id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.19101
|
Vaishali Dhanoa
|
Vaishali Dhanoa, Anton Wolter, Gabriela Molina Le\'on, Hans-J\"org Schulz, Niklas Elmqvist
|
Agentic Visualization: Extracting Agent-based Design Patterns from Visualization Systems
| null | null | null | null |
cs.HC
|
http://creativecommons.org/licenses/by/4.0/
|
Autonomous agents powered by Large Language Models are transforming AI, creating an imperative for the visualization field to embrace agentic frameworks. However, our field's focus on a human in the sensemaking loop raises critical questions about autonomy, delegation, and coordination for such \textit{agentic visualization} that preserve human agency while amplifying analytical capabilities. This paper addresses these questions by reinterpreting existing visualization systems with semi-automated or fully automatic AI components through an agentic lens. Based on this analysis, we extract a collection of design patterns for agentic visualization, including agentic roles, communication and coordination. These patterns provide a foundation for future agentic visualization systems that effectively harness AI agents while maintaining human insight and control.
|
[
{
"version": "v1",
"created": "Sun, 25 May 2025 11:40:38 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:18:22 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dhanoa",
"Vaishali",
""
],
[
"Wolter",
"Anton",
""
],
[
"León",
"Gabriela Molina",
""
],
[
"Schulz",
"Hans-Jörg",
""
],
[
"Elmqvist",
"Niklas",
""
]
] |
2505.19219
|
Wenhao Li
|
Shiyue Wang, Haozheng Xu, Yuhan Zhang, Jingran Lin, Changhong Lu, Xiangfeng Wang, Wenhao Li
|
Where Paths Collide: A Comprehensive Survey of Classic and Learning-Based Multi-Agent Pathfinding
|
112 pages, 21 figures, 20 tables. The project website is: https://wangsh1yue.github.io/Where-Paths-Collide
| null | null | null |
cs.AI cs.LG cs.MA math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
Multi-Agent Path Finding (MAPF) is a fundamental problem in artificial intelligence and robotics, requiring the computation of collision-free paths for multiple agents navigating from their start locations to designated goals. As autonomous systems become increasingly prevalent in warehouses, urban transportation, and other complex environments, MAPF has evolved from a theoretical challenge to a critical enabler of real-world multi-robot coordination. This comprehensive survey bridges the long-standing divide between classical algorithmic approaches and emerging learning-based methods in MAPF research. We present a unified framework that encompasses search-based methods (including Conflict-Based Search, Priority-Based Search, and Large Neighborhood Search), compilation-based approaches (SAT, SMT, CSP, ASP, and MIP formulations), and data-driven techniques (reinforcement learning, supervised learning, and hybrid strategies). Through systematic analysis of experimental practices across 200+ papers, we uncover significant disparities in evaluation methodologies, with classical methods typically tested on larger-scale instances (up to 200 by 200 grids with 1000+ agents) compared to learning-based approaches (predominantly 10-100 agents). We provide a comprehensive taxonomy of evaluation metrics, environment types, and baseline selections, highlighting the need for standardized benchmarking protocols. Finally, we outline promising future directions including mixed-motive MAPF with game-theoretic considerations, language-grounded planning with large language models, and neural solver architectures that combine the rigor of classical methods with the flexibility of deep learning. This survey serves as both a comprehensive reference for researchers and a practical guide for deploying MAPF solutions in increasingly complex real-world applications.
|
[
{
"version": "v1",
"created": "Sun, 25 May 2025 16:28:06 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 14:16:44 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Shiyue",
""
],
[
"Xu",
"Haozheng",
""
],
[
"Zhang",
"Yuhan",
""
],
[
"Lin",
"Jingran",
""
],
[
"Lu",
"Changhong",
""
],
[
"Wang",
"Xiangfeng",
""
],
[
"Li",
"Wenhao",
""
]
] |
2505.20348
|
Steven Samuels
|
Steven Samuels, William Campbell, Michael E. Tobar, and Maxim Goryachev
|
Cryogenic Microwave Whispering Gallery Mode Spectroscopy of Paramagnetic Impurities in High-Purity Crystalline LiF
| null | null | null | null |
quant-ph cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
A low-noise cryogenic microwave spectroscopy experiment was performed on a high-purity lithium fluoride (LiF) crystal. The spectroscopy data revealed avoided level crossing interactions in whispering gallery modes, indicative of electron spin resonance (ESR) coupling with paramagnetic impurities. Analysis of the interaction spectra identified distinct spin systems corresponding to $(S = 3/2, I = 7/2)$, $(S = 1, I = 7/2)$, and $(S = 3/2, I = 0)$. The number of hyperfine splittings observed, together with the natural abundance of ions possessing the appropriate nuclear spin values, suggest that V$^{2+}$ and V$^{3+}$ impurities, exhibiting orthorhombic distortion, are the most likely sources of the narrow interaction features. This interpretation is supported by earlier ESR studies and established manufacturing records for LiF crystal growth. Additionally, a separate set of broader interaction points is consistent with an orthorhombic model involving a $(S = 3/2, I = 0)$ spin system, although the specific impurity responsible for this interaction remains unidentified.
|
[
{
"version": "v1",
"created": "Sun, 25 May 2025 14:42:02 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:31:08 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Samuels",
"Steven",
""
],
[
"Campbell",
"William",
""
],
[
"Tobar",
"Michael E.",
""
],
[
"Goryachev",
"Maxim",
""
]
] |
2505.20389
|
Damiano Francesco Giuseppe Fiorillo
|
Damiano F. G. Fiorillo, Georg G. Raffelt
|
Dispersion relation of the neutrino plasma: Unifying fast, slow, and collisional instabilities
|
49 pages, 7 figures; added analytical discussion and numerical validation of narrow slow instabilities
| null | null | null |
hep-ph astro-ph.CO astro-ph.HE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In neutrino-dense astrophysical environments, these particles exchange flavor through a coherent weak field, forming a collisionless neutrino plasma with collective flavor dynamics. Instabilities, which grow and affect the environment, may arise from neutrino-neutrino refraction alone (fast limit), vacuum energy splittings caused by masses (slow limit), or neutrino-matter scattering (collisional limit). We present a comprehensive analytical description of the dispersion relation governing these unstable modes. Treating vacuum energy splittings and collision rates as small perturbations, we construct a unified framework for fast, slow, and collisional instabilities. We classify modes into gapped, where collective excitations are already present in the fast limit but rendered unstable by slow or collisional effects, and gapless, which are purely generated by these effects. For each class, we derive approximate dispersion relations for generic energy and angle distributions, which reveal the order of magnitude of the growth rates and the nature of the instabilities without solving directly the dispersion relation. This approach confirms that slow and collisionally unstable waves generally grow much more slowly than they oscillate. Consequently, the common fast-mode approximation of local evolution within small boxes is unjustified. Even for fast modes, neglecting large-distance propagation of growing waves, as usually done, may be a poor approximation. Our unified framework provides an intuitive understanding of the linear phase of flavor evolution across all regimes and paves the way for a quasi-linear treatment of the instability's nonlinear development.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 18:00:00 GMT"
},
{
"version": "v2",
"created": "Mon, 02 Jun 2025 05:27:06 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 14:07:02 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Fiorillo",
"Damiano F. G.",
""
],
[
"Raffelt",
"Georg G.",
""
]
] |
2505.20553
|
Goncalo Oliveira
|
Lu\'is Carvalho, Jo\~ao L. Costa, Jos\'e Mour\~ao, Gon\c{c}alo Oliveira
|
A ZeNN architecture to avoid the Gaussian trap
|
New experiments involving PiNNs for solving Schr\"odinger and Bessel-type equations
| null | null | null |
cs.LG math.PR
|
http://creativecommons.org/licenses/by/4.0/
|
We propose a new simple architecture, Zeta Neural Networks (ZeNNs), in order to overcome several shortcomings of standard multi-layer perceptrons (MLPs). Namely, in the large width limit, MLPs are non-parametric, they do not have a well-defined pointwise limit, they lose non-Gaussian attributes and become unable to perform feature learning; moreover, finite width MLPs perform poorly in learning high frequencies. The new ZeNN architecture is inspired by three simple principles from harmonic analysis:
i) Enumerate the perceptons and introduce a non-learnable weight to enforce convergence;
ii) Introduce a scaling (or frequency) factor;
iii) Choose activation functions that lead to near orthogonal systems.
We will show that these ideas allow us to fix the referred shortcomings of MLPs. In fact, in the infinite width limit, ZeNNs converge pointwise, they exhibit a rich asymptotic structure beyond Gaussianity, and perform feature learning. Moreover, when appropriate activation functions are chosen, (finite width) ZeNNs excel at learning high-frequency features of functions with low dimensional domains.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 22:26:55 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 11:11:42 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Carvalho",
"Luís",
""
],
[
"Costa",
"João L.",
""
],
[
"Mourão",
"José",
""
],
[
"Oliveira",
"Gonçalo",
""
]
] |
2505.20884
|
Xu Wang
|
Weichao Pan, Bohan Xu, Xu Wang, Chengze Lv, Shuoyang Wang, Zhenke Duan
|
YOLO-FireAD: Efficient Fire Detection via Attention-Guided Inverted Residual Learning and Dual-Pooling Feature Preservation
|
2025 International Conference on Intelligent Computing (ICIC 2025)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Fire detection in dynamic environments faces continuous challenges, including the interference of illumination changes, many false detections or missed detections, and it is difficult to achieve both efficiency and accuracy. To address the problem of feature extraction limitation and information loss in the existing YOLO-based models, this study propose You Only Look Once for Fire Detection with Attention-guided Inverted Residual and Dual-pooling Downscale Fusion (YOLO-FireAD) with two core innovations: (1) Attention-guided Inverted Residual Block (AIR) integrates hybrid channel-spatial attention with inverted residuals to adaptively enhance fire features and suppress environmental noise; (2) Dual Pool Downscale Fusion Block (DPDF) preserves multi-scale fire patterns through learnable fusion of max-average pooling outputs, mitigating small-fire detection failures. Extensive evaluation on two public datasets shows the efficient performance of our model. Our proposed model keeps the sum amount of parameters (1.45M, 51.8% lower than YOLOv8n) (4.6G, 43.2% lower than YOLOv8n), and mAP75 is higher than the mainstream real-time object detection models YOLOv8n, YOL-Ov9t, YOLOv10n, YOLO11n, YOLOv12n and other YOLOv8 variants 1.3-5.5%. For more details, please visit our repository: https://github.com/JEFfersusu/YOLO-FireAD
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 08:29:07 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 04:04:04 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Pan",
"Weichao",
""
],
[
"Xu",
"Bohan",
""
],
[
"Wang",
"Xu",
""
],
[
"Lv",
"Chengze",
""
],
[
"Wang",
"Shuoyang",
""
],
[
"Duan",
"Zhenke",
""
]
] |
2505.20980
|
Micha{\l} Czuba Mr
|
Micha{\l} Czuba, Mateusz Stolarski, Adam Pir\'og, Piotr Bielak, Piotr Br\'odka
|
Identifying Super Spreaders in Multilayer Networks
| null | null | null | null |
cs.SI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Identifying super-spreaders can be framed as a subtask of the influence maximisation problem. It seeks to pinpoint agents within a network that, if selected as single diffusion seeds, disseminate information most effectively. Multilayer networks, a specific class of heterogeneous graphs, can capture diverse types of interactions (e.g., physical-virtual or professional-social), and thus offer a more accurate representation of complex relational structures. In this work, we introduce a novel approach to identifying super-spreaders in such networks by leveraging graph neural networks. To this end, we construct a dataset by simulating information diffusion across hundreds of networks - to the best of our knowledge, the first of its kind tailored specifically to multilayer networks. We further formulate the task as a variation of the ranking prediction problem based on a four-dimensional vector that quantifies each agent's spreading potential: (i) the number of activations; (ii) the duration of the diffusion process; (iii) the peak number of activations; and (iv) the simulation step at which this peak occurs. Our model, TopSpreadersNetwork, comprises a relationship-agnostic encoder and a custom aggregation layer. This design enables generalisation to previously unseen data and adapts to varying graph sizes. In an extensive evaluation, we compare our model against classic centrality-based heuristics and competitive deep learning methods. The results, obtained across a broad spectrum of real-world and synthetic multilayer networks, demonstrate that TopSpreadersNetwork achieves superior performance in identifying high-impact nodes, while also offering improved interpretability through its structured output.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 10:14:14 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:48:42 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Czuba",
"Michał",
""
],
[
"Stolarski",
"Mateusz",
""
],
[
"Piróg",
"Adam",
""
],
[
"Bielak",
"Piotr",
""
],
[
"Bródka",
"Piotr",
""
]
] |
2505.21567
|
Zihao Zheng
|
Feng Jiang, Zihao Zheng, Xiuping Cui, Maoliang Li, JIayu Chen, Xiang Chen
|
EaqVLA: Encoding-aligned Quantization for Vision-Language-Action Models
|
There is an error in this paper, and as the author, I request retraction
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
With the development of Embodied Artificial intelligence, the end-to-end control policy such as Vision-Language-Action (VLA) model has become the mainstream. Existing VLA models faces expensive computing/storage cost, which need to be optimized. Quantization is considered as the most effective method which can not only reduce the memory cost but also achieve computation acceleration. However, we find the token alignment of VLA models hinders the application of existing quantization methods. To address this, we proposed an optimized framework called EaqVLA, which apply encoding-aligned quantization to VLA models. Specifically, we propose an complete analysis method to find the misalignment in various granularity. Based on the analysis results, we propose a mixed precision quantization with the awareness of encoding alignment. Experiments shows that the porposed EaqVLA achieves better quantization performance (with the minimal quantization loss for end-to-end action control and xxx times acceleration) than existing quantization methods.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 05:42:21 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:30:45 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Jiang",
"Feng",
""
],
[
"Zheng",
"Zihao",
""
],
[
"Cui",
"Xiuping",
""
],
[
"Li",
"Maoliang",
""
],
[
"Chen",
"JIayu",
""
],
[
"Chen",
"Xiang",
""
]
] |
2505.22186
|
Mark Fortune
|
Mark Fortune, Neale P. Gibson, Hannah Diamond-Lowe, Jo\~ao M. Mendon\c{c}a, Am\'elie Gressier, Daniel Kitzmann, Natalie H. Allen, Prune C. August, Jegug Ih, Erik Meier Vald\'es, Merlin Zgraggen, Lars A. Buchhave, Brice-Olivier Demory, N\'estor Espinoza, Kevin Heng, Kathryn Jones, Alexander D. Rathcke
|
Hot Rocks Survey III: A deep eclipse for LHS 1140c and a new Gaussian process method to account for correlated noise in individual pixels
|
Accepted for publication in Astronomy & Astrophysics (updated version after language edits). 35 pages, 28 figures
| null | null | null |
astro-ph.EP astro-ph.IM
|
http://creativecommons.org/licenses/by/4.0/
|
Time-series photometry at mid-infrared wavelengths is becoming a common technique to search for atmospheres around rocky exoplanets. This method constrains the brightness temperature of the planet to determine whether heat redistribution is taking place - indicative of an atmosphere - or whether the heat is reradiated from a low albedo bare rock. By observing at 15$\mu$m we are also highly sensitive to CO$_2$ absorption. We observed three eclipses of the rocky super-Earth LHS 1140c using MIRI/Imaging with the F1500W filter. We found significant variation in the initial settling ramp for these observations and identify a potential trend between detector settling and the previous filter used by MIRI. We analysed our data using aperture photometry but also developed a novel approach which joint-fits pixel light curves directly using a shared eclipse model and a flexible multi-dimensional Gaussian process which models changes in the PSF over time. We demonstrate using simulated data that our method has the ability to weight away from particular pixels which show increased systematics, allowing for the recovery of eclipse depths in a more robust and precise way. Both methods and an independent analysis detect the eclipse at $>5\sigma$ and are highly consistent with a low albedo bare rock. We recover a dayside brightness temperature of $T_\mathrm{day} = 561\pm44$ K, close to the theoretical maximum of $T_\text{day; max} = 537\pm9$ K. We rule out a wide range of atmospheric forward models to $>3\sigma$ including pure CO$_2$ atmospheres with surface pressure $\ge10$ mbar and pure H$_2$O atmospheres with surface pressure $\ge1$ bar. Our strict constraints on potential atmospheric composition, in combination with future observations of the exciting outer planet LHS 1140b, could provide a powerful benchmark to understand atmospheric escape around M dwarfs.
