan",
""
],
[
"Yuan",
"Qiang",
""
]
] |
2508.15296
|
Kei Kimura
|
Ryota Takeshima, Kei Kimura, Ayumu Kuroki, Temma Wakasugi, Makoto Yokoo
|
A New Relaxation of Fairness in Two-Sided Matching Respecting Acquaintance Relationships
|
To appear in Proceedings of ECAI-25
| null | null | null |
cs.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Two-sided matching, such as matching between students and schools, has been applied to various aspects of real life and has been the subject of much research, however, it has been plagued by the fact that efficiency and fairness are incompatible. In particular, Pareto efficiency and justified-envy-freeness are known to be incompatible even in the simplest one-to-one matching, i.e., the stable marriage problem. In previous research, the primary approach to improving efficiency in matchings has been to tolerate students' envy, thereby relaxing fairness constraints. In this study, we take a different approach to relaxing fairness. Specifically, it focuses on addressing only the envy that students may experience or prioritize more highly and seeks matchings without such envy. More specifically, this study assumes that envy towards students who are not acquaintances has less impact compared to envy towards students who are acquaintances. Accordingly, we assume that the students know each other or not, represented by an undirected graph, and define a local envy as a justified envy toward an acquaintance or a neighbor in the graph. We then propose the property that there is no local envy as a new relaxed concept of fairness, called local envy-freeness. We analyze whether Pareto-efficient matching can be achieved while maintaining local envy-freeness by meaningfully restricting the graph structure and the school's preferences. To analyze in detail the fairness that can achieve Pareto-efficient matching, we introduce a local version of the relaxed fairness recently proposed by Cho et al. (AAMAS 2024), which parameterizes the level of local envy-freeness by nonnegative integers. We then clarify the level of local envy-freeness that can be achieved by Pareto-efficient mechanisms for graphs that are ``close'' to trees and single-peaked preferences on the graphs.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:34:17 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Takeshima",
"Ryota",
""
],
[
"Kimura",
"Kei",
""
],
[
"Kuroki",
"Ayumu",
""
],
[
"Wakasugi",
"Temma",
""
],
[
"Yokoo",
"Makoto",
""
]
] |
2508.15297
|
Zhu Wang
|
Zhu Wang, Homaira Huda Shomee, Sathya N. Ravi, Sourav Medya
|
DesignCLIP: Multimodal Learning with CLIP for Design Patent Understanding
|
Accepted by EMNLP 2025. 22 pages, 14 figures
| null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
In the field of design patent analysis, traditional tasks such as patent classification and patent image retrieval heavily depend on the image data. However, patent images -- typically consisting of sketches with abstract and structural elements of an invention -- often fall short in conveying comprehensive visual context and semantic information. This inadequacy can lead to ambiguities in evaluation during prior art searches. Recent advancements in vision-language models, such as CLIP, offer promising opportunities for more reliable and accurate AI-driven patent analysis. In this work, we leverage CLIP models to develop a unified framework DesignCLIP for design patent applications with a large-scale dataset of U.S. design patents. To address the unique characteristics of patent data, DesignCLIP incorporates class-aware classification and contrastive learning, utilizing generated detailed captions for patent images and multi-views image learning. We validate the effectiveness of DesignCLIP across various downstream tasks, including patent classification and patent retrieval. Additionally, we explore multimodal patent retrieval, which provides the potential to enhance creativity and innovation in design by offering more diverse sources of inspiration. Our experiments show that DesignCLIP consistently outperforms baseline and SOTA models in the patent domain on all tasks. Our findings underscore the promise of multimodal approaches in advancing patent analysis. The codebase is available here: https://anonymous.4open.science/r/PATENTCLIP-4661/README.md.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:36:24 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Zhu",
""
],
[
"Shomee",
"Homaira Huda",
""
],
[
"Ravi",
"Sathya N.",
""
],
[
"Medya",
"Sourav",
""
]
] |
2508.15298
|
Darya Taratynova
|
Darya Taratynova, Alya Almsouti, Beknur Kalmakhanbet, Numan Saeed, Mohammad Yaqub
|
TPA: Temporal Prompt Alignment for Fetal Congenital Heart Defect Classification