|
[
{
"version": "v1",
"created": "Wed, 28 May 2025 09:59:03 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:43:53 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Fortune",
"Mark",
""
],
[
"Gibson",
"Neale P.",
""
],
[
"Diamond-Lowe",
"Hannah",
""
],
[
"Mendonça",
"João M.",
""
],
[
"Gressier",
"Amélie",
""
],
[
"Kitzmann",
"Daniel",
""
],
[
"Allen",
"Natalie H.",
""
],
[
"August",
"Prune C.",
""
],
[
"Ih",
"Jegug",
""
],
[
"Valdés",
"Erik Meier",
""
],
[
"Zgraggen",
"Merlin",
""
],
[
"Buchhave",
"Lars A.",
""
],
[
"Demory",
"Brice-Olivier",
""
],
[
"Espinoza",
"Néstor",
""
],
[
"Heng",
"Kevin",
""
],
[
"Jones",
"Kathryn",
""
],
[
"Rathcke",
"Alexander D.",
""
]
] |
2505.23574
|
Adam Czarnecki
|
Adam J. Czarnecki, Nikola L. Kolev, Patrick See, Nick Sullivan, Wyatt A. Behn, Neil J. Curson, Taylor J.Z. Stock, Peter Gr\"utter
|
Hydrogen Passivation Effects on Spatially Resolved Charge Trap Densities in Si(100)-SiO$_2$
|
8 pages, 11 figures, modified title, added new Fig 5 to directly compare counting methods to manual counts, added details in the discussion in section III and IV, added details about DL method in supplementary section, fixed typos
| null | null | null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
As silicon-based devices continue to shrink to the nanoscale, traps at the Si-SiO$_2$ interface pose increasing challenges to device performance. These traps reduce channel carrier mobility and shift threshold voltages in integrated circuits, and introduce charge noise in quantum systems, reducing their coherence times. Knowledge of the precise location of such traps aids in understanding their influence on device performance. In this work, we demonstrate that frequency-modulated atomic force microscopy (fm-AFM) allows the detection of individual traps. We use this to study how sample preparation, specifically the introduction of a buried hydrogen termination layer, and post-processing annealing in forming gas (N$_2$+H$_2$), affects the density of donor-like traps in Si(100)-SiO$_2$ systems. We spatially map and quantify traps in both conventionally prepared ("pristine") silicon samples and those processed under ultra-high vacuum for hydrogen resist lithography (HRL). We confirm previous studies demonstrating hydrogen passivation of traps and find that hydrogen termination further reduces the donor-like trap density. We also observe a significant reduction in two-level donor-like traps in the hydrogen-terminated samples compared to pristine silicon samples. These findings suggest that HRL-prepared silicon may offer advantages for high-performance nanoscale and atomic-scale devices due to reduced trap densities.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 15:46:07 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 23:00:25 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Czarnecki",
"Adam J.",
""
],
[
"Kolev",
"Nikola L.",
""
],
[
"See",
"Patrick",
""
],
[
"Sullivan",
"Nick",
""
],
[
"Behn",
"Wyatt A.",
""
],
[
"Curson",
"Neil J.",
""
],
[
"Stock",
"Taylor J. Z.",
""
],
[
"Grütter",
"Peter",
""
]
] |
2505.23628
|
Jiaxin Bai
|
Jiaxin Bai, Wei Fan, Qi Hu, Qing Zong, Chunyang Li, Hong Ting Tsang, Hongyu Luo, Yauwai Yim, Haoyu Huang, Xiao Zhou, Feng Qin, Tianshi Zheng, Xi Peng, Xin Yao, Huiwen Yang, Leijie Wu, Yi Ji, Gong Zhang, Renhai Chen, Yangqiu Song
|
AutoSchemaKG: Autonomous Knowledge Graph Construction through Dynamic Schema Induction from Web-Scale Corpora
|
9 pages, preprint, code: https://github.com/HKUST-KnowComp/AutoSchemaKG
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
We present AutoSchemaKG, a framework for fully autonomous knowledge graph construction that eliminates the need for predefined schemas. Our system leverages large language models to simultaneously extract knowledge triples and induce comprehensive schemas directly from text, modeling both entities and events while employing conceptualization to organize instances into semantic categories. Processing over 50 million documents, we construct ATLAS (Automated Triple Linking And Schema induction), a family of knowledge graphs with 900+ million nodes and 5.9 billion edges. This approach outperforms state-of-the-art baselines on multi-hop QA tasks and enhances LLM factuality. Notably, our schema induction achieves 95\% semantic alignment with human-crafted schemas with zero manual intervention, demonstrating that billion-scale knowledge graphs with dynamically induced schemas can effectively complement parametric knowledge in large language models.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 16:34:58 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 06:40:31 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bai",
"Jiaxin",
""
],
[
"Fan",
"Wei",
""
],
[
"Hu",
"Qi",
""
],
[
"Zong",
"Qing",
""
],
[
"Li",
"Chunyang",
""
],
[
"Tsang",
"Hong Ting",
""
],
[
"Luo",
"Hongyu",
""
],
[
"Yim",
"Yauwai",
""
],
[
"Huang",
"Haoyu",
""
],
[
"Zhou",
"Xiao",
""
],
[
"Qin",
"Feng",
""
],
[
"Zheng",
"Tianshi",
""
],
[
"Peng",
"Xi",
""
],
[
"Yao",
"Xin",
""
],
[
"Yang",
"Huiwen",
""
],
[
"Wu",
"Leijie",
""
],
[
"Ji",
"Yi",
""
],
[
"Zhang",
"Gong",
""
],
[
"Chen",
"Renhai",
""
],
[
"Song",
"Yangqiu",
""
]
] |
2505.23998
|
Ali Enayat
|
Ali Enayat
|
The Mostowski Bridge
|
16 pages; minor imperfections of the previous version are corrected in this one, and the typography is improved
| null | null | null |
math.LO
|
http://creativecommons.org/licenses/by/4.0/
|
In 1950, Novak and Mostowski showed that GB (G\"odel-Bernays theory of classes) is conservative over ZF, and therefore by G\"odel's second incompleteness theorem the consistency of ZF is unprovable in GB. In the same year Mostowski unveiled a contrasting result: GB provides a truth-definition for ZF-formulae. Here we first give an expository account of Mostowski's construction and surrounding results, and then we show that the construction bridges the domain of Tarski-style truth theories over PA with certain subsystems of second order arithmetic.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 20:47:36 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Jun 2025 16:15:50 GMT"
},
{
"version": "v3",
"created": "Sat, 05 Jul 2025 20:01:57 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 17:48:54 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Enayat",
"Ali",
""
]
] |
2505.24329
|
Zepeng Huang
|
Yingsen Zeng, Zepeng Huang, Yujie Zhong, Chengjian Feng, Jie Hu, Lin Ma, Yang Liu
|
DisTime: Distribution-based Time Representation for Video Large Language Models
|
Accepted by ICCV 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Despite advances in general video understanding, Video Large Language Models (Video-LLMs) face challenges in precise temporal localization due to discrete time representations and limited temporally aware datasets. Existing methods for temporal expression either conflate time with text-based numerical values, add a series of dedicated temporal tokens, or regress time using specialized temporal grounding heads. To address these issues, we introduce DisTime, a lightweight framework designed to enhance temporal comprehension in Video-LLMs. DisTime employs a learnable token to create a continuous temporal embedding space and incorporates a Distribution-based Time Decoder that generates temporal probability distributions, effectively mitigating boundary ambiguities and maintaining temporal continuity. Additionally, the Distribution-based Time Encoder re-encodes timestamps to provide time markers for Video-LLMs. To overcome temporal granularity limitations in existing datasets, we propose an automated annotation paradigm that combines the captioning capabilities of Video-LLMs with the localization expertise of dedicated temporal models. This leads to the creation of InternVid-TG, a substantial dataset with 1.25M temporally grounded events across 179k videos, surpassing ActivityNet-Caption by 55 times. Extensive experiments demonstrate that DisTime achieves state-of-the-art performance across benchmarks in three time-sensitive tasks while maintaining competitive performance in Video QA tasks. Code and data are released at https://github.com/josephzpng/DisTime.
|
[
{
"version": "v1",
"created": "Fri, 30 May 2025 08:10:18 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:03:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zeng",
"Yingsen",
""
],
[
"Huang",
"Zepeng",
""
],
[
"Zhong",
"Yujie",
""
],
[
"Feng",
"Chengjian",
""
],
[
"Hu",
"Jie",
""
],
[
"Ma",
"Lin",
""
],
[
"Liu",
"Yang",
""
]
] |
2506.00068
|
Afrozah Nadeem
|
Afrozah Nadeem, Mark Dras, and Usman Naseem
|
Framing Political Bias in Multilingual LLMs Across Pakistani Languages
|
Preprint
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Large Language Models (LLMs) increasingly shape public discourse, yet most evaluations of political and economic bias have focused on high-resource, Western languages and contexts. This leaves critical blind spots in low-resource, multilingual regions such as Pakistan, where linguistic identity is closely tied to political, religious, and regional ideologies. We present a systematic evaluation of political bias in 13 state-of-the-art LLMs across five Pakistani languages: Urdu, Punjabi, Sindhi, Pashto, and Balochi. Our framework integrates a culturally adapted Political Compass Test (PCT) with multi-level framing analysis, capturing both ideological stance (economic/social axes) and stylistic framing (content, tone, emphasis). Prompts are aligned with 11 socio-political themes specific to the Pakistani context. Results show that while LLMs predominantly reflect liberal-left orientations consistent with Western training data, they exhibit more authoritarian framing in regional languages, highlighting language-conditioned ideological modulation. We also identify consistent model-specific bias patterns across languages. These findings show the need for culturally grounded, multilingual bias auditing frameworks in global NLP.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 15:15:42 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 04:41:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Nadeem",
"Afrozah",
""
],
[
"Dras",
"Mark",
""
],
[
"Naseem",
"Usman",
""
]
] |
2506.00497
|
Anders Malthe Westerkam
|
Anders Malthe Westerkam, Alba Spliid Damkj{\ae}r, Rasmus Erik Villadsen, Magnus {\O}rum Bastrup Poulsen, Troels Pedersen
|
Second-Order Characterization of Micro Doppler Radar Signatures of Drone Swarms
| null | null | null | null |
eess.SP
|
http://creativecommons.org/licenses/by/4.0/
|
We investigate the second-order characteristics of the radar return signal from a swarm of rotor drones. We consider the case of a swarm of identical drones, with each a number of rotors comprised of a number of rotor blades. By considering the orientation and speed of each rotor as stochastic variables, we derive expressions for the autocorrelation function (ACF) and power spectral density (PSD). The ACF and PSD are in the form of an infinite series with coefficients that drop to zero at a predictable limit. Thus in practical applications, the series may be truncated. As a special case, we show that for deterministic rotor speed, the ACF can be expressed in closed form. We further investigate how system parameters (Blade length, Rotor speed, number of blades, and number of drones) influence the derived expressions for the ACF and PSD.
|
[
{
"version": "v1",
"created": "Sat, 31 May 2025 10:30:03 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Jun 2025 14:00:12 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 15:07:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Westerkam",
"Anders Malthe",
""
],
[
"Damkjær",
"Alba Spliid",
""
],
[
"Villadsen",
"Rasmus Erik",
""
],
[
"Poulsen",
"Magnus Ørum Bastrup",
""
],
[
"Pedersen",
"Troels",
""
]
] |
2506.00540
|
Wenzhang Liu
|
Wenzhang Liu, Muqaddar Abbas, Jiawei Lai and Pei Zhang
|
Manipulation of photonic spin Hall effect in the Rydberg atomic medium
|
We have noticed obvious mistakes in methods and figures and we are working on it, I want to withdraw this manuscript first
| null | null | null |
quant-ph physics.optics
|
http://creativecommons.org/licenses/by/4.0/
|
We present a theoretical study demonstrating enhanced tunability of the photonic spin Hall effect (PSHE) using a strongly interacting Rydberg atomic medium under electromagnetically induced transparency (EIT) conditions. In contrast to conventional approaches that rely on static refractiveindex profiles or metamaterials, here the PSHE is controlled via a nonlocal third-order nonlinear susceptibility arising from long range Rydberg-Rydberg interactions. We show that this nonlocal nonlinearity enables dynamic modulation of spin-dependent light trajectories, amplifying the normally weak PSHE into a readily observable and adjustable effect. These results pave the way for new capabilities in photonic information processing and sensing. In particular, an adjustable PSHE may enable beam steering based on photon spin, improve the sensitivity of precision measurements, and support photonic devices whose functionality can be reconfigured in real time.
|
[
{
"version": "v1",
"created": "Sat, 31 May 2025 12:55:55 GMT"
},
{
"version": "v2",
"created": "Thu, 05 Jun 2025 09:12:06 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 10:48:23 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Liu",
"Wenzhang",
""
],
[
"Abbas",
"Muqaddar",
""
],
[
"Lai",
"Jiawei",
""
],
[
"Zhang",
"Pei",
""
]
] |
2506.00639
|
Zhuoyi Pang
|
Yao Ji, Zhuoyi Pang, Fei Yao, and Jian-Hui Zhang
|
Regularization Prescription for the Mixing Between Nonlocal Gluon and Quark Operators
|
12 pages, 0 figure
| null | null | null |
hep-ph hep-lat nucl-ex nucl-th
|
http://creativecommons.org/licenses/by/4.0/
|
It is well-known that in the study of mixing between nonlocal gluon and quark bilinear operators there exists an ambiguity when relating coordinate space and momentum space results, which can be conveniently resolved through Mellin moments matching in both spaces. In this work, we show that this ambiguity is due to the lack of a proper regularization prescription of the singularity that arises when the separation between the gluon/quark fields approaches zero. We then demonstrate that dimensional regularization resolves this issue and yields consistent results in both coordinate and momentum space. This prescription is also compatible with lattice extractions of parton distributions from nonlocal operators.
|
[
{
"version": "v1",
"created": "Sat, 31 May 2025 17:04:15 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 06:36:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ji",
"Yao",
""
],
[
"Pang",
"Zhuoyi",
""
],
[
"Yao",
"Fei",
""
],
[
"Zhang",
"Jian-Hui",
""
]
] |
2506.00956
|
YoungJoon Yoo
|
Geonu Lee, Yujeong Oh, Geonhui Jang, Soyoung Lee, Jeonghyo Song, Sungmin Cha, YoungJoon Yoo
|
Continual-MEGA: A Large-scale Benchmark for Generalizable Continual Anomaly Detection
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we introduce a new benchmark for continual learning in anomaly detection, aimed at better reflecting real-world deployment scenarios. Our benchmark, Continual-MEGA, includes a large and diverse dataset that significantly expands existing evaluation settings by combining carefully curated existing datasets with our newly proposed dataset, ContinualAD. In addition to standard continual learning with expanded quantity, we propose a novel scenario that measures zero-shot generalization to unseen classes, those not observed during continual adaptation. This setting poses a new problem setting that continual adaptation also enhances zero-shot performance. We also present a unified baseline algorithm that improves robustness in few-shot detection and maintains strong generalization. Through extensive evaluations, we report three key findings: (1) existing methods show substantial room for improvement, particularly in pixel-level defect localization; (2) our proposed method consistently outperforms prior approaches; and (3) the newly introduced ContinualAD dataset enhances the performance of strong anomaly detection models. We release the benchmark and code in https://github.com/Continual-Mega/Continual-Mega.
|
[
{
"version": "v1",
"created": "Sun, 01 Jun 2025 11:00:24 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:31:51 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lee",
"Geonu",
""
],
[
"Oh",
"Yujeong",
""
],
[
"Jang",
"Geonhui",
""
],
[
"Lee",
"Soyoung",
""
],
[
"Song",
"Jeonghyo",
""
],
[
"Cha",
"Sungmin",
""
],
[
"Yoo",
"YoungJoon",
""
]
] |
2506.01447
|
Klaas De Kinder
|
Klaas De Kinder, Amir Bahrami and Christophe Caloz
|
Doppler Pulse Amplification
|
8 pages, 5 figures
| null | null | null |
physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The ability to amplify ultrashort pulses has revolutionized modern laser science, driving advances in various fields such as ultrafast optics and spectroscopy. A pivotal development in this field is chirped pulse amplification (CPA), which stretches, amplifies and recompresses ultrashort optical pulses using dispersive elements to overcome amplification limits. However, CPA faces limitations due to gain narrowing, restricting the final pulse duration. Here, we propose Doppler pulse amplification (DoPA), a novel approach for amplifying ultrashort pulses. While DoPA shares similarities with CPA in that it also stretches, amplifies and recompresses pulses, it differs in how it achieves this temporal compansion. Unlike CPA, DoPA exploits Doppler shifts induced by space-time modulated interfaces through a space-time wedge implementation without chirping. We show that DoPA dynamically shifts the pulse spectrum, effectively mitigating the gain narrowing issue of CPA. Additionally, we show that DoPA enables more compact amplification systems via a space-time Fresnel implementation. This approach may pave the way for more efficient, high-intensity laser systems and expand the potential for applications in both laboratory research and practical environments.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 09:05:36 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 09:30:40 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 08:00:30 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"De Kinder",
"Klaas",
""
],
[
"Bahrami",
"Amir",
""
],
[
"Caloz",
"Christophe",
""
]
] |
2506.01932
|
Diego Catalano Ferraioli
|
Diego Catalano Ferraioli, Tarc\'isio Castro Silva
|
Nonlocal pseudosymmetries and B\"acklund transformations as $\mathcal{C}$-morphisms
| null | null | null | null |
math.DG math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we show how factorisation with respect to nonlocal pseudosymmetries allows one to obtain B\"acklund transformations, interpreted as nonlocal $\mathcal{C}$-morphisms of differential equations. According to this approach, which is illustrated through several examples, the B\"acklund transformations are determined by basic invariants of the exploited nonlocal pseudosymmetries.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 17:50:34 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Jun 2025 02:44:44 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 14:22:32 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ferraioli",
"Diego Catalano",
""
],
[
"Silva",
"Tarcísio Castro",
""
]
] |
2506.02169
|
Parakh Manoj Gupta
|
Parakh M. Gupta, Ond\v{r}ej Proch\'azka, Jan H\v{r}ebec, Matej Novosad, Robert P\v{e}ni\v{c}ka, Martin Saska
|
LoL-NMPC: Low-Level Dynamics Integration in Nonlinear Model Predictive Control for Unmanned Aerial Vehicles
|
Accepted to IROS 2025
| null | null | null |
cs.RO
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
[Accepted to IROS 2025] In this paper, we address the problem of tracking high-speed agile trajectories for Unmanned Aerial Vehicles(UAVs), where model inaccuracies can lead to large tracking errors. Existing Nonlinear Model Predictive Controller(NMPC) methods typically neglect the dynamics of the low-level flight controllers such as underlying PID controller present in many flight stacks, and this results in sub-optimal tracking performance at high speeds and accelerations. To this end, we propose a novel NMPC formulation, LoL-NMPC, which explicitly incorporates low-level controller dynamics and motor dynamics in order to minimize trajectory tracking errors while maintaining computational efficiency. By leveraging linear constraints inside low-level dynamics, our approach inherently accounts for actuator constraints without requiring additional reallocation strategies. The proposed method is validated in both simulation and real-world experiments, demonstrating improved tracking accuracy and robustness at speeds up to 98.57 km/h and accelerations of 3.5 g. Our results show an average 21.97 % reduction in trajectory tracking error over standard NMPC formulation, with LoL-NMPC maintaining real-time feasibility at 100 Hz on an embedded ARM-based flight computer.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 18:55:03 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:25:01 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gupta",
"Parakh M.",
""
],
[
"Procházka",
"Ondřej",
""
],
[
"Hřebec",
"Jan",
""
],
[
"Novosad",
"Matej",
""
],
[
"Pěnička",
"Robert",
""
],
[
"Saska",
"Martin",
""
]
] |
2506.02844
|
Ahmad Jafar Arifi
|
Ahmad Jafar Arifi, Ho-Meoyng Choi, Chueng-Ryong Ji
|
Beyond leading twist: $\rho$ meson decay constants and distribution amplitudes in a self-consistent light-front quark model
|
14 pages, 3 figures, 4 tables. To appear in PRD
| null | null | null |
hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
In this study, we present a comprehensive analysis of decay constants and chiral-even and chiral-odd distribution amplitudes (DAs) up to twist 4 for the $\rho$ meson in the standard light-front quark model (LFQM) based on the Bakamjian-Thomas (BT) construction. For the $\rho$ meson, which possesses both longitudinal $(h=0)$ and transverse $(h=\pm 1)$ polarizations, two types of decay constants, $f_{\rho}^{\parallel}$ and $f_{\rho}^{\perp}$, arises accordingly. We demonstrate that these decay constants can be self-consistently extracted from both local ($z^\mu=0$) and nonlocal ($z^\mu\neq 0$) matrix elements $\langle 0 | \bar{q}(z)\, \Gamma\, q(-z) | \rho(P,h) \rangle$, with $\Gamma = (\gamma^\mu, \sigma^{\mu\nu}, \gamma^\mu\gamma_5, \mathbf{1})$, in a manner independent of current components, polarizations, and reference frames. In particular, we emphasize the role of nonlocal matrix elements involving axial-vector and scalar currents, where mixing between $f_\rho^{\parallel}$ and $f_\rho^{\perp}$ occurs. We show that this mixing is consistently resolved through the BT construction, ensuring the proper extraction of these decay constants. Additionally, we investigate the structure of chiral-even DAs ($\phi_{2;\mathrm{V}}^\parallel, \phi_{3;\mathrm{V}}^\parallel, \psi_{3;\mathrm{A}}^\perp, \phi_{4;\mathrm{V}}^\parallel$) and chiral-odd DAs ($\phi_{2;\mathrm{T}}^\perp, \phi_{3;\mathrm{T}}^\perp, \psi_{3;\mathrm{S}}^\parallel, \phi_{4;\mathrm{T}}^\perp$) beyond leading twists, and present their corresponding $\xi$-moments and Gegenbauer moments. These results provide deeper insight into the nonperturbative structure of vector mesons and demonstrate the robustness and self-consistency of the LFQM based on the BT framework.
|
[
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:14:54 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 01:01:32 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Arifi",
"Ahmad Jafar",
""
],
[
"Choi",
"Ho-Meoyng",
""
],
[
"Ji",
"Chueng-Ryong",
""
]
] |
2506.03347
|
Paul Zivich
|
Paul N Zivich, Haidong Lu
|
Constructing g-computation estimators: two case studies in selection bias
| null | null | null | null |
stat.ME
|
http://creativecommons.org/licenses/by/4.0/
|
G-computation is a useful estimation method that can be adapted to address various biases in epidemiology. However, these adaptations may not be obvious for some complex causal structures. This challenge is an example of the much wider issue of translating a causal diagram into a novel estimation strategy. To highlight these challenges, we consider two recent cases from the selection bias literature: treatment-induced selection and co-occurrence of biases that lack a joint adjustment set. For each case study, we show how g-computation can be adapted, describe how to implement that adaptation, show some general statistical properties, and illustrate the estimator using simulation. To simplify both the theoretical study and practical application of our estimators, we express the proposed g-computation estimators as stacked estimating equations. These examples illustrate how epidemiologists can translate identification results into a g-computation estimator and study the theoretical and finite-sample properties of a novel estimator.
|
[
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:38:43 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 14:18:36 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zivich",
"Paul N",
""
],
[
"Lu",
"Haidong",
""
]
] |
2506.03457
|
Lei Bill Wang
|
Lei Bill Wang, Sooa Ahn
|
Attention vs Choice in Incomplete Welfare Take-Up: What Works for WIC?