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Congenital heart defect (CHD) detection in ultrasound videos is hindered by image noise and probe positioning variability. While automated methods can reduce operator dependence, current machine learning approaches often neglect temporal information, limit themselves to binary classification, and do not account for prediction calibration. We propose Temporal Prompt Alignment (TPA), a method leveraging foundation image-text model and prompt-aware contrastive learning to classify fetal CHD on cardiac ultrasound videos. TPA extracts features from each frame of video subclips using an image encoder, aggregates them with a trainable temporal extractor to capture heart motion, and aligns the video representation with class-specific text prompts via a margin-hinge contrastive loss. To enhance calibration for clinical reliability, we introduce a Conditional Variational Autoencoder Style Modulation (CVAESM) module, which learns a latent style vector to modulate embeddings and quantifies classification uncertainty. Evaluated on a private dataset for CHD detection and on a large public dataset, EchoNet-Dynamic, for systolic dysfunction, TPA achieves state-of-the-art macro F1 scores of 85.40% for CHD diagnosis, while also reducing expected calibration error by 5.38% and adaptive ECE by 6.8%. On EchoNet-Dynamic's three-class task, it boosts macro F1 by 4.73% (from 53.89% to 58.62%). Temporal Prompt Alignment (TPA) is a framework for fetal congenital heart defect (CHD) classification in ultrasound videos that integrates temporal modeling, prompt-aware contrastive learning, and uncertainty quantification.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:36:47 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Taratynova",
"Darya",
""
],
[
"Almsouti",
"Alya",
""
],
[
"Kalmakhanbet",
"Beknur",
""
],
[
"Saeed",
"Numan",
""
],
[
"Yaqub",
"Mohammad",
""
]
] |
2508.15299
|
Kentaro Yoshioka
|
Ryunosuke Hayashi, Kohei Torimi, Rokuto Nagata, Kazuma Ikeda, Ozora Sako, Taichi Nakamura, Masaki Tani, Yoshimitsu Aoki, Kentaro Yoshioka
|
BasketLiDAR: The First LiDAR-Camera Multimodal Dataset for Professional Basketball MOT
|
Accepted to MMSports
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Real-time 3D trajectory player tracking in sports plays a crucial role in tactical analysis, performance evaluation, and enhancing spectator experience. Traditional systems rely on multi-camera setups, but are constrained by the inherently two-dimensional nature of video data and the need for complex 3D reconstruction processing, making real-time analysis challenging. Basketball, in particular, represents one of the most difficult scenarios in the MOT field, as ten players move rapidly and complexly within a confined court space, with frequent occlusions caused by intense physical contact.
To address these challenges, this paper constructs BasketLiDAR, the first multimodal dataset in the sports MOT field that combines LiDAR point clouds with synchronized multi-view camera footage in a professional basketball environment, and proposes a novel MOT framework that simultaneously achieves improved tracking accuracy and reduced computational cost. The BasketLiDAR dataset contains a total of 4,445 frames and 3,105 player IDs, with fully synchronized IDs between three LiDAR sensors and three multi-view cameras. We recorded 5-on-5 and 3-on-3 game data from actual professional basketball players, providing complete 3D positional information and ID annotations for each player. Based on this dataset, we developed a novel MOT algorithm that leverages LiDAR's high-precision 3D spatial information. The proposed method consists of a real-time tracking pipeline using LiDAR alone and a multimodal tracking pipeline that fuses LiDAR and camera data. Experimental results demonstrate that our approach achieves real-time operation, which was difficult with conventional camera-only methods, while achieving superior tracking performance even under occlusion conditions. The dataset is available upon request at: https://sites.google.com/keio.jp/keio-csg/projects/basket-lidar
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:40:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Hayashi",
"Ryunosuke",
""
],
[
"Torimi",
"Kohei",
""
],
[
"Nagata",
"Rokuto",
""
],
[
"Ikeda",
"Kazuma",
""
],
[
"Sako",
"Ozora",
""
],
[
"Nakamura",
"Taichi",
""
],
[
"Tani",
"Masaki",
""
],
[
"Aoki",
"Yoshimitsu",
""
],
[
"Yoshioka",
"Kentaro",
""
]
] |
2508.15300
|
William McDonald
|
William McDonald, Cedric Le Gentil, Jennifer Wakulicz and Teresa Vidal-Calleja
|
Mag-Match: Magnetic Vector Field Features for Map Matching and Registration
|
To be published in IROS: IEEE/RSJ International Conference on Intelligent Robots and Systems, 2025
| null | null | null |
cs.RO
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Map matching and registration are essential tasks in robotics for localisation and integration of multi-session or multi-robot data. Traditional methods rely on cameras or LiDARs to capture visual or geometric information but struggle in challenging conditions like smoke or dust. Magnetometers, on the other hand, detect magnetic fields, revealing features invisible to other sensors and remaining robust in such environments. In this paper, we introduce Mag-Match, a novel method for extracting and describing features in 3D magnetic vector field maps to register different maps of the same area. Our feature descriptor, based on higher-order derivatives of magnetic field maps, is invariant to global orientation, eliminating the need for gravity-aligned mapping. To obtain these higher-order derivatives map-wide given point-wise magnetometer data, we leverage a physics-informed Gaussian Process to perform efficient and recursive probabilistic inference of both the magnetic field and its derivatives. We evaluate Mag-Match in simulated and real-world experiments against a SIFT-based approach, demonstrating accurate map-to-map, robot-to-map, and robot-to-robot transformations - even without initial gravitational alignment.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:41:41 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"McDonald",
"William",
""
],
[
"Gentil",
"Cedric Le",
""
],
[
"Wakulicz",
"Jennifer",
""
],
[
"Vidal-Calleja",
"Teresa",
""
]
] |
2508.15301
|
Huijie Qiao
|
Ying Ma and Huijie Qiao
|
Well-posedness for path-dependent multivalued McKean-Vlasov stochastic differential equations
|
25 pages
| null | null | null |
math.PR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This work concerns a type of path-dependent multivalued McKean-Vlasov stochastic differential equations. First of all, we prove the well-posedness for path-dependent multivalued stochastic differential equations under the Lipschitz conditions. Then by constructing Lipschitz approximation sequences, we generalize the result to the case of the non-Lipschitz conditions. Finally, based on the obtained results, the well-posedness for path-dependent multivalued McKean-Vlasov stochastic differential equations under the non-Lipschitz conditions is established by iterating in distributions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:46:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ma",
"Ying",
""
],
[
"Qiao",
"Huijie",
""
]
] |
2508.15302
|
Yong Zhang
|
Yi Shao, Yong Zhang, Xu-Jia Ouyang, Chuan-Peng Zhang
|
The Missing Giant: Do FAST Spectroscopic Observations Reveal a Scarcity of Large Polycyclic Aromatic Hydrocarbons in Astronomical Environments?
|
7 pages, 4 figures. Accepted for publication in MNRAS
| null | null | null |
astro-ph.GA astro-ph.SR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The search for large polycyclic aromatic hydrocarbons (PAHs) with over 100 carbon atoms is crucial to resolving the origin of unidentified infrared emission (UIE) bands. These bands are commonly observed in nebulae and the interstellar medium, yet their spectroscopic assignment has remained unknown for decades. Using the Five-hundred-meter Aperture Spherical Radio Telescope (FAST), the world's most sensitive instrument operating in the decimeter-wavelength range, we conducted a search for rotational transitions of large, quasi-symmetric PAHs. Our sample included two prototypical UIE sources, NGC 7027 and TMC-1, along with a non-UIE source, IRC+10216, for comparison. A matched filter technique was employed to isolate comb-like spectral features from quasi-symmetric PAHs containing 138 to 194 carbon atoms in the FAST spectra. This method significantly enhanced detection sensitivity to these astrophysically critical molecular signatures. Although no such features were detected, we derived upper limits on the abundance of large PAHs based on simplifying assumptions. These upper limits are lower than the values predicted by theoretical models, which might tentatively suggest that large PAHs may not be the primary carriers of UIE bands. However, this conclusion should be treated as tentative, given that it rests on simplistic assumptions which have not been empirically validated.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:47:39 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Shao",
"Yi",
""
],
[
"Zhang",
"Yong",
""
],
[
"Ouyang",
"Xu-Jia",
""
],
[
"Zhang",
"Chuan-Peng",
""
]
] |
2508.15303
|
Louie Bernhardt
|
Louie Bernhardt
|
Future stability of solutions of the Einstein-nonlinear scalar field system with decelerated expansion
|
50 pages
| null | null | null |
gr-qc math.AP math.DG
|
http://creativecommons.org/licenses/by/4.0/
|
We study solutions to the Einstein equations coupled to a nonlinear scalar field with exponential potential. This system admits Friedmann-Lema\^itre-Robertson-Walker solutions undergoing decelerated expansion, with $\mathbb{T}^3$ spatial topology and scale factor $a(t) = t^p$ for $1/3 < p < 1$. For each $p \in (2/3,1)$, we prove that the corresponding FLRW spacetime is future-stable as a solution to the Einstein-nonlinear scalar field system. Given initial data on a spacelike hypersurface that is sufficiently close to the FLRW data, we show the resulting solution is future-causal geodesically complete, and remains close to the FLRW solution for all time. Moreover, we show the perturbed metric components and scalar field converge to spatially homogeneous functions as $t \rightarrow \infty$. A key feature of our analysis is the decomposition of the metric and scalar field perturbations into their spatial averages and oscillatory remainders with zero average.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:48:16 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bernhardt",