| null | null | null | null |
econ.GN q-fin.EC
|
http://creativecommons.org/licenses/by/4.0/
|
We propose two approaches to decompose the reasons for welfare program non-participation among eligible households into two latent mechanisms: attention and choice. Approach I applies the attention-choice model from the consumer economics literature to the welfare take-up problem. Approach II incorporates random effects into Approach I, allowing for unobserved household-level heterogeneity and conditional dependence between attention and choice. We show that critical aspects of Approach II are semiparametrically identifiable through a novel strategy we call coherent shuffle. Applied to NLSY data for WIC participation, both approaches show that a choice-inducing solution is more effective in promoting take-up than an attention-raising solution. We validate our claims using the Vermont WIC2FIVE quasi-experimental data. The permutation test and DiD estimate strongly support that choice-inducing text messages are highly effective in retaining eligible households in the WIC program, while attention-raising text messages hardly have any effect on the retention rate.
|
[
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:40:32 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 02:41:24 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Lei Bill",
""
],
[
"Ahn",
"Sooa",
""
]
] |
2506.03936
|
Quentin Vigneron
|
Quentin Vigneron, Hamed Barzegar, James Read
|
Affine connections for Galilean and Carrollian structures: a unified perspective
|
18 + 4 pages, 3 tables, 1 figure; v2 minor corrections; Accepted in CQG
| null | null | null |
gr-qc hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We develop a classification of general Carrollian structures, permitting affine connections with both torsion and non-metricity. We compare with a recent classification of general Galilean structures in order to present a unified perspective on both. Moreover, we demonstrate how both sets of structures emerge from the most general possible Lorentzian structures in their respective limits, and we highlight the role of global hyperbolicity in constraining both structures. We then leverage this work in order to construct for the first time an ultra-relativistic geometric trinity of gravitational theories, and consider connections which are simultaneously compatible with Galilean and Carrollian structures. We close by outlining a number of open questions and future prospects.
|
[
{
"version": "v1",
"created": "Wed, 04 Jun 2025 13:29:39 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:45:39 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Vigneron",
"Quentin",
""
],
[
"Barzegar",
"Hamed",
""
],
[
"Read",
"James",
""
]
] |
2506.03956
|
Aojun Lu
|
Aojun Lu, Tao Feng, Hangjie Yuan, Chunhui Ding and Yanan Sun
|
Adapt before Continual Learning
| null | null | null | null |
cs.LG cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Continual Learning (CL) seeks to enable neural networks to incrementally acquire new knowledge (plasticity) while retaining existing knowledge (stability). Although pre-trained models (PTMs) have provided a strong foundation for CL, existing approaches face a fundamental challenge in balancing these two competing objectives. Current methods typically address stability by freezing the PTM backbone, which severely limits the model's plasticity, particularly when incoming data distribution diverges largely from the pre-training data. Alternatively, sequentially fine-tuning the entire PTM can adapt to new knowledge but often leads to catastrophic forgetting, highlighting the critical stability-plasticity trade-off in PTM-based CL. To address this limitation, we propose Adapting PTMs before the core CL} process (ACL), a novel framework that introduces a plug-and-play adaptation phase prior to learning each new task. During this phase, ACL refines the PTM backbone by aligning embeddings with their original class prototypes while distancing them from irrelevant classes. This mechanism theoretically and empirically demonstrates desirable balance between stability and plasticity, significantly improving CL performance across benchmarks and integrated methods. Code is available at https://github.com/byyx666/ACL_code.
|
[
{
"version": "v1",
"created": "Wed, 04 Jun 2025 13:46:33 GMT"
},
{
"version": "v2",
"created": "Thu, 05 Jun 2025 08:01:57 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 03:04:31 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lu",
"Aojun",
""
],
[
"Feng",
"Tao",
""
],
[
"Yuan",
"Hangjie",
""
],
[
"Ding",
"Chunhui",
""
],
[
"Sun",
"Yanan",
""
]
] |
2506.04329
|
Jie Chen
|
Jie Chen, Linhua Jiang, Shengxiu Sun, Zijian Zhang, Mouyuan Sun
|
Estimating Bolometric Luminosities of Type 1 Quasars with Self-Organizing Maps
|
18 pages, 13 figures, accepted for publication in ApJ. Code QSOLbol is available at https://github.com/ChenJiemi/QSOLbol
| null | null | null |
astro-ph.CO astro-ph.GA astro-ph.HE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a new method to calculate bolometric luminosities for unobscured, type 1 quasars with multi-band photometric data. Bolometric luminosity is a fundamental property to understand quasars and it is commonly estimated from monochromatic luminosities using bolometric corrections that often neglect quasar SED diversity. We take advantage of the fact that most quasars now have multi-band observations from UV to mid-IR, and construct SEDs for a well-defined sample of SDSS quasars at $0.5\leq z\leq 2$. Based on this fiducial sample, we explore quasar SEDs, their diversity, and their relations with bolometric luminosities. We then use unsupervised neural network self-organizing maps (SOM) to describe the SED diversity and compute the bolometric luminosities with a fully-trained SOM model. This method reduces systematical uncertainties compared to the traditional method. In addition, we update the multi-linear regression relations between bolometric luminosity and monochromatic luminosities at restframe 1450\r{A}, 3000\r{A}, and 5100\r{A}. Our method is applicable to large quasar samples with a wide range of luminosity and redshift. We have applied it to the SDSS DR16 quasars. We have also made our code publicly available.
|
[
{
"version": "v1",
"created": "Wed, 04 Jun 2025 18:00:04 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:19:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Chen",
"Jie",
""
],
[
"Jiang",
"Linhua",
""
],
[
"Sun",
"Shengxiu",
""
],
[
"Zhang",
"Zijian",
""
],
[
"Sun",
"Mouyuan",
""
]
] |
2506.04754
|
Mohd Tahseen
|
Mohd Tahseen, Vivekanand Dabade
|
Revisiting the cofactor conditions: Elimination of transition layers in compound domains
| null | null | null | null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
This paper investigates the conditions necessary for the elimination of transition layers at interfaces involving compound domains, extending the classical framework of cofactor conditions. Although cofactor conditions enable stress-free phase boundaries between Type I/II domains and austenite, their applicability to compound domains has remained limited. Here, we present a comprehensive theoretical framework to characterize all compatible interfaces, highlighting the fundamental importance of the commutation property among martensitic variants. By establishing necessary and sufficient algebraic conditions, referred to as extreme compatibility conditions, we demonstrate the simultaneous elimination of transition layers at phase interfaces for both Type I/II and compound laminates, across all volume fractions of the martensitic variants. We also investigate the possibility of achieving supercompatibility in non-conventional twins, recently observed in the NiMnGa system. The focus of our work is on cubic-to-orthorhombic and cubic-to-monoclinic~II phase transformations, for which the extreme compatibility conditions are explicitly derived and systematically analyzed. The theory predicts novel zero-elastic-energy microstructures, including an increased number of triple clusters, spearhead-shaped martensitic nuclei, stress-free inclusions of austenite within martensite, and distinctive four-fold martensitic clusters. This significantly expands the possible modes of forming stress-free interfaces between phases and reveals new energy-minimizing microstructures that can facilitate the nucleation of martensite within austenite and vice versa. These configurations highlight significant enhancements in transformation reversibility and material durability, guiding the rational design of next-generation shape memory alloys with optimized functional properties.
|
[
{
"version": "v1",
"created": "Thu, 05 Jun 2025 08:40:09 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:08:54 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Tahseen",
"Mohd",
""
],
[
"Dabade",
"Vivekanand",
""
]
] |
2506.05588
|
Rishona Daniels
|
Rishona Daniels, Duna Wattad, Ronny Ronen, David Saad, and Shahar Kvatinsky
|
Preprocessing Methods for Memristive Reservoir Computing for Image Recognition
|
6 pages, 5 figures, Accepted for presentation in IEEE MetroXRAINE 2025 conference
| null | null | null |
cs.NE cs.AR cs.ET
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Reservoir computing (RC) has attracted attention as an efficient recurrent neural network architecture due to its simplified training, requiring only its last perceptron readout layer to be trained. When implemented with memristors, RC systems benefit from their dynamic properties, which make them ideal for reservoir construction. However, achieving high performance in memristor-based RC remains challenging, as it critically depends on the input preprocessing method and reservoir size. Despite growing interest, a comprehensive evaluation that quantifies the impact of these factors is still lacking. This paper systematically compares various preprocessing methods for memristive RC systems, assessing their effects on accuracy and energy consumption. We also propose a parity-based preprocessing method that improves accuracy by 2-6% while requiring only a modest increase in device count compared to other methods. Our findings highlight the importance of informed preprocessing strategies to improve the efficiency and scalability of memristive RC systems.
|
[
{
"version": "v1",
"created": "Thu, 05 Jun 2025 21:06:10 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 07:55:03 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 14:03:16 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Daniels",
"Rishona",
""
],
[
"Wattad",
"Duna",
""
],
[
"Ronen",
"Ronny",
""
],
[
"Saad",
"David",
""
],
[
"Kvatinsky",
"Shahar",
""
]
] |
2506.05630
|
Niccolo' Di Lalla
|
Niccol\`o Di Lalla, Nicola Omodei, Niccol\`o Bucciantini, Jack T. Dinsmore, Nicol\`o Cibrario, Stefano Silvestri, Josephine Wong, Patrick Slane, Tsunefumi Mizuno, Michela Negro, Roger W. Romani, Riccardo Ferrazzoli, Stephen Chi-Yung Ng, Miltiadis Michailidis, Yi-Jung Yang, Fei Xie, Martin C. Weisskopf, Philip Kaaret, Iv\'an Agudo, L. A. Antonelli, Matteo Bachetti, Luca Baldini, Wayne H. Baumgartner, Ronaldo Bellazzini, Stefano Bianchi, Stephen D. Bongiorno, Raffaella Bonino, Alessandro Brez, Fiamma Capitanio, Simone Castellano, Elisabetta Cavazzuti, Chien-Ting J. Chen, Stefano Ciprini, Enrico Costa, Alessandra De Rosa, Ettore Del Monte, Laura Di Gesu, Alessandro Di Marco, Immacolata Donnarumma, Victor Doroshenko, Michal Dovciak, Steven R. Ehlert, Teruaki Enoto, Yuri Evangelista, Sergio Fabiani, Javier A. Garcia, Shuichi Gunji, Jeremy Heyl, Wataru Iwakiri, Svetlana G. Jorstad, Vladimir Karas, Fabian Kislat, Takao Kitaguchi, Jeffery J. Kolodziejczak, Henric Krawczynski, Fabio La Monaca, Luca Latronico, Ioannis Liodakis, Simone Maldera, Alberto Manfreda, Fre'de'ric Marin, Andrea Marinucci, Alan P. Marscher, Herman L. Marshall, Francesco Massaro, Giorgio Matt, Ikuyuki Mitsuishi, Fabio Muleri, Stephen L. O'Dell, Chiara Oppedisano, Alessandro Papitto, George Pavlov, Abel L. Peirson, Matteo Perri, Melissa Pesce-Rollins, Pierre-Olivier Petrucci, Maura Pilia, Andrea Possenti, Juri Poutanen, Simonetta Puccetti, Brian Ramsey, John Rankin, Ajay Ratheesh, Oliver J. Roberts, Carmelo Sgro, Paolo Soffitta, Gloria Spandre, Douglas A. Swartz, Toru Tamagawa, Fabrizio Tavecchio, Roberto Taverna, Yuzuru Tawara, Allyn F. Tennant, Nicholas Thomas, Francesco Tombesi, Alessio Trois, Sergey S. Tsygankov, Roberto Turolla, Jacco Vink, Kinwah Wu, Silvia Zane
|
X-ray Polarization Detection of the Pulsar Wind Nebula in G21.5-0.9 with IXPE
|
Accepted for publication in ApJ. 15 pages, 6 figures
| null | null | null |
astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
We present the X-ray polarization observation of G21.5-0.9, a young Galactic supernova remnant (SNR), conducted with the Imaging X-ray Polarimetry Explorer (IXPE) in October 2023, with a total livetime of approximately 837 ks. Using different analysis methods, such as a space-integrated study of the entire region of the PWN and a space-resolved polarization map, we detect significant polarization from the pulsar wind nebula (PWN) at the center of the SNR, with an average polarization degree of ~10% oriented at ~33{\deg} (north through east). No significant energy-dependent variation in polarization is observed across the IXPE band (2-8 keV). The polarization map, corrected for the effect of polarization leakage, reveals a consistent pattern in both degree and angle, with little change across the nebula. Our findings indicate the presence of a highly polarized central torus, suggesting low levels of turbulence at particle acceleration sites. Unlike Vela, but similar to the Crab Nebula, we observe substantial differences between radio and X-ray polarization maps. This suggests a clear separation in energy of the emitting particle populations and hints at an important, yet poorly understood, role of instabilities in the turbulence dynamics of PWNe.
|
[
{
"version": "v1",
"created": "Thu, 05 Jun 2025 23:23:23 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 17:54:51 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Di Lalla",
"Niccolò",
""
],
[
"Omodei",
"Nicola",
""
],
[
"Bucciantini",
"Niccolò",
""
],
[
"Dinsmore",
"Jack T.",
""
],
[
"Cibrario",
"Nicolò",
""
],
[
"Silvestri",
"Stefano",
""
],
[
"Wong",
"Josephine",
""
],
[
"Slane",
"Patrick",
""
],
[
"Mizuno",
"Tsunefumi",
""
],
[
"Negro",
"Michela",
""
],
[
"Romani",
"Roger W.",
""
],
[
"Ferrazzoli",
"Riccardo",
""
],
[
"Ng",
"Stephen Chi-Yung",
""
],
[
"Michailidis",
"Miltiadis",
""
],
[
"Yang",
"Yi-Jung",
""
],
[
"Xie",
"Fei",
""
],
[
"Weisskopf",
"Martin C.",
""
],
[
"Kaaret",
"Philip",
""
],
[
"Agudo",
"Iván",
""
],
[
"Antonelli",
"L. A.",
""
],
[
"Bachetti",
"Matteo",
""
],
[
"Baldini",
"Luca",
""
],
[
"Baumgartner",
"Wayne H.",
""
],
[
"Bellazzini",
"Ronaldo",
""
],
[
"Bianchi",
"Stefano",
""
],
[
"Bongiorno",
"Stephen D.",
""
],
[
"Bonino",
"Raffaella",
""
],
[
"Brez",
"Alessandro",
""
],
[
"Capitanio",
"Fiamma",
""
],
[
"Castellano",
"Simone",
""
],
[
"Cavazzuti",
"Elisabetta",
""
],
[
"Chen",
"Chien-Ting J.",
""
],
[
"Ciprini",
"Stefano",
""
],
[
"Costa",
"Enrico",
""
],
[
"De Rosa",
"Alessandra",
""
],
[
"Del Monte",
"Ettore",
""
],
[
"Di Gesu",
"Laura",
""
],
[
"Di Marco",
"Alessandro",
""
],
[
"Donnarumma",
"Immacolata",
""
],
[
"Doroshenko",
"Victor",
""
],
[
"Dovciak",
"Michal",
""
],
[
"Ehlert",
"Steven R.",
""
],
[
"Enoto",
"Teruaki",
""
],
[
"Evangelista",
"Yuri",
""
],
[
"Fabiani",
"Sergio",
""
],
[
"Garcia",
"Javier A.",
""
],
[
"Gunji",
"Shuichi",
""
],
[
"Heyl",
"Jeremy",
""
],
[
"Iwakiri",
"Wataru",
""
],
[
"Jorstad",
"Svetlana G.",
""
],
[
"Karas",
"Vladimir",
""
],
[
"Kislat",
"Fabian",
""
],
[
"Kitaguchi",
"Takao",
""
],
[
"Kolodziejczak",
"Jeffery J.",
""
],
[
"Krawczynski",
"Henric",
""
],
[
"La Monaca",
"Fabio",
""
],
[
"Latronico",
"Luca",
""
],
[
"Liodakis",
"Ioannis",
""
],
[
"Maldera",
"Simone",
""
],
[
"Manfreda",
"Alberto",
""
],
[
"Marin",
"Fre'de'ric",
""
],
[
"Marinucci",
"Andrea",
""
],
[
"Marscher",
"Alan P.",
""
],
[
"Marshall",
"Herman L.",
""
],
[
"Massaro",
"Francesco",
""
],
[
"Matt",
"Giorgio",
""
],
[
"Mitsuishi",
"Ikuyuki",
""
],
[
"Muleri",
"Fabio",
""
],
[
"O'Dell",
"Stephen L.",
""
],
[
"Oppedisano",
"Chiara",
""
],
[
"Papitto",
"Alessandro",
""
],
[
"Pavlov",
"George",
""
],
[
"Peirson",
"Abel L.",
""
],
[
"Perri",
"Matteo",
""
],
[
"Pesce-Rollins",
"Melissa",
""
],
[
"Petrucci",
"Pierre-Olivier",
""
],
[
"Pilia",
"Maura",
""
],
[
"Possenti",
"Andrea",
""
],
[
"Poutanen",
"Juri",
""
],
[
"Puccetti",
"Simonetta",
""
],
[
"Ramsey",
"Brian",
""
],
[
"Rankin",
"John",
""
],
[
"Ratheesh",
"Ajay",
""
],
[
"Roberts",
"Oliver J.",
""
],
[
"Sgro",
"Carmelo",
""
],
[
"Soffitta",
"Paolo",
""
],
[
"Spandre",
"Gloria",
""
],
[
"Swartz",
"Douglas A.",
""
],
[
"Tamagawa",
"Toru",
""
],
[
"Tavecchio",
"Fabrizio",
""
],
[
"Taverna",
"Roberto",
""
],
[
"Tawara",
"Yuzuru",
""
],
[
"Tennant",
"Allyn F.",
""
],
[
"Thomas",
"Nicholas",
""
],
[
"Tombesi",
"Francesco",
""
],
[
"Trois",
"Alessio",
""
],
[
"Tsygankov",
"Sergey S.",
""
],
[
"Turolla",
"Roberto",
""
],
[
"Vink",
"Jacco",
""
],
[
"Wu",
"Kinwah",
""
],
[
"Zane",
"Silvia",
""
]
] |
2506.05750
|
Indu Dihingia Kalpa
|
Indu K. Dihingia, Akhil Uniyal, and Yosuke Mizuno
|
Imprints of Different Types of Low-Angular-Momentum Accretion Flow Solutions in General Relativistic Hydrodynamic Simulations
|
12 pages, 7 figures; Accepted for publication ApJ
| null | null | null |
astro-ph.HE
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Depending on the astrophysical source and its environment, the accretion flows can exhibit a variety of behaviors and characteristics in accordance with the type of solutions. We study low-angular-momentum accretion flows onto black holes using two-dimensional general relativistic hydrodynamic (GRHD) simulations to find imprints of different types of accretion solutions. Such flows, relevant to X-ray binaries and wind-fed low-luminosity active galactic nuclei, often lack sufficient angular momentum to form standard accretion disks. We initialize simulations with semi-analytical transonic solutions defined by specific energy (${\cal E}_0$) and angular momentum ($\lambda_0$), allowing a systematic classification of flow types with: (i) an outer sonic point, (ii) an inner sonic point, and (iii) both, exhibiting shock transitions. Only solutions with two sonic points produce hot, thermally driven bipolar jets/outflows with Lorentz factors up to $\gamma\sim2$, despite the absence of magnetic fields. Using a general relativistic radiation transfer calculation, we compute broadband spectra and images at X-ray ($1 \, \rm keV$) from bremsstrahlung emission. Radiative properties depend strongly on the type of accretion solution. Solutions with inner sonic points produce the brightest and most extended X-ray emission, while outer-point solutions produce compact, fainter signals. These multidimensional models are thus essential for predicting radiative signatures and will enable the development of semi-analytical tools for interpreting X-ray binaries and possibly Sgr~A$^*$ in weak magnetic field regimes.
|
[
{
"version": "v1",
"created": "Fri, 06 Jun 2025 05:20:18 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 01:48:50 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dihingia",
"Indu K.",
""
],
[
"Uniyal",
"Akhil",
""
],
[
"Mizuno",
"Yosuke",
""
]
] |
2506.05898
|
Kenan Turbic
|
Kenan Turbic and Slawomir Stanczak
|
On Level Crossings and Fade Durations in von Mises-Fisher Scattering Channels
|
Accepted for presentation at WSA 2025 (Track 2)
| null | null | null |
eess.SP stat.AP
|
http://creativecommons.org/licenses/by/4.0/
|
This paper investigates the second-order statistics of multipath fading channels with von Mises-Fisher (vMF) distributed scatters. Simple closed-form expressions for the mean Doppler shift and Doppler spread are derived as the key spectral moments that capture the impact of mobility and scattering characteristics on level crossings and fade durations. These expressions are then used to analyze the influence of vMF parameters on the Level-Crossing Rate (LCR) and Average Fade Duration (AFD). The results show that isotropic scattering yields the highest LCR and the lowest AFD, while fading dynamics reduce with the decreasing angular spread of scatterers. Moreover, mobile antenna motion parallel to the mean scattering direction results in a lower LCR than the perpendicular motion, with the difference between the two cases increasing with the higher concentration of scatterers.