"Louie",
""
]
] |
2508.15304
|
Yuzhuo Dang
|
Yuzhuo Dang, Xin Zhang, Zhiqiang Pan, Yuxiao Duan, Wanyu Chen, Fei Cai and Honghui Chen
|
MLLMRec: Exploring the Potential of Multimodal Large Language Models in Recommender Systems
| null | null | null | null |
cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Multimodal recommendation typically combines the user behavioral data with the modal features of items to reveal user's preference, presenting superior performance compared to the conventional recommendations. However, existing methods still suffer from two key problems: (1) the initialization methods of user multimodal representations are either behavior-unperceived or noise-contaminated, and (2) the KNN-based item-item graph contains noisy edges with low similarities and lacks audience co-occurrence relationships. To address such issues, we propose MLLMRec, a novel MLLM-driven multimodal recommendation framework with two item-item graph refinement strategies. On the one hand, the item images are first converted into high-quality semantic descriptions using an MLLM, which are then fused with the textual metadata of items. Then, we construct a behavioral description list for each user and feed it into the MLLM to reason about the purified user preference containing interaction motivations. On the other hand, we design the threshold-controlled denoising and topology-aware enhancement strategies to refine the suboptimal item-item graph, thereby enhancing the item representation learning. Extensive experiments on three publicly available datasets demonstrate that MLLMRec achieves the state-of-the-art performance with an average improvement of 38.53% over the best baselines.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:50:00 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Dang",
"Yuzhuo",
""
],
[
"Zhang",
"Xin",
""
],
[
"Pan",
"Zhiqiang",
""
],
[
"Duan",
"Yuxiao",
""
],
[
"Chen",
"Wanyu",
""
],
[
"Cai",
"Fei",
""
],
[
"Chen",
"Honghui",
""
]
] |
2508.15305
|
Wei Yang
|
Wei Yang, Jinwei Xiao, Hongming Zhang, Qingyang Zhang, Yanna Wang, Bo Xu
|
Coarse-to-Fine Grounded Memory for LLM Agent Planning
|
Accepted to EMNLP 2025 Main Conference;27 pages,15 figures
| null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Recent advancements in Large Language Models (LLMs) have driven growing interest in LLM-based agents for complex planning tasks. To avoid costly agent training, many studies adopted memory mechanism that enhances LLM with offline experiences or online trajectory analysis. However, existing works focus on single-granularity memory derived from dynamic environmental interactions, which are inherently constrained by the quality of the collected experiences. This limitation, in turn, constrain the diversity of knowledge and the flexibility of planning. We propose Coarse-to-Fine Grounded Memory (\Ours{}), a novel framework that grounds coarse-to-fine memories with LLM, thereby fully leverage them for flexible adaptation to diverse scenarios. \Ours{} grounds environmental information into coarse-grained focus points to guide experience collection in training tasks, followed by grounding of actionable hybrid-grained tips from each experience. At inference, \Ours{} retrieves task-relevant experiences and tips to support planning. When facing environmental anomalies, the LLM grounds the current situation into fine-grained key information, enabling flexible self-QA reflection and plan correction.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:50:23 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yang",
"Wei",
""
],
[
"Xiao",
"Jinwei",
""
],
[
"Zhang",
"Hongming",
""
],
[
"Zhang",
"Qingyang",
""
],
[
"Wang",
"Yanna",
""
],
[
"Xu",
"Bo",
""
]
] |
2508.15306
|
Laszlo Erdodi
|
Henrietta Hegyi and Laszlo Erdodi
|
Connected and Exposed: Cybersecurity Risks, Regulatory Gaps, and Public Perception in Internet-Connected Vehicles
| null | null | null | null |
cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
The rapid advancement of Internet-connected vehicle technologies has introduced a new era of smart mobility, while simultaneously raising significant cybersecurity and privacy concerns. This paper explores the evolving threat landscape associated with connected vehicles, focusing on risks such as unauthorized remote access and the potential leakage of personal data. To assess the current state of protection, we conducted a comprehensive analysis of 16 international standards and regulations, evaluating them from multiple perspectives including regulatory strength, technical specificity, treatment of supply chain risks, and approaches to personal data handling.