|
[
{
"version": "v1",
"created": "Fri, 06 Jun 2025 09:11:35 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 00:21:36 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Turbic",
"Kenan",
""
],
[
"Stanczak",
"Slawomir",
""
]
] |
2506.07106
|
Samir Abdaljalil
|
Samir Abdaljalil, Hasan Kurban, Khalid Qaraqe, Erchin Serpedin
|
Theorem-of-Thought: A Multi-Agent Framework for Abductive, Deductive, and Inductive Reasoning in Language Models
|
ACL 2025 KnowFM
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Large language models (LLMs) have shown strong performance across natural language reasoning tasks, yet their reasoning processes remain brittle and difficult to interpret. Prompting techniques like Chain-of-Thought (CoT) enhance reliability by eliciting intermediate reasoning steps or aggregating multiple outputs. However, they lack mechanisms for enforcing logical structure and assessing internal coherence. We introduce Theorem-of-Thought (ToTh), a novel framework that models reasoning as collaboration among three parallel agents, each simulating a distinct mode of inference: abductive, deductive, and inductive. Each agent produces a reasoning trace, which is structured into a formal reasoning graph. To evaluate consistency, we apply Bayesian belief propagation guided by natural language inference (NLI), assigning confidence scores to each step. The most coherent graph is selected to derive the final answer. Experiments on symbolic (WebOfLies) and numerical (MultiArith) reasoning benchmarks show that ToTh consistently outperforms CoT, Self-Consistency, and CoT-Decoding across multiple LLMs, while producing interpretable and logically grounded reasoning chains. Our findings suggest a promising direction for building more robust and cognitively inspired LLM reasoning. The implementation is available at https://github.com/KurbanIntelligenceLab/theorem-of-thought.
|
[
{
"version": "v1",
"created": "Sun, 08 Jun 2025 12:28:38 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:33:35 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Abdaljalil",
"Samir",
""
],
[
"Kurban",
"Hasan",
""
],
[
"Qaraqe",
"Khalid",
""
],
[
"Serpedin",
"Erchin",
""
]
] |
2506.07287
|
Alexander Veretennikov
|
Alexander Veretennikov and Aisha Nurieva
|
On conditions for Dobrushin's Central limit theorem for non-homogeneous Markov chains
|
9 pages, 16 references
| null | null | null |
math.PR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A new sufficient condition is proposed in the problem of Central Limit Theorem in the array scheme for non-homogeneous Markov Chains.
|
[
{
"version": "v1",
"created": "Sun, 08 Jun 2025 21:11:34 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 17:12:41 GMT"
},
{
"version": "v3",
"created": "Fri, 11 Jul 2025 15:04:43 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 09:59:14 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Veretennikov",
"Alexander",
""
],
[
"Nurieva",
"Aisha",
""
]
] |
2506.07528
|
Qisheng Hu
|
Qisheng Hu, Quanyu Long, Wenya Wang
|
Coordinating Search-Informed Reasoning and Reasoning-Guided Search in Claim Verification
|
Work in progress
| null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Multi-hop claim verification is inherently challenging, requiring multi-step reasoning to construct verification chains while iteratively searching for information to uncover hidden bridging facts. This process is fundamentally interleaved, as effective reasoning relies on dynamically retrieved evidence, while effective search demands reasoning to refine queries based on partial information. To achieve this, we propose Hierarchical Agent Reasoning and Information Search (HARIS), explicitly modeling the coordinated process of reasoning-driven searching and search-informed reasoning. HARIS consists of a high-level reasoning agent that focuses on constructing the main verification chain, generating factual questions when more information is needed, and a low-level search agent that iteratively retrieves more information, refining its search based on intermediate findings. This design allows each agent to specialize in its respective task, enhancing verification accuracy and interpretability. HARIS is trained using reinforcement learning with outcome-based rewards. Experimental results on the EX-FEVER and HOVER benchmarks demonstrate that HARIS achieves strong performance, greatly advancing multi-hop claim verification.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 08:11:43 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 17:12:54 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Hu",
"Qisheng",
""
],
[
"Long",
"Quanyu",
""
],
[
"Wang",
"Wenya",
""
]
] |
2506.08050
|
Takuya Sakasai
|
Carl-Fredrik Nyberg-Brodda, Takuya Sakasai, Yuuki Tadokoro, Kokoro Tanaka
|
On the structure of groups defined by Kim and Manturov
|
23pages, one reference added
| null | null | null |
math.GR math.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study the structure of a series of groups $\Gamma_n^4$ defined by Kim and Manturov. We show that the groups are finite for all $n \ge 6$ and in fact they are 2-step nilpotent $2$-groups.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 00:46:22 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:16:53 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Nyberg-Brodda",
"Carl-Fredrik",
""
],
[
"Sakasai",
"Takuya",
""
],
[
"Tadokoro",
"Yuuki",
""
],
[
"Tanaka",
"Kokoro",
""
]
] |
2506.08184
|
Chupei Wang
|
Chupei Wang (University of Virginia), Jiaqiu Vince Sun (New York University)
|
Unable to Forget: Proactive Interference Reveals Working Memory Limits in LLMs Beyond Context Length
|
Accepted at ICML 2025 Workshop on Long Context Foundation Models (ICFM). Code: https://github.com/zhuangziGiantfish/Unable-to-Forget
| null | null | null |
cs.CL cs.AI q-bio.NC
|
http://creativecommons.org/licenses/by/4.0/
|
Information retrieval in Large Language Models (LLMs) is increasingly recognized as intertwined with generation capabilities rather than mere lookup. While longer contexts are often assumed to improve retrieval, the effects of intra-context interference remain understudied. To address this, we adapt the proactive interference (PI) paradigm from cognitive science, where earlier information disrupts recall of newer updates. In humans, susceptibility to such interference is inversely linked to working memory capacity. We introduce PI-LLM, an evaluation that sequentially streams semantically related key-value updates and queries only the final values. Although these final values are clearly positioned just before the query, LLM retrieval accuracy declines log-linearly toward zero as interference accumulates; errors arise from retrieving previously overwritten values. Attempts to mitigate interference via prompt engineering (e.g., instructing models to ignore earlier input) yield limited success. These findings reveal a fundamental constraint on LLMs' ability to disentangle interference and flexibly manipulate information, suggesting a working memory bottleneck beyond mere context access. This calls for approaches that strengthen models' ability to suppress irrelevant content during retrieval.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 19:49:11 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Jun 2025 09:53:49 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 16:45:51 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Chupei",
"",
"University of Virginia"
],
[
"Sun",
"Jiaqiu Vince",
"",
"New York University"
]
] |
2506.08193
|
Vasilis Oikonomou
|
S.D. Odintsov, V.K. Oikonomou
|
GW170817 Viable Einstein-Gauss-Bonnet Inflation Compatible with the Atacama Cosmology Telescope Data
|
PLB in press
| null | null | null |
gr-qc astro-ph.CO hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this work we investigate several Einstein-Gauss-Bonnet models that are compatible with the GW170817 event, the Atacama Cosmology Telescope data and the BICEP/Keck updated Planck constraints on the tensor-to-scalar ratio. We consider two distinct classes of Einstein-Gauss-Bonnet theories, which are equally successful for GW170817-compatible model building and we examine their viability against the Atacama Cosmology Telescope data and the updated Planck constraints on the tensor-to-scalar ratio. The two models are distinct since the first class relates directly the non-minimal Gauss-Bonnet scalar coupling function with the scalar potential and yields $c_T^2=1$ while in the second class, the non-minimal Gauss-Bonnet scalar coupling function and the scalar potential are freely but conveniently chosen and the class of models respects the constraint $\left| c_T^2 - 1 \right| < 6 \times 10^{-15}$. We provide several examples of models belonging to both the two classes of GW170817-compatible Einstein-Gauss-Bonnet theories and we demonstrate that Einstein-Gauss-Bonnet theories provide a promising theoretical framework for inflationary dynamics.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 20:08:22 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 20:09:57 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Odintsov",
"S. D.",
""
],
[
"Oikonomou",
"V. K.",
""
]
] |
2506.10006
|
Wei Yang
|
Jie Qin, Wei Yang, Yan Su, Yiran Zhu, Weizhen Li, Yunyue Pan, Chengchang Pan, Honggang Qi
|
HER2 Expression Prediction with Flexible Multi-Modal Inputs via Dynamic Bidirectional Reconstruction
|
8 pages,6 figures,3 tables,accepted by the 33rd ACM International Conference on Multimedia(ACM MM 2025)
| null | null | null |
cs.MM cs.AI cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
In breast cancer HER2 assessment, clinical evaluation relies on combined H&E and IHC images, yet acquiring both modalities is often hindered by clinical constraints and cost. We propose an adaptive bimodal prediction framework that flexibly supports single- or dual-modality inputs through two core innovations: a dynamic branch selector activating modality completion or joint inference based on input availability, and a cross-modal GAN (CM-GAN) enabling feature-space reconstruction of missing modalities. This design dramatically improves H&E-only accuracy from 71.44% to 94.25%, achieves 95.09% with full dual-modality inputs, and maintains 90.28% reliability under single-modality conditions. The "dual-modality preferred, single-modality compatible" architecture delivers near-dual-modality accuracy without mandatory synchronized acquisition, offering a cost-effective solution for resource-limited regions and significantly improving HER2 assessment accessibility.
|
[
{
"version": "v1",
"created": "Sat, 12 Apr 2025 11:24:06 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:57:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Qin",
"Jie",
""
],
[
"Yang",
"Wei",
""
],
[
"Su",
"Yan",
""
],
[
"Zhu",
"Yiran",
""
],
[
"Li",
"Weizhen",
""
],
[
"Pan",
"Yunyue",
""
],
[
"Pan",
"Chengchang",
""
],
[
"Qi",
"Honggang",
""
]
] |
2506.10129
|
Lucas Bonomo
|
Lucas A Bonomo, Edward J Brambley, Julio A Cordioli
|
Effects of the sheared flow velocity profile on impedance eduction in a 2D duct
|
25 pages, Preprint submitted to Acta Acustica
| null | null | null |
physics.flu-dyn
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Impedance eduction methods are the current standard approach to access the impedance of acoustic liner under sheared grazing flow. The dedicated facilities for these methods consists on a waveguide with rectangular cross-section, which implies in a sheared grazing flow. A current debate in the literature is the effect of this sheared flow in the impedance eduction methods. We assess the impact of the flow profile shape on acoustic propagation in a two-dimensional duct within the typical operating range of impedance eduction facilities. Firstly, a numerical experiment is proposed in which the Pridmore--Brown equation is assumed to represent the true physical behaviour, and is used with both simplified flow profiles commonly used in the literature and a realistic representation of a turbulent boundary layer using a van Driest universal law of the wall model. The data from these numerical experiments are then used with a traditional impedance eduction process, and the resulting variation in obtained impedances are investigated. Secondly, we apply a less-traditional impedance eduction method that incorporates the sheared velocity profile to data obtained from real-world experiments. The results suggest that the Ingard--Myers boundary condition remains a good approximation to a realistic boundary layer profile, such as the universal law of the wall, at least in the two-dimensional case. However, it is also shown that the simplified flow profiles often used in the literature can lead to significant deviations from the results obtained using a realistic velocity distribution.
|
[
{
"version": "v1",
"created": "Wed, 11 Jun 2025 19:18:05 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:35:56 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bonomo",
"Lucas A",
""
],
[
"Brambley",
"Edward J",
""
],
[
"Cordioli",
"Julio A",
""
]
] |
2506.10619
|
Xiong-Hui Cao
|
Xiong-Hui Cao, Feng-Kun Guo, Zhi-Hui Guo and Qu-Zhi Li
|
Revisiting Roy-Steiner-equation analysis of pion-kaon scattering from lattice QCD data
|
66 pages, 19 figures; discussions added; revised version published in Phys.Rev.D
| null | null | null |
hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A comprehensive analysis of $\pi K\rightarrow \pi K$ and $\pi\pi\rightarrow K\bar K$ amplitudes at large unphysical pion mass for all important partial waves is presented. A set of crossing-symmetric partial-wave hyperbolic dispersion relations is used to describe lattice QCD data at $m_\pi=391$ MeV. In the present analysis, the amplitudes for the $S$- and $P$-waves are formulated by combining the constraints of analyticity, unitarity, and crossing symmetry, fulfilling Roy-Steiner-type equations. We use these results to investigate the low-lying strange-meson resonances and resolve the instability problem tied to analytic continuation in prior lattice QCD studies based on the $K$-matrix formalism. At $m_\pi=391$ MeV, the rigorous Roy-Steiner-type equation approach allows us to determine the $S$-wave scattering lengths, $m_\pi a_0^{1/2}=\left(0.92_{-0.28}^{+0.06}\right)$, $m_\pi a_0^{3/2}=-\left(0.32_{-0.02}^{+0.05}\right)$, and the $\kappa$ (also known as $K_0^*(700)$) pole position, $\sqrt{s_\kappa}=\left(966_{-24}^{+41}-i 198_{-17}^{+38}\right)$ MeV. We also provide a detailed analysis of the complex validity domain of the Roy-Steiner-type equations.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 12:02:46 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:39:30 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cao",
"Xiong-Hui",
""
],
[
"Guo",
"Feng-Kun",
""
],
[
"Guo",
"Zhi-Hui",
""
],
[
"Li",
"Qu-Zhi",
""
]
] |
2506.11290
|
Viktor Stasenko
|
Viktor Stasenko
|
Merger History of Clustered Primordial Black Holes
|
9 pages, 4 figures
| null | null | null |
astro-ph.CO gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Primordial black hole (PBH) binaries experience strong gravitational perturbations in the case of their initial clustering, which significantly affects the dynamics of their mergers. In this work, we develop a new formalism to account for these perturbations and track the evolution of the binary orbital parameters distribution. Based on this approach, we calculate the merger rate of PBH binaries and demonstrate that its temporal evolution differs greatly from that of isolated binary systems. Moreover, PBH clustering produces distinctive features in the stochastic gravitational-wave background: the canonical $2/3$ spectral slope transforms to $\Omega_{\rm gw} \propto \nu^{-65/28}$ in a certain frequency band. These predictions can be probed in future gravitational wave observations, opening up new opportunities to test the clustering of PBHs and their contribution to dark matter.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 20:46:45 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 18:57:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Stasenko",
"Viktor",
""
]
] |
2506.11298
|
Piotr Sowinski
|
Piotr Sowinski and Karolina Bogacka and Anastasiya Danilenka and Nikita Kozlov
|
Jelly: a Fast and Convenient RDF Serialization Format
|
Developers Workshop, co-located with SEMANTiCS'25: International Conference on Semantic Systems, September 3-5, 2025, Vienna, Austria
| null | null | null |
cs.DB cs.NI
|
http://creativecommons.org/licenses/by/4.0/
|
Existing RDF serialization formats such as Turtle, N-Quads, and JSON-LD are widely used for communication and storage in knowledge graph and Semantic Web applications. However, they suffer from limitations in performance, compression ratio, and lack of native support for RDF streams. To address these shortcomings, we introduce Jelly, a fast and convenient binary serialization format for RDF data that supports both batch and streaming use cases. Jelly is designed to maximize serialization throughput, reduce file size with lightweight streaming compression, and minimize compute resource usage. Built on Protocol Buffers, Jelly is easy to integrate with modern programming languages and RDF libraries. To maximize reusability, Jelly has an open protocol specification, open-source implementations in Java and Python integrated with popular RDF libraries, and a versatile command-line tool. To illustrate its usefulness, we outline concrete use cases where Jelly can provide tangible benefits. We consider that by combining practical usability with state-of-the-art efficiency, Jelly is an important contribution to the Semantic Web tool stack.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 20:58:39 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:49:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Sowinski",
"Piotr",
""
],
[
"Bogacka",
"Karolina",
""
],
[
"Danilenka",
"Anastasiya",
""
],
[
"Kozlov",
"Nikita",
""
]
] |
2506.12284
|
Thomas Walker
|
Thomas Walker, Ahmed Imtiaz Humayun, Randall Balestriero, Richard Baraniuk
|
GrokAlign: Geometric Characterisation and Acceleration of Grokking
|
23 pages, 11 figures, 3 tables
| null | null | null |
cs.LG stat.ML
|
http://creativecommons.org/licenses/by/4.0/
|
A key challenge for the machine learning community is to understand and accelerate the training dynamics of deep networks that lead to delayed generalisation and emergent robustness to input perturbations, also known as grokking. Prior work has associated phenomena like delayed generalisation with the transition of a deep network from a linear to a feature learning regime, and emergent robustness with changes to the network's functional geometry, in particular the arrangement of the so-called linear regions in deep networks employing continuous piecewise affine nonlinearities. Here, we explain how grokking is realised in the Jacobian of a deep network and demonstrate that aligning a network's Jacobians with the training data (in the sense of cosine similarity) ensures grokking under a low-rank Jacobian assumption. Our results provide a strong theoretical motivation for the use of Jacobian regularisation in optimizing deep networks -- a method we introduce as GrokAlign -- which we show empirically to induce grokking much sooner than more conventional regularizers like weight decay. Moreover, we introduce centroid alignment as a tractable and interpretable simplification of Jacobian alignment that effectively identifies and tracks the stages of deep network training dynamics. Accompanying webpage (https://thomaswalker1.github.io/blog/grokalign.html) and code (https://github.com/ThomasWalker1/grokalign).
|
[
{
"version": "v1",
"created": "Sat, 14 Jun 2025 00:16:11 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:15:21 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Walker",
"Thomas",
""
],
[
"Humayun",
"Ahmed Imtiaz",
""
],
[
"Balestriero",
"Randall",
""
],
[
"Baraniuk",
"Richard",
""
]
] |
2506.12365
|
Asifullah Khan
|
Asifullah Khan, Muhammad Zaeem Khan, Saleha Jamshed, Sadia Ahmad, Aleesha Zainab, Kaynat Khatib, Faria Bibi, and Abdul Rehman
|
Advances in LLMs with Focus on Reasoning, Adaptability, Efficiency and Ethics
| null | null | null | null |
cs.CL cs.DB
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
This survey paper outlines the key developments in the field of Large Language Models (LLMs), including enhancements to their reasoning skills, adaptability to various tasks, increased computational efficiency, and the ability to make ethical decisions. The techniques that have been most effective in bridging the gap between human and machine communications include the Chain-of-Thought prompting, Instruction Tuning, and Reinforcement Learning from Human Feedback. The improvements in multimodal learning and few-shot or zero-shot techniques have further empowered LLMs to handle complex jobs with minor input. A significant focus is placed on efficiency, detailing scaling strategies, optimization techniques, and the influential Mixture-of-Experts (MoE) architecture, which strategically routes inputs to specialized subnetworks to boost predictive accuracy, while optimizing resource allocation. This survey also offers a broader perspective on recent advancements in LLMs, going beyond isolated aspects such as model architecture or ethical concerns. Additionally, it explores the role of LLMs in Agentic AI and their use as Autonomous Decision-Making Systems, and categorizes emerging methods that enhance LLM reasoning, efficiency, and ethical alignment. The survey also identifies underexplored areas such as interpretability, cross-modal integration, and sustainability. While significant advancements have been made in LLMs, challenges such as high computational costs, biases, and ethical risks remain. Overcoming these requires a focus on bias mitigation, transparent decision-making, and explicit ethical guidelines. Future research will generally focus on enhancing the model's ability to handle multiple inputs, thereby making it more intelligent, safe, and reliable.
|
[
{
"version": "v1",
"created": "Sat, 14 Jun 2025 05:55:19 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:09:27 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Khan",
"Asifullah",
""
],
[
"Khan",
"Muhammad Zaeem",
""
],
[
"Jamshed",
"Saleha",
""
],
[
"Ahmad",
"Sadia",
""
],
[
"Zainab",
"Aleesha",
""
],
[
"Khatib",
"Kaynat",
""
],
[
"Bibi",
"Faria",
""
],
[
"Rehman",
"Abdul",
""
]
] |
2506.13531
|
Quinlan Lee
|
Christian Gourieroux and Quinlan Lee
|
Identification of Impulse Response Functions for Nonlinear Dynamic Models
|
36 pages, 1 figure; This submission is a split from a previous version posted as [arXiv:2305.18145], with some overlapping content in the background and methods
| null | null | null |
econ.EM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We explore the issues of identification for nonlinear Impulse Response Functions in nonlinear dynamic models and discuss the settings in which the problem can be mitigated. In particular, we introduce the nonlinear autoregressive representation with Gaussian innovations and characterize the identified set. This set arises from the multiplicity of nonlinear innovations and transformations which leave invariant the standard normal density. We then discuss possible identifying restrictions, such as non-Gaussianity of independent sources, or identifiable parameters by means of learning algorithms, and the possibility of identification in nonlinear dynamic factor models when the underlying latent factors have different dynamics. We also explain how these identification results depend ultimately on the set of series under consideration.
|
[
{
"version": "v1",
"created": "Mon, 16 Jun 2025 14:22:53 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 02:29:13 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gourieroux",
"Christian",
""
],
[
"Lee",
"Quinlan",
""
]
] |
2506.14662
|
Young-Ho Cho
|
Young-ho Cho, Min-Seung Ko, Hao Zhu
|
PGLib-CO2: A Power Grid Library for Computing and Optimizing Carbon Emissions
| null | null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by/4.0/
|
A sustainable electricity infrastructure requires the explicit integration of carbon emissions into power system modeling and optimization paradigms. However, existing open-source datasets for power system R&D lack generator-level carbon emission profiling, limiting the ability to benchmark and compare various carbon-aware grid operational strategies. To address this gap, this work introduces PGLib-CO2, an open-source extension to the widely adopted PGLib-OPF test case library. PGLib-CO2 enriches standard network cases with CO2 and CO2-equivalent emission intensity factors by expanding the fuel-type categorization used by PGLib-OPF, attaining a realistic generator-level carbon profiling. It is also packaged for both Python's pandapower and Julia's PowerModels.jl, for a seamless, user-friendly integration of emission modeling into grid computation and optimization tasks. The dataset produced by PGLib-CO2 can support grid-based carbon accounting, emission metric evaluation, and integration into AC optimal power flow (OPF) and optimal load shifting (OLS) formulations. We demonstrate PGLib-CO2's utility through case studies that quantify cost-emission trade-offs and optimize a carbon-aware objective function. By standardizing carbon-enhanced test cases, PGLib-CO2 provides an open-source, reproducible foundation for benchmarking carbon-aware computation, facilitating future research in sustainable power system operation.
|
[
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:52:47 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2025 19:41:39 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 17:33:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cho",
"Young-ho",
""
],
[
"Ko",
"Min-Seung",
""
],
[
"Zhu",
"Hao",
""
]
] |
2506.14781
|
Corentin Delacour
|
Corentin Delacour, M Mahmudul Hasan Sajeeb, Joao P. Hespanha, Kerem Y. Camsari
|
Two-dimensional Parallel Tempering for Constrained Optimization
|
Added references in Introduction
|
Physical Review E (2025)
|
10.1103/mr2n-qqrb
| null |
cs.LG cond-mat.stat-mech math.OC stat.ML
|
http://creativecommons.org/licenses/by/4.0/
|
Sampling Boltzmann probability distributions plays a key role in machine learning and optimization, motivating the design of hardware accelerators such as Ising machines. While the Ising model can in principle encode arbitrary optimization problems, practical implementations are often hindered by soft constraints that either slow down mixing when too strong, or fail to enforce feasibility when too weak. We introduce a two-dimensional extension of the powerful parallel tempering algorithm (PT) that addresses this challenge by adding a second dimension of replicas interpolating the penalty strengths. This scheme ensures constraint satisfaction in the final replicas, analogous to low-energy states at low temperature. The resulting two-dimensional parallel tempering algorithm (2D-PT) improves mixing in heavily constrained replicas and eliminates the need to explicitly tune the penalty strength. In a representative example of graph sparsification with copy constraints, 2D-PT achieves near-ideal mixing, with Kullback-Leibler divergence decaying as O(1/t). When applied to sparsified Wishart instances, 2D-PT yields orders of magnitude speedup over conventional PT with the same number of replicas. The method applies broadly to constrained Ising problems and can be deployed on existing Ising machines.
|
[
{
"version": "v1",
"created": "Sat, 24 May 2025 20:41:45 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 19:09:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Delacour",
"Corentin",
""
],
[
"Sajeeb",
"M Mahmudul Hasan",
""
],
[
"Hespanha",
"Joao P.",
""
],
[
"Camsari",
"Kerem Y.",
""
]
] |
2506.15692
|
Jaehyun Nam
|
Jaehyun Nam, Jinsung Yoon, Jiefeng Chen, Jinwoo Shin, Sercan \"O. Ar{\i}k, Tomas Pfister
|
MLE-STAR: Machine Learning Engineering Agent via Search and Targeted Refinement
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Agents based on large language models (LLMs) for machine learning engineering (MLE) can automatically implement ML models via code generation. However, existing approaches to build such agents often rely heavily on inherent LLM knowledge and employ coarse exploration strategies that modify the entire code structure at once. This limits their ability to select effective task-specific models and perform deep exploration within specific components, such as experimenting extensively with feature engineering options. To overcome these, we propose MLE-STAR, a novel approach to build MLE agents. MLE-STAR first leverages external knowledge by using a search engine to retrieve effective models from the web, forming an initial solution, then iteratively refines it by exploring various strategies targeting specific ML components. This exploration is guided by ablation studies analyzing the impact of individual code blocks. Furthermore, we introduce a novel ensembling method using an effective strategy suggested by MLE-STAR. Our experimental results show that MLE-STAR achieves medals in 64% of the Kaggle competitions on the MLE-bench Lite, significantly outperforming the best alternative.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 18:11:25 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 21:58:41 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Nam",
"Jaehyun",
""
],
[
"Yoon",
"Jinsung",
""
],
[
"Chen",
"Jiefeng",
""
],
[
"Shin",
"Jinwoo",
""
],
[
"Arık",
"Sercan Ö.",
""
],
[
"Pfister",
"Tomas",
""
]
] |
2506.17114
|
Jiayu Liu
|
Dadi Guo, Jiayu Liu, Zhiyuan Fan, Zhitao He, Haoran Li, Yumeng Wang, Yi R. Fung
|
Mathematical Proof as a Litmus Test: Revealing Failure Modes of Advanced Large Reasoning Models
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large reasoning models (e.g., R1, o3) have demonstrated remarkable mathematical problem-solving abilities. However, the high reported accuracy of these advanced models on popular datasets, reliance on purely numerical evaluation and potential benchmark leakage, often masks their true reasoning shortcomings. To address this, we propose leveraging the inherent rigor and methodological complexity of mathematical proofs as a diagnostic tool to expose these hidden failures. Specifically, we introduce the RFMDataset (Reveal Failure Modes), a collection of 200 diverse mathematical proof problems, and thoroughly evaluate advanced models' performance on it. Our in-depth analysis of their failures uncovers 10 fine-grained error types, which shows fundamental limitations in current large reasoning models: 1) large reasoning models grapple profoundly with mathematical proofs, with some generating entirely correct proofs for less than 20% of problems and failing even on basic ones; 2) models exhibit a diverse spectrum of reasoning failures, prominently demonstrating the lack of guarantees for the correctness and rigor of single-step reasoning; and 3) models show hallucination and incompleteness during the reasoning process. Our findings reveal that models' self-reflection is insufficient to resolve the current logical dilemmas, necessitating formalized and fine-grained logical training.
|
[
{
"version": "v1",
"created": "Fri, 20 Jun 2025 16:14:18 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2025 04:11:44 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 16:23:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Guo",
"Dadi",
""
],
[
"Liu",
"Jiayu",
""
],
[
"Fan",
"Zhiyuan",
""
],
[
"He",
"Zhitao",
""
],
[
"Li",
"Haoran",
""
],
[
"Wang",
"Yumeng",
""
],
[
"Fung",
"Yi R.",
""
]
] |
2506.17176
|
Pierfrancesco Guarino
|
Pierfrancesco Guarino and Gabriel Ziegler
|
Capturing Misalignment
| null | null | null | null |
econ.TH
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
We introduce and formalize misalignment, a phenomenon of interactive environments perceived from an analyst's perspective where an agent holds beliefs about another agent's beliefs that do not correspond to the actual beliefs of the latter. We demonstrate that standard frameworks, such as type structures, fail to capture misalignment, necessitating new tools to analyze this phenomenon. To this end, we characterize misalignment through non-belief-closed state spaces and introduce agent-dependent type structures, which provide a flexible tool to understand the varying degrees of misalignment. Furthermore, we establish that appropriately adapted modal operators on agent-dependent type structures behave consistently with standard properties, enabling us to explore the implications of misalignment for interactive reasoning. Finally, we show how speculative trade can arise under misalignment, even when imposing the corresponding assumptions that rule out such trades in standard environments.
|
[
{
"version": "v1",
"created": "Fri, 20 Jun 2025 17:29:34 GMT"
},
{
"version": "v2",
"created": "Tue, 08 Jul 2025 20:12:39 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 20:10:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Guarino",
"Pierfrancesco",
""
],
[
"Ziegler",
"Gabriel",
""
]
] |
2506.17247
|
Yiting Liu
|
Andrew B. Kahng, Yiting Liu, Zhiang Wang
|
Recursive Learning-Based Virtual Buffering for Analytical Global Placement
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Due to the skewed scaling of interconnect versus cell delay in modern technology nodes, placement with buffer porosity (i.e., cell density) awareness is essential for timing closure in physical synthesis flows. However, existing approaches face two key challenges: (i) traditional van Ginneken-Lillis-style buffering approaches are computationally expensive during global placement; and (ii) machine learning-based approaches, such as BufFormer, lack a thorough consideration of Electrical Rule Check (ERC) violations and fail to "close the loop" back into the physical design flow. In this work, we propose MLBuf-RePlAce, the first open-source learning-driven virtual buffering-aware analytical global placement framework, built on top of the OpenROAD infrastructure. MLBuf-RePlAce adopts an efficient recursive learning-based generative buffering approach to predict buffer types and locations, addressing ERC violations during global placement. We compare MLBuf-RePlAce against the default virtual buffering-based timing-driven global placer in OpenROAD, using open-source testcases from the TILOS MacroPlacement and OpenROAD-flow-scripts repositories. Without degradation of post-route power, MLBuf-RePlAce achieves (maximum, average) improvements of (56%, 31%) in total negative slack (TNS) within the open-source OpenROAD flow. When evaluated by completion in a commercial flow, MLBuf-RePlAce achieves (maximum, average) improvements of (53%, 28%) in TNS with an average of 0.2% improvement in post-route power.
|
[
{
"version": "v1",
"created": "Sat, 07 Jun 2025 18:22:31 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 18:51:25 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kahng",
"Andrew B.",
""
],
[
"Liu",
"Yiting",
""
],
[
"Wang",
"Zhiang",
""
]
] |
2506.18029
|
Hans-Peter Schr\"ocker
|
Z\"ulal Derin Yaqub, Hans-Peter Schr\"ocker
|
Rational Motions of Minimal Quaternionic Degree with Prescribed Line Trajectories
| null | null |
10.1016/j.cam.2025.116949
| null |
math.RA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we study how to find rational motions that move a line along a given rational ruled surface. Our goal is to find motions with the lowest possible degree using dual quaternions. While similar problems for point trajectories are well known, the case of line trajectories is more complicated and has not been studied. We explain when such motions exist and how to compute them. Our method gives explicit formulas for constructing these motions and shows that, in many cases, the solution is unique. We also show examples and explain how to use these results to design simple mechanisms that move a line in the desired way. This work helps to better understand the relationship between rational motions and ruled surfaces and may be useful for future research in mechanism design.
|
[
{
"version": "v1",
"created": "Sun, 22 Jun 2025 13:15:22 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Yaqub",
"Zülal Derin",
""
],
[
"Schröcker",
"Hans-Peter",
""
]
] |
2506.18199
|
Bushra Asseri
|
Bushra Asseri, Estabrag Abdelaziz, Areej Al-Wabil
|
Prompt Engineering Techniques for Mitigating Cultural Bias Against Arabs and Muslims in Large Language Models: A Systematic Review
|
Research is incomplete
| null | null | null |
cs.CL cs.AI cs.CY cs.HC
|
http://creativecommons.org/licenses/by/4.0/
|
Large language models have demonstrated remarkable capabilities across various domains, yet concerns about cultural bias - particularly towards Arabs and Muslims - pose significant ethical challenges by perpetuating harmful stereotypes and marginalization. Despite growing recognition of bias in LLMs, prompt engineering strategies specifically addressing Arab and Muslim representation remain understudied. This mixed-methods systematic review examines such techniques, offering evidence-based guidance for researchers and practitioners. Following PRISMA guidelines and Kitchenham's systematic review methodology, we analyzed 8 empirical studies published between 2021-2024 investigating bias mitigation strategies. Our findings reveal five primary prompt engineering approaches: cultural prompting, affective priming, self-debiasing techniques, structured multi-step pipelines, and parameter-optimized continuous prompts. Although all approaches show potential for reducing bias, effectiveness varied substantially across studies and bias types. Evidence suggests that certain bias types may be more resistant to prompt-based mitigation than others. Structured multi-step pipelines demonstrated the highest overall effectiveness, achieving up to 87.7% reduction in bias, though they require greater technical expertise. Cultural prompting offers broader accessibility with substantial effectiveness. These results underscore the accessibility of prompt engineering for mitigating cultural bias without requiring access to model parameters. The limited number of studies identified highlights a significant research gap in this critical area. Future research should focus on developing culturally adaptive prompting techniques, creating Arab and Muslim-specific evaluation resources, and integrating prompt engineering with complementary debiasing methods to address deeper stereotypes while maintaining model utility.
|
[
{
"version": "v1",
"created": "Sun, 22 Jun 2025 23:15:25 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 19:07:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Asseri",
"Bushra",
""
],
[
"Abdelaziz",
"Estabrag",
""
],
[
"Al-Wabil",
"Areej",
""
]
] |
2506.19330
|
Yidi Shao
|
Yidi Shao and Longfei Zhou and Fangshuo Tang and Xinyi Shi and Dalang Chen and Shengtao Xia
|
Comparative Performance of Finetuned ImageNet Pre-trained Models for Electronic Component Classification
|
Due to issues related to author order and some problems in the current version regarding methodology, we would like to withdraw the preprint to avoid potential conflicts
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Electronic component classification and detection are crucial in manufacturing industries, significantly reducing labor costs and promoting technological and industrial development. Pre-trained models, especially those trained on ImageNet, are highly effective in image classification, allowing researchers to achieve excellent results even with limited data. This paper compares the performance of twelve ImageNet pre-trained models in classifying electronic components. Our findings show that all models tested delivered respectable accuracies. MobileNet-V2 recorded the highest at 99.95%, while EfficientNet-B0 had the lowest at 92.26%. These results underscore the substantial benefits of using ImageNet pre-trained models in image classification tasks and confirm the practical applicability of these methods in the electronics manufacturing sector.
|
[
{
"version": "v1",
"created": "Tue, 24 Jun 2025 05:48:39 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:31:54 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Shao",
"Yidi",
""
],
[
"Zhou",
"Longfei",
""
],
[
"Tang",
"Fangshuo",
""
],
[
"Shi",
"Xinyi",
""
],
[
"Chen",
"Dalang",
""
],
[
"Xia",
"Shengtao",
""
]
] |
2506.19737
|
Shuige Liu
|
Shuige Liu, Gabriel Ziegler
|
Reasoning about Bounded Reasoning
| null | null | null | null |
econ.TH
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Interactive decision-making relies on strategic reasoning. Two prominent frameworks are (1) models of bounded reasoning, exemplified by level-$k$ models, which keep reasoning implicit, and (2) epistemic game theory, which makes reasoning explicit. We connect these approaches by "lifting" static complete-information games into incomplete-information settings where payoff types reflect players' reasoning depths as in level-$k$ models. We introduce downward rationalizability, defined via minimal belief restrictions capturing the basic idea common to level-$k$ models, to provide robust and well-founded predictions in games where bounded reasoning matters. We then refine these belief restrictions to analyze the foundations of two seminal models of bounded reasoning: the classic level-$k$ model and the cognitive hierarchy model. Our findings shed light on the distinction between hard cognitive bounds on reasoning and beliefs about co-players' types. Furthermore, they offer insights into robustness issues relevant for market design. Thus, our approach unifies key level-$k$ models building on clear foundations of strategic reasoning stemming from epistemic game theory.
|
[
{
"version": "v1",
"created": "Tue, 24 Jun 2025 15:56:24 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 14:29:44 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 15:10:31 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Liu",
"Shuige",
""
],
[
"Ziegler",
"Gabriel",
""
]
] |
2506.19955
|
Yiming Ma
|
Yiming Ma, Victor Sanchez, Tanaya Guha
|
ZIP: Scalable Crowd Counting via Zero-Inflated Poisson Modeling
|
15 pages, 11 figures
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Most crowd counting methods directly regress blockwise density maps using Mean Squared Error (MSE) losses. This practice has two key limitations: (1) it fails to account for the extreme spatial sparsity of annotations - over 95% of 8x8 blocks are empty across standard benchmarks, so supervision signals in informative regions are diluted by the predominant zeros; (2) MSE corresponds to a Gaussian error model that poorly matches discrete, non-negative count data. To address these issues, we introduce ZIP, a scalable crowd counting framework that models blockwise counts with a Zero-Inflated Poisson likelihood: a zero-inflation term learns the probability a block is structurally empty (handling excess zeros), while the Poisson component captures expected counts when people are present (respecting discreteness). We provide a generalization analysis showing a tighter risk bound for ZIP than MSE-based losses and DMCount provided that the training resolution is moderately large. To assess the scalability of ZIP, we instantiate it on backbones spanning over 100x in parameters/compute. Experiments on ShanghaiTech A & B, UCF-QNRF, and NWPU-Crowd demonstrate that ZIP consistently surpasses state-of-the-art methods across all model scales.
|
[
{
"version": "v1",
"created": "Tue, 24 Jun 2025 19:06:22 GMT"
},
{
"version": "v2",
"created": "Wed, 16 Jul 2025 22:29:43 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 14:46:10 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ma",
"Yiming",
""
],
[
"Sanchez",
"Victor",
""
],
[
"Guha",
"Tanaya",
""
]
] |
2506.20547
|
Jordan Cotler
|
Jordan Cotler, Kristan Jensen
|
Norm of the no-boundary state
|
38+10 pages, 6 figures; v2: results generalized, typos corrected, references added
| null | null | null |
hep-th gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We consider Einstein gravity with positive cosmological constant coupled to matter in an asymptotically de Sitter universe with sphere boundary at timelike infinity. In this setting we show that, to one-loop order and at late time, the norm of the no-boundary state vanishes, going as $e^{S_0} \frac{Z S_0^{-d(d+1)/4}}{\text{vol}(SO(d,1))}$ with $S_0$ the tree-level entropy of the static patch, $d$ the spacetime dimension, and $Z$ non-negative. We show that the presence of an observer stabilizes the norm to a large, positive value.
|
[
{
"version": "v1",
"created": "Wed, 25 Jun 2025 15:44:56 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:52:08 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cotler",
"Jordan",
""
],
[
"Jensen",
"Kristan",
""
]
] |
2506.20947
|
Dejie Yang
|
Dejie Yang and Zhu Xu and Xinjie Gao and Yang Liu
|
Hierarchical Sub-action Tree for Continuous Sign Language Recognition
| null |
ICME 2025
| null | null |
cs.CV cs.MM
|
http://creativecommons.org/licenses/by/4.0/
|
Continuous sign language recognition (CSLR) aims to transcribe untrimmed videos into glosses, which are typically textual words. Recent studies indicate that the lack of large datasets and precise annotations has become a bottleneck for CSLR due to insufficient training data. To address this, some works have developed cross-modal solutions to align visual and textual modalities. However, they typically extract textual features from glosses without fully utilizing their knowledge. In this paper, we propose the Hierarchical Sub-action Tree (HST), termed HST-CSLR, to efficiently combine gloss knowledge with visual representation learning. By incorporating gloss-specific knowledge from large language models, our approach leverages textual information more effectively. Specifically, we construct an HST for textual information representation, aligning visual and textual modalities step-by-step and benefiting from the tree structure to reduce computational complexity. Additionally, we impose a contrastive alignment enhancement to bridge the gap between the two modalities. Experiments on four datasets (PHOENIX-2014, PHOENIX-2014T, CSL-Daily, and Sign Language Gesture) demonstrate the effectiveness of our HST-CSLR.