In parallel, we carried out a user-focused survey designed to map consumer attitudes toward smart cars. The survey investigated which types of vehicles users trust and prefer, the reasons behind rejecting certain car types, their awareness of data-related risks, and whether they feel adequately informed about how their vehicles handle data. By combining regulatory analysis with user perception insights, this study aims to contribute to a more holistic understanding of the challenges and expectations surrounding connected vehicle ecosystems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 06:51:54 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Hegyi",
"Henrietta",
""
],
[
"Erdodi",
"Laszlo",
""
]
] |
2508.15307
|
Xiangtong Wang
|
Xiangtong Wang, Wei Li, Menglong Yang and Songchen Han
|
Unlocking the Performance Potential of Mega-Constellation Networks: An Exploration of Structure-Building Paradigms
| null | null | null | null |
cs.NI
|
http://creativecommons.org/licenses/by/4.0/
|
The network structure design plays a vital role in the mega-constellation network (MSN) to coordinate massive network nodes to ensure the effectiveness and reliability of operations and services for future space wireless communications networks.
One of the critical issues in MCN is how to design an optimal network control structure by configuring the most stable inter-satellite link (ISL) to achieve high available MCN within a limited average transmission delays.
To address this problem, this paper introduces a novel MCN structure design paradigm: Structure = Motif + Lattice (SML), which decouples MCN design into local motifs design and global lattices design. Specifically, we formulate the High-Availability and Low-Latency Mega-Constellation Design (HALLMD) problem, aimed at maximizing ISL availability while minimizing the transmission latency. To solve HALLMD, we propose SMLOP, a heuristic algorithm that efficiently finds optimal network structures in polynomial time. Experimental validation on four public state-of-the-art constellations demonstrates significant improvements, including enhanced capacity by $5\sim 18\%$, increased throughput by $1\sim 12\%$, reduced path stretch by $12\sim 23\%$, and Round-Trip Time (RTT) by $8\sim 77\%$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:00:52 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Xiangtong",
""
],
[
"Li",
"Wei",
""
],
[
"Yang",
"Menglong",
""
],
[
"Han",
"Songchen",
""
]
] |
2508.15308
|
Haibo Xing
|
Haibo Xing, Hao Deng, Yucheng Mao, Jinxin Hu, Yi Xu, Hao Zhang, Jiahao Wang, Shizhun Wang, Yu Zhang, Xiaoyi Zeng, Jing Zhang
|
REG4Rec: Reasoning-Enhanced Generative Model for Large-Scale Recommendation Systems
| null | null | null | null |
cs.IR
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Sequential recommendation aims to predict a user's next action in large-scale recommender systems. While traditional methods often suffer from insufficient information interaction, recent generative recommendation models partially address this issue by directly generating item predictions. To better capture user intents, recent studies have introduced a reasoning process into generative recommendation, significantly improving recommendation performance. However, these approaches are constrained by the singularity of item semantic representations, facing challenges such as limited diversity in reasoning pathways and insufficient reliability in the reasoning process. To tackle these issues, we introduce REG4Rec, a reasoning-enhanced generative model that constructs multiple dynamic semantic reasoning paths alongside a self-reflection process, ensuring high-confidence recommendations. Specifically, REG4Rec utilizes an MoE-based parallel quantization codebook (MPQ) to generate multiple unordered semantic tokens for each item, thereby constructing a larger-scale diverse reasoning space. Furthermore, to enhance the reliability of reasoning, we propose a training reasoning enhancement stage, which includes Preference Alignment for Reasoning (PARS) and a Multi-Step Reward Augmentation (MSRA) strategy. PARS uses reward functions tailored for recommendation to enhance reasoning and reflection, while MSRA introduces future multi-step actions to improve overall generalization. During inference, Consistency-Oriented Self-Reflection for Pruning (CORP) is proposed to discard inconsistent reasoning paths, preventing the propagation of erroneous reasoning. Lastly, we develop an efficient offline training strategy for large-scale recommendation. Experiments on real-world datasets and online evaluations show that REG4Rec delivers outstanding performance and substantial practical value.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:02:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xing",