|
[
{
"version": "v1",
"created": "Thu, 26 Jun 2025 02:27:50 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Yang",
"Dejie",
""
],
[
"Xu",
"Zhu",
""
],
[
"Gao",
"Xinjie",
""
],
[
"Liu",
"Yang",
""
]
] |
2506.21292
|
Oleksii Ivanytskyi
|
Oleksii Ivanytskyi, David Blaschke, Gerd R\"opke
|
Speed of sound of QCD matter at chiral crossover
|
7 pages, 2 figuer, contribution to the Special Issue "Aspects of Criticality II"
|
J.Subatomic Part.Cosmol. 4 (2025) 100096
|
10.1016/j.jspc.2025.100096
| null |
nucl-th hep-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Based on a generalized Beth-Uhlenbeck approach to thermodynamics of QCD motivated by cluster decomposition we present a unified equation of state of hot strongly interacting matter and analyze its properties in a wide range of temperatures. The hadrons are treated as color singlet multiquark clusters in medium with a background gluon field in the Polyakov gauge. The confining aspect of QCD is accounted for by the Polyakov loop mechanism and by a large vacuum quark mass motivated by a confining density functional approach. We demonstrate that an abrupt switching between hadronic and partonic degrees of freedom, which is one of striking manifestations of dynamical restoration of chiral symmetry, is accompanied by a smooth behavior of entropy density at chiral crossover. Individual contributions of different components of strongly interacting matter to its speed of sound are analyzed for the first time. It is shown that restoration of chiral symmetry drives speed of sound of hadron gas to negative values, manifesting its mechanical instability and being in a strike disagreements with the lattice QCD data. Accounting for the partonic excitations naturally resolves this contradiction.
|
[
{
"version": "v1",
"created": "Thu, 26 Jun 2025 14:11:19 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ivanytskyi",
"Oleksii",
""
],
[
"Blaschke",
"David",
""
],
[
"Röpke",
"Gerd",
""
]
] |
2506.21509
|
Jiahe Chen
|
Jiahe Chen and Jiaying He and Qian Shao and Qiyuan Chen and Jiahe Ying and Hongxia Xu and Jintai Chen and Jianwei Zheng and Jian Wu
|
Mitigating Hallucination of Large Vision-Language Models via Dynamic Logits Calibration
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large Vision-Language Models (LVLMs) have demonstrated significant advancements in multimodal understanding, yet they are frequently hampered by hallucination-the generation of text that contradicts visual input. Existing training-free decoding strategies exhibit critical limitations, including the use of static constraints that do not adapt to semantic drift during generation, inefficiency stemming from the need for multiple forward passes, and degradation of detail due to overly rigid intervention rules. To overcome these challenges, this paper introduces Dynamic Logits Calibration (DLC), a novel training-free decoding framework designed to dynamically align text generation with visual evidence at inference time. At the decoding phase, DLC step-wise employs CLIP to assess the semantic alignment between the input image and the generated text sequence. Then, the Relative Visual Advantage (RVA) of candidate tokens is evaluated against a dynamically updated contextual baseline, adaptively adjusting output logits to favor tokens that are visually grounded. Furthermore, an adaptive weighting mechanism, informed by a real-time context alignment score, carefully balances the visual guidance while ensuring the overall quality of the textual output. Extensive experiments conducted across diverse benchmarks and various LVLM architectures (such as LLaVA, InstructBLIP, and MiniGPT-4) demonstrate that DLC significantly reduces hallucinations, outperforming current methods while maintaining high inference efficiency by avoiding multiple forward passes. Overall, we present an effective and efficient decoding-time solution to mitigate hallucinations, thereby enhancing the reliability of LVLMs for more practices. Code will be released on Github.
|
[
{
"version": "v1",
"created": "Thu, 26 Jun 2025 17:35:40 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:02:33 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Chen",
"Jiahe",
""
],
[
"He",
"Jiaying",
""
],
[
"Shao",
"Qian",
""
],
[
"Chen",
"Qiyuan",
""
],
[
"Ying",
"Jiahe",
""
],
[
"Xu",
"Hongxia",
""
],
[
"Chen",
"Jintai",
""
],
[
"Zheng",
"Jianwei",
""
],
[
"Wu",
"Jian",
""
]
] |
2506.21875
|
Jian Zhang
|
Jian Zhang, Linhao Zhang, Bokai Lei, Chuhan Wu, Wei Jia, Xiao Zhou
|
WildSpeech-Bench: Benchmarking Audio LLMs in Natural Speech Conversation
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Recent multi-modal Large Language Models (LLMs) such as GPT-4o have demonstrated strong capabilities of direct speech interaction. However, the lack of specialized and comprehensive benchmarks for end-to-end speech LLM evaluation hinders optimizing the user experience of Audio LLMs in real-world applications. Existing evaluation methods often adapt text-based benchmarks, overlooking speech's unique characteristics and challenges, including prosody, homophones, stuttering, and differing user expectations. Here, we present a novel approach to thoroughly evaluate LLMs in practical speech conversations. We systematically curate real-world chat data relevant to spoken scenarios, introduce diversity in speaker attributes and acoustic conditions, and augment the dataset with speech-specific phenomena. We further design a query-aware evaluation method to use customized evaluation checklists and prompts to enhance the accuracy of automatic evaluation. We conduct comprehensive testing and detailed analysis of various mainstream speech models, revealing significant differences in model performance across different speech scenarios. The use of query-aware evaluation further enables a finer-grained assessment under various speech-specific scenarios. Our benchmark can provide valuable insights for speech model development and evaluation.
|
[
{
"version": "v1",
"created": "Fri, 27 Jun 2025 03:18:45 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:23:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhang",
"Jian",
""
],
[
"Zhang",
"Linhao",
""
],
[
"Lei",
"Bokai",
""
],
[
"Wu",
"Chuhan",
""
],
[
"Jia",
"Wei",
""
],
[
"Zhou",
"Xiao",
""
]
] |
2506.23232
|
Yu-Zhu Sun
|
Yu-Zhu Sun, Hong-Xin Zhang, Elias Brinks, Rory Smith, Fujia Li, Minsu Kim, Se-Heon Oh, Zesen Lin, Jaebeom Kim, Weibin Sun, Tie Li, Patrick C\^ot\'e, Alessandro Boselli, Lijun Chen, Pierre-Alain Duc, Sanjaya Paudel, Matthew A. Taylor, Kaixiang Wang, Enci Wang, Lanyue Zhang, Yinghe Zhao
|
Unveiling the nature and fate of the almost-dark cloud AGC 226178 through HI mapping
|
18 pages, 14 figures. Accepted for Publication in A&A
| null | null | null |
astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
The origin of extragalactic, almost dark HI clouds with extreme gas-to-stellar mass ratios remains poorly understood. We investigate the nature and fate of the "almost dark" cloud AGC 226178, projected within the Virgo cluster, with an HI-to-stellar mass ratio of ~1000. We present deep single-dish HI mapping from the Five-hundred-meter Aperture Spherical Telescope (FAST), complemented by high-resolution interferometric data from the Very Large Array (VLA), as part of the Atomic gas in Virgo Interacting Dwarf galaxies (AVID) project. These observations provide the highest-quality HI analysis to date of such a cloud, combining resolution and sensitivity. FAST data reveal a short, low-velocity tail toward the dwarf galaxy VCC 2034, previously proposed as a possible origin for AGC 226178. However, VCC 2034 shows a line-of-sight asymmetric HI feature and cometary morphology indicating a stripping event unrelated to AGC 226178. VLA data reveal a velocity gradient across AGC 226178 and a clumpy internal structure. The velocity dispersion exceeds the thermal linewidth, implying turbulence or unresolved motions. The cloud cannot be gravitationally bound by atomic gas alone. The resolved HI clumps follow standard HI mass-star formation rate and mass-size relations, with those forming stars reaching surface densities above the threshold for self-shielding. We conclude that AGC 226178 is a free-floating HI cloud of unknown origin. The system appears to be in the process of disintegration. It is likely located well outside the Virgo cluster, as the preservation of its extended HI morphology within the cluster environment would otherwise require a substantial reservoir of unseen molecular gas with a mass exceeding that of the observed HI content. While confinement pressure from the hot intracluster medium may aid its stability, it is unlikely to be the dominant factor preventing its disruption.
|
[
{
"version": "v1",
"created": "Sun, 29 Jun 2025 13:32:27 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:22:09 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Sun",
"Yu-Zhu",
""
],
[
"Zhang",
"Hong-Xin",
""
],
[
"Brinks",
"Elias",
""
],
[
"Smith",
"Rory",
""
],
[
"Li",
"Fujia",
""
],
[
"Kim",
"Minsu",
""
],
[
"Oh",
"Se-Heon",
""
],
[
"Lin",
"Zesen",
""
],
[
"Kim",
"Jaebeom",
""
],
[
"Sun",
"Weibin",
""
],
[
"Li",
"Tie",
""
],
[
"Côté",
"Patrick",
""
],
[
"Boselli",
"Alessandro",
""
],
[
"Chen",
"Lijun",
""
],
[
"Duc",
"Pierre-Alain",
""
],
[
"Paudel",
"Sanjaya",
""
],
[
"Taylor",
"Matthew A.",
""
],
[
"Wang",
"Kaixiang",
""
],
[
"Wang",
"Enci",
""
],
[
"Zhang",
"Lanyue",
""
],
[
"Zhao",
"Yinghe",
""
]
] |
2507.00059
|
Ranjan Naik
|
Ranjan N Naik
|
Computational Verification of the Buratti--Horak--Rosa Conjecture for Small Integers and Inductive Approaches
|
This result supports the results by Mariusz Meszka for all primes up to 23 (included) with the aid of a computer. Additional results on Coprime BHR Conjecture verifications for p < 31 and Inductive Approaches are included in this revision
| null | null | null |
cs.DM cs.DS math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
This paper presents a comprehensive computational approach to verify and inductively construct Hamiltonian paths for the Buratti--Horak--Rosa (BHR) Conjecture. The conjecture posits that for any multiset $L$ of $p-1$ positive integers not exceeding $\lfloor p/2 \rfloor$, there exists a Hamiltonian path in the complete graph $K_p$ with vertex-set $\{0, 1, \dots, p-1\}$ whose edge lengths (under the cyclic metric) match $L$, if and only if for every divisor $d$ of $p$, the number of multiples of $d$ appearing in $L$ is at most $p - d$.
Building upon prior computational work by Mariusz Meszka, which verified the conjecture for all primes up to $p=23$, our Python program extends this verification significantly. We approach the problem by systematically generating frequency partitions (FPs) of edge lengths and employing a recursive backtracking algorithm. We report successful computational verification for all frequency partitions for integers $p < 32$, specifically presenting results for $p=31$ and a composite $p=26$. For the composite number $p=30$, the Python code took approximately 11 hours to verify on a Lenovo laptop. For $p=16$, $167,898$ valid multisets were processed, taking around 20 hours on Google Colab Pro+.
Furthermore, we introduce and implement two constructive, inductive strategies for building Hamiltonian paths: (1) increasing the multiplicity of an existing edge length, and (2) adding a new edge length. These methods, supported by a reuse-insertion heuristic and backtracking search, demonstrate successful constructions for evolving FPs up to $p=40$. Through these empirical tests and performance metrics, we provide strong computational evidence for the validity of the BHR conjecture within the scope tested, and outline the scalability of our approach for higher integer values.
|
[
{
"version": "v1",
"created": "Thu, 26 Jun 2025 23:33:25 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Jul 2025 18:11:59 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Jul 2025 23:19:51 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 13:56:28 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Naik",
"Ranjan N",
""
]
] |
2507.00578
|
Dizhan Du
|
Dizhan Du and Bei You and Zhen Yan and Xinwu Cao and Jean-Marie Hameury and Yue Wu
|
A comprehensive study of time delay between optical/near-infrared and X-ray emissions in black hole X-ray binaries
|
submitted to MNRAS
| null | null | null |
astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
We conducted a comprehensive study of daily delays using multi-wavelength data from a sample of well-studied black hole X-ray binaries, specifically focusing on the sources GX 339-4, 4U 1543-47, and XTE J1550-564. The Interpolated-Correlation Function method was employed to investigate the temporal relationship between the X-ray (Compton component) and optical-infrared (OIR) emissions. Our results show that during the rising hard state, the Compton emission consistently lags behind OIR emission for several days. In contrast, during the decaying hard state, the OIR emission lags behind the Compton emission by approximately 6 to 35 days. This measurement can potentially be used in models of accretion physics and disk instability. We explore the underlying mechanisms responsible for these time delays, highlighting the critical role of viscous heating in the accretion disk in generating OIR luminosity for these sources. The observed time delays during both the rising and decaying hard states are well explained by the disk instability model.
|
[
{
"version": "v1",
"created": "Tue, 01 Jul 2025 08:59:44 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:53:11 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Du",
"Dizhan",
""
],
[
"You",
"Bei",
""
],
[
"Yan",
"Zhen",
""
],
[
"Cao",
"Xinwu",
""
],
[
"Hameury",
"Jean-Marie",
""
],
[
"Wu",
"Yue",
""
]
] |
2507.01631
|
Camille Billouard
|
Camille Billouard, Dawa Derksen, Alexandre Constantin, Bruno Vallet
|
Tile and Slide : A New Framework for Scaling NeRF from Local to Global 3D Earth Observation
|
Accepted at ICCV 2025 Workshop 3D-VAST (From street to space: 3D Vision Across Altitudes). Our code will be made public after the conference at https://github.com/Ellimac0/Snake-NeRF
| null | null | null |
cs.CV cs.AI cs.GR cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Neural Radiance Fields (NeRF) have recently emerged as a paradigm for 3D reconstruction from multiview satellite imagery. However, state-of-the-art NeRF methods are typically constrained to small scenes due to the memory footprint during training, which we study in this paper. Previous work on large-scale NeRFs palliate this by dividing the scene into NeRFs. This paper introduces Snake-NeRF, a framework that scales to large scenes. Our out-of-core method eliminates the need to load all images and networks simultaneously, and operates on a single device. We achieve this by dividing the region of interest into NeRFs that 3D tile without overlap. Importantly, we crop the images with overlap to ensure each NeRFs is trained with all the necessary pixels. We introduce a novel $2\times 2$ 3D tile progression strategy and segmented sampler, which together prevent 3D reconstruction errors along the tile edges. Our experiments conclude that large satellite images can effectively be processed with linear time complexity, on a single GPU, and without compromise in quality.
|
[
{
"version": "v1",
"created": "Wed, 02 Jul 2025 11:59:36 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:32:03 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Billouard",
"Camille",
""
],
[
"Derksen",
"Dawa",
""
],
[
"Constantin",
"Alexandre",
""
],
[
"Vallet",
"Bruno",
""
]
] |
2507.01689
|
Nikolaos Mavromatos
|
Panagiotis Dorlis, Nick E. Mavromatos, Sarben Sarkar, Sotirios-Neilos Vlachos
|
Superradiant Axionic Black-Hole Clouds as Seeds for Graviton Squeezing
|
7 pages revtex, four pdf figures incorporated
| null | null |
KCL-PH-TH/2025-15
|
gr-qc hep-ph hep-th quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
It is shown that both, standard general relativity (GR) and Chern-Simons (CS) gravity, the latter containing chiral gravitational anomaly terms, seed the production of pairs of entangled gravitons in a multi-mode squeezed state. This involves the interaction of gravitons with the axionic cloud surrounding a superradiant Kerr (rotating) black hole background. The order of magnitude of the squeezing effect, specifically the number of graviton excitations in the squeezed vacuum, is estimated in the non-relativistic limit, relevant for the superradiance process. It is found analytically that the squeezing from the GR process of annihilation of two axions into two gravitons, dominates, by many orders of magnitude, that coming from the axion decay into two gravitons, induced by the higher-derivative CS term. It is also shown that significant squeezing effects are produced in the case of long-lived axionic clouds, whose lifetimes are much longer than the timescale for which superradiance is effective.
|
[
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:17:22 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dorlis",
"Panagiotis",
""
],
[
"Mavromatos",
"Nick E.",
""
],
[
"Sarkar",
"Sarben",
""
],
[
"Vlachos",
"Sotirios-Neilos",
""
]
] |
2507.01694
|
Hanlin Cai
|
Hanlin Cai, Haofan Dong, Houtianfu Wang, Kai Li, Ozgur B. Akan
|
Graph Representation-based Model Poisoning on Federated Large Language Models
|
7 pages, 5 figures (Submitted to IEEE Communication Magazine)
| null | null | null |
cs.CR cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Federated large language models (FedLLMs) enable powerful generative capabilities within wireless networks while preserving data privacy. Nonetheless, FedLLMs remain vulnerable to model poisoning attacks. This article first reviews recent advancements in model poisoning techniques and existing defense mechanisms for FedLLMs, underscoring critical limitations, especially when dealing with non-IID textual data distributions. Current defense strategies predominantly employ distance or similarity-based outlier detection mechanisms, relying on the assumption that malicious updates markedly differ from benign statistical patterns. However, this assumption becomes inadequate against adaptive adversaries targeting billion-parameter LLMs. The article further investigates graph representation-based model poisoning (GRMP), an emerging attack paradigm that exploits higher-order correlations among benign client gradients to craft malicious updates indistinguishable from legitimate ones. GRMP can effectively circumvent advanced defense systems, causing substantial degradation in model accuracy and overall performance. Moreover, the article outlines a forward-looking research roadmap that emphasizes the necessity of graph-aware secure aggregation methods, specialized vulnerability metrics tailored for FedLLMs, and evaluation frameworks to enhance the robustness of federated language model deployments.
|
[
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:20:52 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:30:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cai",
"Hanlin",
""
],
[
"Dong",
"Haofan",
""
],
[
"Wang",
"Houtianfu",
""
],
[
"Li",
"Kai",
""
],
[
"Akan",
"Ozgur B.",
""
]
] |
2507.01910
|
Mikhail Shneider
|
Mikhail Pekker and Mikhail N. Shneider
|
Sub-barrier cavitation in liquid helium
| null | null | null | null |
cond-mat.soft
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, the tunneling mechanism of cavitation in liquid helium for 3He and 4He is considered on the basis of the Schr\"odinger-like equation. It is assumed that the pairwise interactions of helium atoms are determined by the Lennard-Jones potential. The kinetics of nucleation and the mechanism that limits the growth of cavitation bubbles in liquid helium are considered, taking into account their growth in a negative pressure field.
|
[
{
"version": "v1",
"created": "Wed, 25 Jun 2025 23:23:54 GMT"
},
{
"version": "v2",
"created": "Tue, 08 Jul 2025 22:03:48 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 18:19:40 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Pekker",
"Mikhail",
""
],
[
"Shneider",
"Mikhail N.",
""
]
] |
2507.02473
|
Jebarathinam Chellasamy Dr
|
Chellasamy Jebarathinam
|
Popescu-Rohrlich box fraction of dimensionally restricted nonlocality and secure quantum key distribution
|
4 pages. The proof of Proposition 1 has been revised
| null | null | null |
quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
With a measure of dimensionally restricted nonlocality in the bipartite Bell scenario with two inputs and two outputs, a different decomposition of any nonsignaling box that is a convex mixture of a single Popescu-Rohrlich box and a Bell-local box, with the measure being zero, is introduced. It is shown that any nonsignaling box shared by Alice and Bob that has dimensionally restricted nonlocality contains secret correlations against any third party, Eve, who is also dimensionally restricted. In this context, for the specific Bell scenario considered, it is demonstrated that the Popescu-Rohrlich box fraction can be used as a resource for secure quantum key distribution even if entanglement is not certified.
|
[
{
"version": "v1",
"created": "Thu, 03 Jul 2025 09:30:26 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:08:41 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Jebarathinam",
"Chellasamy",
""
]
] |
2507.02552
|
Haeran Cho Dr
|
Haeran Cho and Housen Li
|
Covariance scanning for adaptively optimal change point detection in high-dimensional linear models
| null | null | null | null |
math.ST stat.ME stat.TH
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
This paper investigates the detection and estimation of a single change in high-dimensional linear models. We derive minimax lower bounds for the detection boundary and the estimation rate, which uncover a phase transition governed the sparsity of the covariance-weighted differential parameter. This form of "inherent sparsity" captures a delicate interplay between the covariance structure of the regressors and the change in regression coefficients on the detectability of a change point. Complementing the lower bounds, we introduce two covariance scanning-based methods, McScan and QcSan, which achieve minimax optimal performance (up to possible logarithmic factors) in the sparse and the dense regimes, respectively. In particular, QcScan is the first method shown to achieve consistency in the dense regime and further, we devise a combined procedure which is adaptively minimax optimal across sparse and dense regimes without the knowledge of the sparsity. Computationally, covariance scanning-based methods avoid costly computation of Lasso-type estimators and attain worst-case computation complexity that is linear in the dimension and sample size. Additionally, we consider the post-detection estimation of the differential parameter and the refinement of the change point estimator. Simulation studies support the theoretical findings and demonstrate the computational and statistical efficiency of the proposed covariance scanning methods.
|
[
{
"version": "v1",
"created": "Thu, 03 Jul 2025 11:53:31 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:49:33 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cho",
"Haeran",
""
],
[
"Li",
"Housen",
""
]
] |
2507.02708
|
Ananya Rao
|
Ananya Rao, Alyssa Hargis, David Wettergreen, Howie Choset
|
Optimizing Start Locations in Ergodic Search for Disaster Response
| null | null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In disaster response scenarios, deploying robotic teams effectively is crucial for improving situational awareness and enhancing search and rescue operations. The use of robots in search and rescue has been studied but the question of where to start robot deployments has not been addressed. This work addresses the problem of optimally selecting starting locations for robots with heterogeneous capabilities by formulating a joint optimization problem. To determine start locations, this work adds a constraint to the ergodic optimization framework whose minimum assigns robots to start locations. This becomes a little more challenging when the robots are heterogeneous (equipped with different sensing and motion modalities) because not all robots start at the same location, and a more complex adaptation of the aforementioned constraint is applied. Our method assumes access to potential starting locations, which can be obtained from expert knowledge or aerial imagery. We experimentally evaluate the efficacy of our joint optimization approach by comparing it to baseline methods that use fixed starting locations for all robots. Our experimental results show significant gains in coverage performance, with average improvements of 35.98% on synthetic data and 31.91% on real-world data for homogeneous and heterogeneous teams, in terms of the ergodic metric.
|
[
{
"version": "v1",
"created": "Thu, 03 Jul 2025 15:24:54 GMT"
},
{
"version": "v2",
"created": "Fri, 04 Jul 2025 14:53:28 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 22:45:04 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Rao",
"Ananya",
""
],
[
"Hargis",
"Alyssa",
""
],
[
"Wettergreen",
"David",
""
],
[
"Choset",
"Howie",
""
]
] |
2507.03398
|
Roberto Emparan
|
Roberto Emparan and Stefano Trezzi
|
Quantum Transparency of Near-extremal Black Holes
|
22 pages, 8 figures. v2: 23 pages, 8 figures. Small improvements, incl mention of di-quanta absorption, and refined quantum vs semiclassical comparison
| null | null | null |
hep-th gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We investigate the scattering of electromagnetic and gravitational waves off a Reissner-Nordstr\"om black hole in the low-temperature regime where the near-horizon throat experiences large quantum fluctuations. We find that the black hole is transparent to electromagnetic and gravitational radiation of fixed helicity below a certain frequency threshold. This phenomenon arises because the angular momentum of the black hole is quantized, creating an energy gap between the spinless black hole state and the first excited spinning states. Radiation with angular momentum -- such as photons, gravitons, and partial waves of a massless scalar field, which we also study -- must supply enough energy to bridge this gap to be absorbed. Below this threshold, no absorption can occur, rendering the black hole transparent. For frequencies above the gap, the scarcity of black hole states continues to suppress the absorption cross-section relative to semiclassical predictions, making the black hole translucent rather than completely transparent. Notably, electromagnetic absorption is significantly stronger than gravitational absorption, beyond what differences in spin alone would suggest.
|
[
{
"version": "v1",
"created": "Fri, 04 Jul 2025 09:05:10 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 00:44:41 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Emparan",
"Roberto",
""
],
[
"Trezzi",
"Stefano",
""
]
] |
2507.04382
|
Kumiko Kotera
|
Kumiko Kotera, Ingo Allekotte, Jaime Alvarez-Mu\~niz, Sergio Cabana-Freire, Valentin Decoene, Luciano Ferreyro, Ars\`ene Ferri\`ere, Matias Hampel, Olivier Martineau-Huynh, Valentin Niess, Federico Sanchez, Stephanie Wissel, Andrew Zeolla (for the BEACON and GRAND Collaborations)
|
The Hybrid Elevated Radio Observatory for Neutrinos (HERON) Project
|
Presented at the 39th International Cosmic Ray Conference (ICRC 2023). 8 pages, 6 figures
| null | null |
PoS(ICRC2025)1078
|
astro-ph.IM astro-ph.HE hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Measuring ultra-high energy neutrinos, with energies above $10^{16}$ eV, is the next frontier of the emerging multi-messenger era. Their detection requires building a large-scale detector with 10 times the instantaneous sensitivity of current instruments, sub-degree angular resolution, and wide daily field of view. The Hybrid Elevated Radio Observatory for Neutrinos (HERON) is designed to be that discovery instrument. HERON combines the complementary features of two radio techniques being demonstrated by the BEACON and GRAND prototypes. Its preliminary design consists of 24 compact, elevated phased stations with 24 antennas each, embedded in a sparse array of 360 standalone antennas. This setup tunes the energy threshold to below 100 PeV, where the neutrino flux should be high. The sensitivity of the phased stations combines with the powerful reconstruction capacities of the standalone antennas to produce an optimal detector. HERON is planned to be installed at an elevation of 1,000 m across a 72 km-long mountain range overlooking a valley in Argentina's San Juan province. It would be connected to the worldwide network of multimessenger observatories and search for neutrino bursts from candidate sources of cosmic rays, like gamma-ray bursts and other powerful transients. With HERON's deep sensitivity, this strategy targets discoveries that cast new light into the inner workings of the most violent astrophysical sources at uncharted energies. We present the preliminary design, performances, and observation strategy of HERON.
|
[
{
"version": "v1",
"created": "Sun, 06 Jul 2025 13:02:28 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 17:06:56 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 10:09:28 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kotera",
"Kumiko",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Allekotte",
"Ingo",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Alvarez-Muñiz",
"Jaime",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Cabana-Freire",
"Sergio",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Decoene",
"Valentin",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Ferreyro",
"Luciano",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Ferrière",
"Arsène",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Hampel",
"Matias",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Martineau-Huynh",
"Olivier",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Niess",
"Valentin",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Sanchez",
"Federico",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Wissel",
"Stephanie",
"",
"for the BEACON and GRAND Collaborations"
],
[
"Zeolla",
"Andrew",
"",
"for the BEACON and GRAND Collaborations"
]
] |
2507.05168
|
Iv\'an Le\'on
|
Iv\'an Le\'on and Diego Paz\'o
|
Dynamics and chaotic properties of the fully disordered Kuramoto model
|
11 pages, 7 figures
| null |
10.1063/5.0272051
| null |
cond-mat.dis-nn nlin.CD
|
http://creativecommons.org/licenses/by/4.0/
|
Frustrated random interactions are a key ingredient of spin glasses. From this perspective, we study the dynamics of the Kuramoto model with quenched random couplings: the simplest oscillator ensemble with fully disordered interactions. We answer some open questions by means of extensive numerical simulations and a perturbative calculation (the cavity method). We show frequency entrainment is not realized in the thermodynamic limit and that chaotic dynamics are pervasive in parameter space. In the weak coupling regime, we find closed formulas for the frequency shift and the dissipativeness of the model. Interestingly, the largest Lyapunov exponent is found to exhibit the same asymptotic dependence on the coupling constant irrespective of the coupling asymmetry, within the numerical accuracy.
|
[
{
"version": "v1",
"created": "Mon, 07 Jul 2025 16:23:19 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"León",
"Iván",
""
],
[
"Pazó",
"Diego",
""
]
] |
2507.05487
|
Christian Jooss
|
V. Moshnyaga, Ch. Jooss, P. E. Bl\"ochl, V. Bruchmann-Bamberg, A. Dehning, L. Allen-Rump, C. Hausmann, M. Kr\"uger, A. Rathnakaran, S. Rajpurohit, D. Steil, C. Flathmann, J. Hoffmann, M. Seibt, C. Volkert
|
Elementary Steps of Energy Conversion in Strongly Correlated Systems: Beyond Single Quasiparticles and Rigid Bands
| null | null | null | null |
cond-mat.str-el
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Energy conversion in materials can be considered as a sequence of elementary steps initiated by a primary excitation. While these steps are quite well understood in classical semiconductors in terms of quasiparticle (QP) excitations and interactions, their understanding in strongly correlated materials is still elusive. Here, we review the progress which has been achieved over recent years by studies of manganite perovskite oxides as a model system for materials with strong correlations. They show a subtle interplay of different types of correlations, i.e., electron-phonon, electron-electron and spin-spin, resulting in rich physical phenomena due to competition between different ground states accompanied by temperature- and field-induced phase transitions. They strongly impact various types of energy conversion and transport processes including friction at surfaces, thermal transport, time-, energy- and power-dependent optical excitations as well as photovoltaic energy conversion. The underlying microscopic processes can be broken down to the behavior of the low-energy thermal and high-energy optical excitations, their interactions, transport and conversion which are theoretically analyzed by using models of interacting and tunable QPs: Their nature and interactions can change during excitation, transport and phase transitions, thus modifying electronic structure. At sufficiently high stimulation, QP excitations can even induce or actuate phase transitions. As a result, we obtained a comprehensive understanding of energy conversion steps going far beyond single QP pictures and rigid band approximations well-known for conventional semiconductors.
|
[
{
"version": "v1",
"created": "Mon, 07 Jul 2025 21:19:09 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 20:53:57 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Moshnyaga",
"V.",
""
],
[
"Jooss",
"Ch.",
""
],
[
"Blöchl",
"P. E.",
""
],
[
"Bruchmann-Bamberg",
"V.",
""
],
[
"Dehning",
"A.",
""
],
[
"Allen-Rump",
"L.",
""
],
[
"Hausmann",
"C.",
""
],
[
"Krüger",
"M.",
""
],
[
"Rathnakaran",
"A.",
""
],
[
"Rajpurohit",
"S.",
""
],
[
"Steil",
"D.",
""
],
[
"Flathmann",
"C.",
""
],
[
"Hoffmann",
"J.",
""
],
[
"Seibt",
"M.",
""
],
[
"Volkert",
"C.",
""
]
] |
2507.05903
|
Gerd Gra{\ss}hoff
|
Gerd Gra{\ss}hoff
|
AI-Reporter: A Path to a New Genre of Scientific Communication
| null | null | null | null |
cs.DL cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
The AI-Reporter represents a paradigmatic shift in scientific publication practice. This document demonstrates through a concrete case study how our system transforms academic presentations into publication-ready chapters -- in less than three minutes. Using Arno Simons' lecture on Large Language Models from the ``Large Language Models for the History, Philosophy, and Sociology of Science'' workshop (NEPI) as an example, we show how technological innovation bridges the gap between ephemeral presentation and permanent scientific documentation.
|
[
{
"version": "v1",
"created": "Tue, 08 Jul 2025 11:41:37 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 06:23:02 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Graßhoff",
"Gerd",
""
]
] |
2507.06005
|
Diana Baumann
|
Diana Baumann and Tim C. Rese and David Bermbach
|
Towards Serverless Processing of Spatiotemporal Big Data Queries
|
Accepted for publication in 13th IEEE International Conference on Cloud Engineering (IC2E 2025)
| null | null | null |
cs.DB cs.DC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Spatiotemporal data are being produced in continuously growing volumes by a variety of data sources and a variety of application fields rely on rapid analysis of such data. Existing systems such as PostGIS or MobilityDB usually build on relational database systems, thus, inheriting their scale-out characteristics. As a consequence, big spatiotemporal data scenarios still have limited support even though many query types can easily be parallelized. In this paper, we propose our vision of a native serverless data processing approach for spatiotemporal data: We break down queries into small subqueries which then leverage the near-instant scaling of Function-as-a-Service platforms to execute them in parallel. With this, we partially solve the scalability needs of big spatiotemporal data processing.
|
[
{
"version": "v1",
"created": "Tue, 08 Jul 2025 14:08:30 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:13:42 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Baumann",
"Diana",
""
],
[
"Rese",
"Tim C.",
""
],
[
"Bermbach",
"David",
""
]
] |
2507.06448
|
Zhenhailong Wang
|
Zhenhailong Wang, Xuehang Guo, Sofia Stoica, Haiyang Xu, Hongru Wang, Hyeonjeong Ha, Xiusi Chen, Yangyi Chen, Ming Yan, Fei Huang, Heng Ji
|
Perception-Aware Policy Optimization for Multimodal Reasoning
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Reinforcement Learning with Verifiable Rewards (RLVR) has proven to be a highly effective strategy for endowing Large Language Models (LLMs) with robust multi-step reasoning abilities. However, its design and optimizations remain tailored to purely textual domains, resulting in suboptimal performance when applied to multimodal reasoning tasks. In particular, we observe that a major source of error in current multimodal reasoning lies in the perception of visual inputs. To address this bottleneck, we propose PAPO, a novel policy gradient algorithm that encourages the model to learn to perceive while learning to reason. Specifically, we introduce the Implicit Perception Loss in the form of a KL divergence term, which can be seamlessly plugged into mainstream RLVR algorithms such as GRPO and DAPO. Notably, PAPO does not rely on additional data curation, reward models, or stronger teacher models. To further enhance the training stability of PAPO, we introduce the Double Entropy Loss, which effectively regularizes the new KL objective without compromising performance. Despite its simplicity, PAPO yields significant overall improvements of 4.4%-17.5% on diverse multimodal benchmarks. The improvements are more pronounced, approaching 8.0%-19.1%, on tasks with high vision dependency. We also observe a substantial reduction of 30.5% in perception errors, indicating improved perceptual capabilities with PAPO. Overall, our work introduces a deeper integration of perception-aware supervision into core learning objectives and lays the groundwork for a new RL framework that encourages visually grounded reasoning. Code and data will be made publicly available for research purposes. Project page: https://mikewangwzhl.github.io/PAPO.
|
[
{
"version": "v1",
"created": "Tue, 08 Jul 2025 23:22:34 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Jul 2025 16:48:37 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 17:54:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Zhenhailong",
""
],
[
"Guo",
"Xuehang",
""
],
[
"Stoica",
"Sofia",
""
],
[
"Xu",
"Haiyang",
""
],
[
"Wang",
"Hongru",
""
],
[
"Ha",
"Hyeonjeong",
""
],
[
"Chen",
"Xiusi",
""
],
[
"Chen",
"Yangyi",
""
],
[
"Yan",
"Ming",
""
],
[
"Huang",
"Fei",
""
],
[
"Ji",
"Heng",
""
]
] |
2507.06460
|
Ravi Chugh
|
Sam Cohen and Ravi Chugh
|
Ragged Blocks: Rendering Structured Text with Style
|
UIST 2025 Paper + Appendices
| null |
10.1145/3746059.3747751
| null |
cs.HC
|
http://creativecommons.org/licenses/by/4.0/
|
Whether it be source code in a programming language, prose in natural language, or otherwise, text is highly structured. Currently, text visualizations are confined either to _flat, line-based_ decorations, which can convey only limited information about textual structure, or _nested boxes_, which convey structure but often destroy the typographic layout of the underlying text. We hypothesize that the lack of rich styling options limits the kinds of information that are displayed alongside text, wherever it may be displayed.
In this paper, we show that it is possible to achieve arbitrarily nested decorations while minimally disturbing the underlying typographic layout. Specifically, we present a layout algorithm that generates _ragged blocks_, or _rocks_, which are rectilinear polygons that allow nested text to be compactly rendered even when styled with borders and padding. Our layout algorithm is evaluated on a benchmark suite comprising representative source code files in multiple programming languages. The (ragged block) layouts produced by our algorithm are substantially more compact than the (rectangular block) layouts produced by conventional techniques, when uniformly styling every element in the syntax tree with borders and padding.
|
[
{
"version": "v1",
"created": "Wed, 09 Jul 2025 00:26:52 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 04:12:31 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cohen",
"Sam",
""
],
[
"Chugh",
"Ravi",
""
]
] |
2507.06557
|
Kaoru Mizuta
|
Kaoru Mizuta
|
On the commutator scaling in Hamiltonian simulation with multi-product formulas
|
23 pages, 1 table
| null | null | null |
quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A multi-product formula (MPF) is a promising approach for Hamiltonian simulation efficiently both in the system size $N$ and the inverse allowable error $1/\varepsilon$ by combining Trotterization and the linear combination of unitaries (LCU). It achieves poly-logarithmic cost in $1/\varepsilon$ like LCU [G. H. Low, V. Kliuchnikov, N. Wiebe, arXiv:1907.11679 (2019)]. The efficiency in $N$ is expected to come from the commutator scaling in Trotterization, and this appears to be confirmed by the error bound of MPF expressed by nested commutators [J. Aftab, D. An, K. Trivisa, arXiv:2403.08922 (2024)]. However, we point out that the efficiency of MPF in the system size $N$ is not exactly resolved yet in that the present error bound expressed by nested commutators is incompatible with the size-efficient complexity reflecting the commutator scaling. The problem is that $q$-fold nested commutators with arbitrarily large $q$ are involved in their requirement and error bound. The benefit of commutator scaling by locality is absent, and the cost efficient in $N$ becomes prohibited in general. In this paper, we show an alternative commutator-scaling error of MPF and derive its size-efficient cost properly inheriting the advantage in Trotterization. The requirement and the error bound in our analysis, derived by techniques from the Floquet-Magnus expansion, have a certain truncation order in the nested commutators and can fully exploit the locality. We prove that Hamiltonian simulation by MPF certainly achieves the cost whose system-size dependence is as large as Trotterization while keeping the $\mathrm{polylog}(1/\varepsilon)$-scaling like the LCU. Our results will provide improved or accurate error and cost also for various algorithms using interpolation or extrapolation of Trotterization.
|
[
{
"version": "v1",
"created": "Wed, 09 Jul 2025 05:25:06 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:13:36 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Mizuta",
"Kaoru",
""
]
] |
2507.06781
|
Juan A. Molina-Calzada
|
J. A. Molina-Calzada, M. A. Lara-L\'opez, J. Gallego, A. M. Hopkins, Benne W. Holwerda and A. R. L\'opez-S\'anchez
|
Galaxy And Mass Assembly (GAMA): From filaments to voids, how extreme environment affects gas metallicity and SFR in galaxies
|
Accepted for publication in A&A
| null |
10.1051/0004-6361/202555871
| null |
astro-ph.GA
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
We analyse the stellar mass-metallicity (M-Z) and stellar mass-star formation rate (M-SFR) relations for star-forming galaxies classified by their environment and compare them with matched control samples of field galaxies. Using data from the Galaxy And Mass Assembly (GAMA) survey and the filament catalogue, which categorises galaxies into filaments, tendrils, and voids, we correct emission lines for dust extinction and select star-forming galaxies based on the BPT diagram. Metallicity and star formation rate are estimated and used to fit the M-Z and M-SFR relations through both Bayesian and least-squares approaches. We find that metallicity increases in denser environments, while star formation rate decreases, with the most notable contrasts seen between filament/tendril galaxies and those in voids. Galaxies in filaments and tendrils that are not group members show little to no deviation from their control samples. Morphological analysis reveals no significant differences. Overall, galaxies in denser environments appear more chemically enriched with lower SFRs, likely due to processed material and reduced cold gas availability, while isolated void galaxies maintain higher SFRs and lower metallicities, possibly due to ongoing cold gas accretion. These results suggest that local environmental conditions, rather than large-scale structure alone, are the main drivers of the observed trends.