"Haibo",
""
],
[
"Deng",
"Hao",
""
],
[
"Mao",
"Yucheng",
""
],
[
"Hu",
"Jinxin",
""
],
[
"Xu",
"Yi",
""
],
[
"Zhang",
"Hao",
""
],
[
"Wang",
"Jiahao",
""
],
[
"Wang",
"Shizhun",
""
],
[
"Zhang",
"Yu",
""
],
[
"Zeng",
"Xiaoyi",
""
],
[
"Zhang",
"Jing",
""
]
] |
2508.15309
|
Teresa Lemainque
|
Teresa Lemainque, Philipp Schad, Simon Westfechtel, Marc von der St\"uck, Robert Siepmann, Axel Honn\'e, Christiane Kuhl, Daniel Truhn, Sven Nebelung
|
T2 Radiomic Features Are More Sensitive Than Mean T2 for Cartilage Load Response: A Stress MRI Study
|
12 pages (main manuscript text), 4 tables, 5 figures, 3 supplementary figures
| null | null | null |
physics.med-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Objective: To assess response-to-loading in a human cadaveric knee joint model under different loading conditions before and after meniscectomy
Design: In this prospective study, stress magnetic resonance imaging was performed using an MR-compatible loading device and quantitative T2 mapping in unloaded (UL), 0${\deg}$ neutrally loaded (LN) and 10${\deg}$ varus loaded (LV) condition before and after meniscectomy. Mean T2 values and four radiomic texture parameters were assessed within the cartilage of medial femur (MF) and medial tibia (MT) for all conditions.
Results: Medial joint space width decreased from UL to LN to LV and after meniscectomy (all p<0.05). T2 values did not show any significant dependency on pressure or meniscectomy (all p>0.05). The radiomic parameter variance could assess loading induced textural T2 changes in the MF (UL vs. LN: p=0.042; UL vs. LV: p<0.001; LN vs. LV: p=0.022), and, in part, in the MT (LN vs. LV: p<0.013). Meniscectomy did not significantly alter the T2 mean values or radiomic parameters, respectively.
Conclusions: T2-based radiomic features were more sensitive to assess cartilage response to loading than T2-mapping alone.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:05:43 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lemainque",
"Teresa",
""
],
[
"Schad",
"Philipp",
""
],
[
"Westfechtel",
"Simon",
""
],
[
"von der Stück",
"Marc",
""
],
[
"Siepmann",
"Robert",
""
],
[
"Honné",
"Axel",
""
],
[
"Kuhl",
"Christiane",
""
],
[
"Truhn",
"Daniel",
""
],
[
"Nebelung",
"Sven",
""
]
] |
2508.15310
|
Hengyu An
|
Hengyu An, Jinghuai Zhang, Tianyu Du, Chunyi Zhou, Qingming Li, Tao Lin, Shouling Ji
|
IPIGuard: A Novel Tool Dependency Graph-Based Defense Against Indirect Prompt Injection in LLM Agents
|
EMNLP 2025
| null | null | null |
cs.CR cs.AI cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large language model (LLM) agents are widely deployed in real-world applications, where they leverage tools to retrieve and manipulate external data for complex tasks. However, when interacting with untrusted data sources (e.g., fetching information from public websites), tool responses may contain injected instructions that covertly influence agent behaviors and lead to malicious outcomes, a threat referred to as Indirect Prompt Injection (IPI). Existing defenses typically rely on advanced prompting strategies or auxiliary detection models. While these methods have demonstrated some effectiveness, they fundamentally rely on assumptions about the model's inherent security, which lacks structural constraints on agent behaviors. As a result, agents still retain unrestricted access to tool invocations, leaving them vulnerable to stronger attack vectors that can bypass the security guardrails of the model. To prevent malicious tool invocations at the source, we propose a novel defensive task execution paradigm, called IPIGuard, which models the agents' task execution process as a traversal over a planned Tool Dependency Graph (TDG). By explicitly decoupling action planning from interaction with external data, IPIGuard significantly reduces unintended tool invocations triggered by injected instructions, thereby enhancing robustness against IPI attacks. Experiments on the AgentDojo benchmark show that IPIGuard achieves a superior balance between effectiveness and robustness, paving the way for the development of safer agentic systems in dynamic environments.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:08:16 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"An",
"Hengyu",
""
],
[
"Zhang",
"Jinghuai",
""
],
[
"Du",
"Tianyu",
""
],
[
"Zhou",
"Chunyi",
""
],
[
"Li",
"Qingming",
""
],
[
"Lin",
"Tao",
""
],
[
"Ji",
"Shouling",
""
]
] |
2508.15311
|
Weijiang Lai
|
Weijiang Lai, Beihong Jin, Yapeng Zhang, Yiyuan Zheng, Rui Zhao, Jian Dong, Jun Lei, Xingxing Wang
|
Modeling Long-term User Behaviors with Diffusion-driven Multi-interest Network for CTR Prediction
| null |
RecSys 2025
|
10.1145/3705328.3748045
| null |
cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
CTR (Click-Through Rate) prediction, crucial for recommender systems and online advertising, etc., has been confirmed to benefit from modeling long-term user behaviors. Nonetheless, the vast number of behaviors and complexity of noise interference pose challenges to prediction efficiency and effectiveness. Recent solutions have evolved from single-stage models to two-stage models. However, current two-stage models often filter out significant information, resulting in an inability to capture diverse user interests and build the complete latent space of user interests. Inspired by multi-interest and generative modeling, we propose DiffuMIN (Diffusion-driven Multi-Interest Network) to model long-term user behaviors and thoroughly explore the user interest space. Specifically, we propose a target-oriented multi-interest extraction method that begins by orthogonally decomposing the target to obtain interest channels. This is followed by modeling the relationships between interest channels and user behaviors to disentangle and extract multiple user interests. We then adopt a diffusion module guided by contextual interests and interest channels, which anchor users' personalized and target-oriented interest types, enabling the generation of augmented interests that align with the latent spaces of user interests, thereby further exploring restricted interest space. Finally, we leverage contrastive learning to ensure that the generated augmented interests align with users' genuine preferences. Extensive offline experiments are conducted on two public datasets and one industrial dataset, yielding results that demonstrate the superiority of DiffuMIN. Moreover, DiffuMIN increased CTR by 1.52% and CPM by 1.10% in online A/B testing. Our source code is available at https://github.com/laiweijiang/DiffuMIN.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:10:01 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lai",
"Weijiang",
""
],
[
"Jin",
"Beihong",
""
],
[
"Zhang",
"Yapeng",
""
],
[
"Zheng",
"Yiyuan",
""
],
[
"Zhao",
"Rui",
""
],
[
"Dong",
"Jian",
""
],
[
"Lei",
"Jun",
""
],
[
"Wang",
"Xingxing",
""
]
] |
2508.15312
|
Xiu-Xiu Zhan
|
Su-Su Zhang, JinFeng Xie, Yang Chen, Min Gao, Cong Li, Chuang Liu, Xiu-Xiu Zhan
|
HIP: Model-Agnostic Hypergraph Influence Prediction via Distance-Centrality Fusion and Neural ODEs
| null | null | null | null |
cs.SI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Predicting user influence in social networks is a critical problem, and hypergraphs, as a prevalent higher-order modeling approach, provide new perspectives for this task. However, the absence of explicit cascade or infection probability data makes it particularly challenging to infer influence in hypergraphs. To address this, we introduce HIP, a unified and model-independent framework for influence prediction without knowing the underlying spreading model. HIP fuses multi-dimensional centrality indicators with a temporally reinterpreted distance matrix to effectively represent node-level diffusion capacity in the absence of observable spreading. These representations are further processed through a multi-hop Hypergraph Neural Network (HNN) to capture complex higher-order structural dependencies, while temporal correlations are modeled using a hybrid module that combines Long Short-Term Memory (LSTM) networks and Neural Ordinary Differential Equations (Neural ODEs). Notably, HIP is inherently modular: substituting the standard HGNN with the advanced DPHGNN, and the LSTM with xLSTM, yields similarly strong performance, showcasing its architectural generality and robustness. Empirical evaluations across 14 real-world hypergraph datasets demonstrate that HIP consistently surpasses existing baselines in prediction accuracy, resilience, and identification of top influencers, all without relying on any diffusion trajectories or prior knowledge of the spreading model. These findings underline HIP's effectiveness and adaptability as a general-purpose solution for influence prediction in complex hypergraph environments.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:12:57 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Su-Su",
""
],
[
"Xie",
"JinFeng",
""
],
[
"Chen",
"Yang",
""
],
[
"Gao",
"Min",
""
],
[
"Li",
"Cong",
""
],
[
"Liu",
"Chuang",
""
],
[
"Zhan",
"Xiu-Xiu",
""
]
] |
2508.15313
|
Wutao Liu
|
Wutao Liu, YiDan Wang, and Pan Gao
|
First RAG, Second SEG: A Training-Free Paradigm for Camouflaged Object Detection
| null | null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Camouflaged object detection (COD) poses a significant challenge in computer vision due to the high similarity between objects and their backgrounds. Existing approaches often rely on heavy training and large computational resources. While foundation models such as the Segment Anything Model (SAM) offer strong generalization, they still struggle to handle COD tasks without fine-tuning and require high-quality prompts to yield good performance. However, generating such prompts manually is costly and inefficient. To address these challenges, we propose \textbf{First RAG, Second SEG (RAG-SEG)}, a training-free paradigm that decouples COD into two stages: Retrieval-Augmented Generation (RAG) for generating coarse masks as prompts, followed by SAM-based segmentation (SEG) for refinement. RAG-SEG constructs a compact retrieval database via unsupervised clustering, enabling fast and effective feature retrieval. During inference, the retrieved features produce pseudo-labels that guide precise mask generation using SAM2. Our method eliminates the need for conventional training while maintaining competitive performance. Extensive experiments on benchmark COD datasets demonstrate that RAG-SEG performs on par with or surpasses state-of-the-art methods. Notably, all experiments are conducted on a \textbf{personal laptop}, highlighting the computational efficiency and practicality of our approach. We present further analysis in the Appendix, covering limitations, salient object detection extension, and possible improvements.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:14:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Wutao",
""
],
[
"Wang",
"YiDan",
""
],
[
"Gao",
"Pan",
""
]
] |
2508.15314
|
Naen Xu
|
Naen Xu, Jinghuai Zhang, Changjiang Li, Zhi Chen, Chunyi Zhou, Qingming Li, Tianyu Du, Shouling Ji
|
VideoEraser: Concept Erasure in Text-to-Video Diffusion Models
|
To appear in the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP)
| null | null | null |
cs.CV cs.AI cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
The rapid growth of text-to-video (T2V) diffusion models has raised concerns about privacy, copyright, and safety due to their potential misuse in generating harmful or misleading content. These models are often trained on numerous datasets, including unauthorized personal identities, artistic creations, and harmful materials, which can lead to uncontrolled production and distribution of such content. To address this, we propose VideoEraser, a training-free framework that prevents T2V diffusion models from generating videos with undesirable concepts, even when explicitly prompted with those concepts. Designed as a plug-and-play module, VideoEraser can seamlessly integrate with representative T2V diffusion models via a two-stage process: Selective Prompt Embedding Adjustment (SPEA) and Adversarial-Resilient Noise Guidance (ARNG). We conduct extensive evaluations across four tasks, including object erasure, artistic style erasure, celebrity erasure, and explicit content erasure. Experimental results show that VideoEraser consistently outperforms prior methods regarding efficacy, integrity, fidelity, robustness, and generalizability. Notably, VideoEraser achieves state-of-the-art performance in suppressing undesirable content during T2V generation, reducing it by 46% on average across four tasks compared to baselines.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:15:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xu",
"Naen",
""
],
[
"Zhang",
"Jinghuai",
""
],
[
"Li",
"Changjiang",
""
],
[
"Chen",
"Zhi",
""
],
[
"Zhou",
"Chunyi",
""
],
[
"Li",
"Qingming",
""
],
[
"Du",
"Tianyu",
""
],
[
"Ji",
"Shouling",
""
]
] |
2508.15315
|
Hikaru Sasaki
|
Takahiro Nishinaka and Hikaru Sasaki
|
On bosonic vertex algebras associated with 3D reductions of Argyres-Douglas theories
|
54 pages, 5 figures, 4 tables
| null | null |
NITEP 259
|
hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We study the bosonic VOA associated with the 3D $\mathcal{N}=4$ abelian linear quiver gauge theories arising from compactifying 4D $\mathcal{N}=2$ Argyres-Douglas theories of $(A_1,A_{2n-1})$ and $(A_1,D_{2n})$ types. These VOAs are obtained by cancelling the gauge anomaly of the H-twisted 3D theory on the half-space by Heisenberg algebras on the boundary. We particularly conjecture a complete set of strong generators of these bosonic VOAs, which contains more than the Virasoro stress tensor and those arising from Higgs branch operators. We also find that these bosonic VOAs contain copies of the $W_3$ vertex algebra at $c=-2$ as sub vertex algebras.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:19:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Nishinaka",
"Takahiro",
""
],
[
"Sasaki",
"Hikaru",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.