|
[
{
"version": "v1",
"created": "Wed, 09 Jul 2025 12:13:18 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 11:19:25 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Molina-Calzada",
"J. A.",
""
],
[
"Lara-López",
"M. A.",
""
],
[
"Gallego",
"J.",
""
],
[
"Hopkins",
"A. M.",
""
],
[
"Holwerda",
"Benne W.",
""
],
[
"López-Sánchez",
"A. R.",
""
]
] |
2507.07122
|
Giuseppe Puglisi Ph.D.
|
M. Bortolami, N. Raffuzzi, L. Pagano, G. Puglisi, A. Anand, A. J. Banday, P. Campeti, G. Galloni, A. I. Lonappan, M. Monelli, M. Tomasi, G. Weymann-Despres, D. Adak, E. Allys, J. Aumont, R. Aurvik, C. Baccigalupi, M. Ballardini, R. B. Barreiro, N. Bartolo, S. Basak, M. Bersanelli, A. Besnard, T. Brinckmann, E. Calabrese, E. Carinos, A. Carones, F. J. Casas, K. Cheung, M. Citran, L. Clermont, F. Columbro, G. Coppi, A. Coppolecchia, F. Cuttaia, P. Dal Bo, P. de Bernardis, E. de la Hoz, M. De Lucia, S. Della Torre, P. Diego-Palazuelos, H. K. Eriksen, T. Essinger-Hileman, C. Franceschet, U. Fuskeland, M. Galloway, M. Gerbino, M. Gervasi, T. Ghigna, S. Giardiello, C. Gimeno-Amo, E. Gjerl{\o}w, A. Gruppuso, M. Hazumi, S. Henrot-Versill\'e, L. T. Hergt, E. Hivon, B. Jost, K. Kohri, L. Lamagna, T. Lari, M. Lattanzi, C. Leloup, F. Levrier, M. L\'opez-Caniego, G. Luzzi, J. Macias-Perez, B. Maffei, E. Mart\'inez-Gonz\'alez, S. Masi, S. Matarrese, T. Matsumura, S. Micheli, L. Montier, G. Morgante, L. Mousset, R. Nagata, T. Namikawa, A. Novelli, F. Noviello, I. Obata, A. Occhiuzzi, A. Paiella, D. Paoletti, G. Pascual-Cisneros, F. Piacentini, M. Pinchera, G. Polenta, L. Porcelli, M. Remazeilles, A. Ritacco, A. Rizzieri, J. A. Rubi\~no-Mart\'in, M. Ruiz-Granda, J. Sanghavi, V. Sauvage, M. Shiraishi, G. Signorelli, S. L. Stever, R. M. Sullivan, Y. Takase, K. Tassis, L. Terenzi, M. Tristram, L. Vacher, B. van Tent, P. Vielva, I. K. Wehus, M. Zannoni, Y. Zhou (for the LiteBIRD Collaboration)
|
First release of LiteBIRD simulations from an end-to-end pipeline
| null | null | null | null |
astro-ph.IM astro-ph.CO
|
http://creativecommons.org/licenses/by/4.0/
|
The LiteBIRD satellite mission aims at detecting Cosmic Microwave Background $B$ modes with unprecedented precision, targeting a total error on the tensor-to-scalar ratio $r$ of $\delta r \sim 0.001$. Operating from the L2 Lagrangian point of the Sun-Earth system, LiteBIRD will survey the full sky across 15 frequency bands (34 to 448 GHz) for 3 years.The current LiteBIRD baseline configuration employs 4508 detectors sampling at 19.1 Hz to achieve an effective polarization sensitivity of $ 2 \mu\mathrm{K-arcmin}$ and an angular resolution of 31 arcmin (at 140 GHz).We describe the first release of the official LiteBIRD simulations, realized with a new simulation pipeline developed using the LiteBIRD Simulation Framework, see https://github.com/litebird/litebird_sim . This pipeline generates 500 full-sky simulated maps at a Healpix resolution of nside=512. The simulations include also one year of Time Ordered Data for approximately one-third of LiteBIRD's total detectors.
|
[
{
"version": "v1",
"created": "Tue, 08 Jul 2025 09:27:35 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Jul 2025 10:09:12 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 12:58:12 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bortolami",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Raffuzzi",
"N.",
"",
"for the LiteBIRD Collaboration"
],
[
"Pagano",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Puglisi",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Anand",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Banday",
"A. J.",
"",
"for the LiteBIRD Collaboration"
],
[
"Campeti",
"P.",
"",
"for the LiteBIRD Collaboration"
],
[
"Galloni",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Lonappan",
"A. I.",
"",
"for the LiteBIRD Collaboration"
],
[
"Monelli",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Tomasi",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Weymann-Despres",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Adak",
"D.",
"",
"for the LiteBIRD Collaboration"
],
[
"Allys",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Aumont",
"J.",
"",
"for the LiteBIRD Collaboration"
],
[
"Aurvik",
"R.",
"",
"for the LiteBIRD Collaboration"
],
[
"Baccigalupi",
"C.",
"",
"for the LiteBIRD Collaboration"
],
[
"Ballardini",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Barreiro",
"R. B.",
"",
"for the LiteBIRD Collaboration"
],
[
"Bartolo",
"N.",
"",
"for the LiteBIRD Collaboration"
],
[
"Basak",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Bersanelli",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Besnard",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Brinckmann",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Calabrese",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Carinos",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Carones",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Casas",
"F. J.",
"",
"for the LiteBIRD Collaboration"
],
[
"Cheung",
"K.",
"",
"for the LiteBIRD Collaboration"
],
[
"Citran",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Clermont",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Columbro",
"F.",
"",
"for the LiteBIRD Collaboration"
],
[
"Coppi",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Coppolecchia",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Cuttaia",
"F.",
"",
"for the LiteBIRD Collaboration"
],
[
"Bo",
"P. Dal",
"",
"for the LiteBIRD Collaboration"
],
[
"de Bernardis",
"P.",
"",
"for the LiteBIRD Collaboration"
],
[
"de la Hoz",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"De Lucia",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Della Torre",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Diego-Palazuelos",
"P.",
"",
"for the LiteBIRD Collaboration"
],
[
"Eriksen",
"H. K.",
"",
"for the LiteBIRD Collaboration"
],
[
"Essinger-Hileman",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Franceschet",
"C.",
"",
"for the LiteBIRD Collaboration"
],
[
"Fuskeland",
"U.",
"",
"for the LiteBIRD Collaboration"
],
[
"Galloway",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Gerbino",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Gervasi",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Ghigna",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Giardiello",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Gimeno-Amo",
"C.",
"",
"for the LiteBIRD Collaboration"
],
[
"Gjerløw",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Gruppuso",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Hazumi",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Henrot-Versillé",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Hergt",
"L. T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Hivon",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Jost",
"B.",
"",
"for the LiteBIRD Collaboration"
],
[
"Kohri",
"K.",
"",
"for the LiteBIRD Collaboration"
],
[
"Lamagna",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Lari",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Lattanzi",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Leloup",
"C.",
"",
"for the LiteBIRD Collaboration"
],
[
"Levrier",
"F.",
"",
"for the LiteBIRD Collaboration"
],
[
"López-Caniego",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Luzzi",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Macias-Perez",
"J.",
"",
"for the LiteBIRD Collaboration"
],
[
"Maffei",
"B.",
"",
"for the LiteBIRD Collaboration"
],
[
"Martínez-González",
"E.",
"",
"for the LiteBIRD Collaboration"
],
[
"Masi",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Matarrese",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Matsumura",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Micheli",
"S.",
"",
"for the LiteBIRD Collaboration"
],
[
"Montier",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Morgante",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Mousset",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Nagata",
"R.",
"",
"for the LiteBIRD Collaboration"
],
[
"Namikawa",
"T.",
"",
"for the LiteBIRD Collaboration"
],
[
"Novelli",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Noviello",
"F.",
"",
"for the LiteBIRD Collaboration"
],
[
"Obata",
"I.",
"",
"for the LiteBIRD Collaboration"
],
[
"Occhiuzzi",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Paiella",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Paoletti",
"D.",
"",
"for the LiteBIRD Collaboration"
],
[
"Pascual-Cisneros",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Piacentini",
"F.",
"",
"for the LiteBIRD Collaboration"
],
[
"Pinchera",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Polenta",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Porcelli",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Remazeilles",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Ritacco",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Rizzieri",
"A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Rubiño-Martín",
"J. A.",
"",
"for the LiteBIRD Collaboration"
],
[
"Ruiz-Granda",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Sanghavi",
"J.",
"",
"for the LiteBIRD Collaboration"
],
[
"Sauvage",
"V.",
"",
"for the LiteBIRD Collaboration"
],
[
"Shiraishi",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Signorelli",
"G.",
"",
"for the LiteBIRD Collaboration"
],
[
"Stever",
"S. L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Sullivan",
"R. M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Takase",
"Y.",
"",
"for the LiteBIRD Collaboration"
],
[
"Tassis",
"K.",
"",
"for the LiteBIRD Collaboration"
],
[
"Terenzi",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"Tristram",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Vacher",
"L.",
"",
"for the LiteBIRD Collaboration"
],
[
"van Tent",
"B.",
"",
"for the LiteBIRD Collaboration"
],
[
"Vielva",
"P.",
"",
"for the LiteBIRD Collaboration"
],
[
"Wehus",
"I. K.",
"",
"for the LiteBIRD Collaboration"
],
[
"Zannoni",
"M.",
"",
"for the LiteBIRD Collaboration"
],
[
"Zhou",
"Y.",
"",
"for the LiteBIRD Collaboration"
]
] |
2507.07426
|
Zerui Yang
|
Zerui Yang and Yuwei Wan and Siyu Yan and Yudai Matsuda and Tong Xie and Bram Hoex and Linqi Song
|
DrugMCTS: a drug repurposing framework combining multi-agent, RAG and Monte Carlo Tree Search
| null | null | null | null |
cs.AI cs.CE
|
http://creativecommons.org/licenses/by/4.0/
|
Recent advances in large language models have demonstrated considerable potential in scientific domains such as drug repositioning. However, their effectiveness remains constrained when reasoning extends beyond the knowledge acquired during pretraining. Conventional approaches, such as fine-tuning or retrieval-augmented generation, face limitations in either imposing high computational overhead or failing to fully exploit structured scientific data. To overcome these challenges, we propose DrugMCTS, a novel framework that synergistically integrates RAG, multi-agent collaboration, and Monte Carlo Tree Search for drug repositioning. The framework employs five specialized agents tasked with retrieving and analyzing molecular and protein information, thereby enabling structured and iterative reasoning. Extensive experiments on the DrugBank and KIBA datasets demonstrate that DrugMCTS achieves substantially higher recall and robustness compared to both general-purpose LLMs and deep learning baselines. Our results highlight the importance of structured reasoning, agent-based collaboration, and feedback-driven search mechanisms in advancing LLM applications for drug repositioning.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 04:39:55 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Jul 2025 08:20:44 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 13:57:25 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Yang",
"Zerui",
""
],
[
"Wan",
"Yuwei",
""
],
[
"Yan",
"Siyu",
""
],
[
"Matsuda",
"Yudai",
""
],
[
"Xie",
"Tong",
""
],
[
"Hoex",
"Bram",
""
],
[
"Song",
"Linqi",
""
]
] |
2507.07526
|
Sheng Zhang
|
Cunhang Fan, Sheng Zhang, Jingjing Zhang, Enrui Liu, Xinhui Li, Minggang Zhao, Zhao Lv
|
DMF2Mel: A Dynamic Multiscale Fusion Network for EEG-Driven Mel Spectrogram Reconstruction
|
Accepted by ACM MM 2025
| null | null | null |
cs.SD eess.AS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Decoding speech from brain signals is a challenging research problem. Although existing technologies have made progress in reconstructing the mel spectrograms of auditory stimuli at the word or letter level, there remain core challenges in the precise reconstruction of minute-level continuous imagined speech: traditional models struggle to balance the efficiency of temporal dependency modeling and information retention in long-sequence decoding. To address this issue, this paper proposes the Dynamic Multiscale Fusion Network (DMF2Mel), which consists of four core components: the Dynamic Contrastive Feature Aggregation Module (DC-FAM), the Hierarchical Attention-Guided Multi-Scale Network (HAMS-Net), the SplineMap attention mechanism, and the bidirectional state space module (convMamba). Specifically, the DC-FAM separates speech-related "foreground features" from noisy "background features" through local convolution and global attention mechanisms, effectively suppressing interference and enhancing the representation of transient signals. HAMS-Net, based on the U-Net framework,achieves cross-scale fusion of high-level semantics and low-level details. The SplineMap attention mechanism integrates the Adaptive Gated Kolmogorov-Arnold Network (AGKAN) to combine global context modeling with spline-based local fitting. The convMamba captures long-range temporal dependencies with linear complexity and enhances nonlinear dynamic modeling capabilities. Results on the SparrKULee dataset show that DMF2Mel achieves a Pearson correlation coefficient of 0.074 in mel spectrogram reconstruction for known subjects (a 48% improvement over the baseline) and 0.048 for unknown subjects (a 35% improvement over the baseline).Code is available at: https://github.com/fchest/DMF2Mel.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 08:15:03 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:31:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Fan",
"Cunhang",
""
],
[
"Zhang",
"Sheng",
""
],
[
"Zhang",
"Jingjing",
""
],
[
"Liu",
"Enrui",
""
],
[
"Li",
"Xinhui",
""
],
[
"Zhao",
"Minggang",
""
],
[
"Lv",
"Zhao",
""
]
] |
2507.07675
|
Darshan Makwana
|
Darshan Makwana
|
Some Theoretical Results on Layerwise Effective Dimension Oscillations in Finite Width ReLU Networks
|
Incomplete citations
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
We analyze the layerwise effective dimension (rank of the feature matrix) in fully-connected ReLU networks of finite width. Specifically, for a fixed batch of $m$ inputs and random Gaussian weights, we derive closed-form expressions for the expected rank of the \$m\times n\$ hidden activation matrices. Our main result shows that $\mathbb{E}[EDim(\ell)]=m[1-(1-2/\pi)^\ell]+O(e^{-c m})$ so that the rank deficit decays geometrically with ratio $1-2 / \pi \approx 0.3634$. We also prove a sub-Gaussian concentration bound, and identify the "revival" depths at which the expected rank attains local maxima. In particular, these peaks occur at depths $\ell_k^*\approx(k+1/2)\pi/\log(1/\rho)$ with height $\approx (1-e^{-\pi/2}) m \approx 0.79m$. We further show that this oscillatory rank behavior is a finite-width phenomenon: under orthogonal weight initialization or strong negative-slope leaky-ReLU, the rank remains (nearly) full. These results provide a precise characterization of how random ReLU layers alternately collapse and partially revive the subspace of input variations, adding nuance to prior work on expressivity of deep networks.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 11:54:18 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:41:53 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Makwana",
"Darshan",
""
]
] |
2507.07695
|
Basem Suleiman PhD
|
Hruday Markondapatnaikuni, Basem Suleiman, Abdelkarim Erradi, Shijing Chen
|
KeyKnowledgeRAG (K^2RAG): An Enhanced RAG method for improved LLM question-answering capabilities
|
21 pages, 14 figures
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Fine-tuning is an immensely resource-intensive process when retraining Large Language Models (LLMs) to incorporate a larger body of knowledge. Although many fine-tuning techniques have been developed to reduce the time and computational cost involved, the challenge persists as LLMs continue to grow in size and complexity. To address this, a new approach to knowledge expansion in LLMs is needed. Retrieval-Augmented Generation (RAG) offers one such alternative by storing external knowledge in a database and retrieving relevant chunks to support question answering. However, naive implementations of RAG face significant limitations in scalability and answer accuracy. This paper introduces KeyKnowledgeRAG (K2RAG), a novel framework designed to overcome these limitations. Inspired by the divide-and-conquer paradigm, K2RAG integrates dense and sparse vector search, knowledge graphs, and text summarization to improve retrieval quality and system efficiency. The framework also includes a preprocessing step that summarizes the training data, significantly reducing the training time. K2RAG was evaluated using the MultiHopRAG dataset, where the proposed pipeline was trained on the document corpus and tested on a separate evaluation set. Results demonstrated notable improvements over common naive RAG implementations. K2RAG achieved the highest mean answer similarity score of 0.57, and reached the highest third quartile (Q3) similarity of 0.82, indicating better alignment with ground-truth answers. In addition to improved accuracy, the framework proved highly efficient. The summarization step reduced the average training time of individual components by 93%, and execution speed was up to 40% faster than traditional knowledge graph-based RAG systems. K2RAG also demonstrated superior scalability, requiring three times less VRAM than several naive RAG implementations tested in this study.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 12:19:03 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:57:40 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Markondapatnaikuni",
"Hruday",
""
],
[
"Suleiman",
"Basem",
""
],
[
"Erradi",
"Abdelkarim",
""
],
[
"Chen",
"Shijing",
""
]
] |
2507.07820
|
Eunsu Baek
|
Eunsu Baek, Keondo Park, Jeonggil Ko, Min-hwan Oh, Taesik Gong and Hyung-Sin Kim
|
AI Should Sense Better, Not Just Scale Bigger: Adaptive Sensing as a Paradigm Shift
| null | null | null | null |
cs.AI cs.LG
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Current AI advances largely rely on scaling neural models and expanding training datasets to achieve generalization and robustness. Despite notable successes, this paradigm incurs significant environmental, economic, and ethical costs, limiting sustainability and equitable access. Inspired by biological sensory systems, where adaptation occurs dynamically at the input (e.g., adjusting pupil size, refocusing vision)--we advocate for adaptive sensing as a necessary and foundational shift. Adaptive sensing proactively modulates sensor parameters (e.g., exposure, sensitivity, multimodal configurations) at the input level, significantly mitigating covariate shifts and improving efficiency. Empirical evidence from recent studies demonstrates that adaptive sensing enables small models (e.g., EfficientNet-B0) to surpass substantially larger models (e.g., OpenCLIP-H) trained with significantly more data and compute. We (i) outline a roadmap for broadly integrating adaptive sensing into real-world applications spanning humanoid, healthcare, autonomous systems, agriculture, and environmental monitoring, (ii) critically assess technical and ethical integration challenges, and (iii) propose targeted research directions, such as standardized benchmarks, real-time adaptive algorithms, multimodal integration, and privacy-preserving methods. Collectively, these efforts aim to transition the AI community toward sustainable, robust, and equitable artificial intelligence systems.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 14:50:32 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:44:36 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Baek",
"Eunsu",
""
],
[
"Park",
"Keondo",
""
],
[
"Ko",
"Jeonggil",
""
],
[
"Oh",
"Min-hwan",
""
],
[
"Gong",
"Taesik",
""
],
[
"Kim",
"Hyung-Sin",
""
]
] |
2507.07953
|
Mihails Milehins
|
Mihails Milehins, Dan B. Marghitu
|
Incremental Collision Laws Based on the Bouc-Wen Model: External Forces and Corner Cases
|
12 pages, 3 figures, see https://gitlab.com/user9716869/EBWCM ; (v2-v4) various amendments; arXiv admin note: text overlap with arXiv:2410.08147
| null | null | null |
physics.class-ph cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In the article titled "The Bouc-Wen Model for Binary Direct Collinear Collisions of Convex Viscoplastic Bodies" and published in the Journal of Computational and Nonlinear Dynamics (Volume 20, Issue 6, June 2025), the authors studied mathematical models of binary direct collinear collisions of convex viscoplastic bodies that employed two incremental collision laws based on the Bouc-Wen differential model of hysteresis. It was shown that the models possess favorable analytical properties, and several model parameter identification studies were conducted, demonstrating that the models can accurately capture the nature of a variety of collision phenomena. In this article, the aforementioned models are augmented by modeling the effects of external forces as time-dependent inputs that belong to a certain function space. Furthermore, the range of the parameters under which the models possess favorable analytical properties is extended to several corner cases that were not considered in the prior publication. Finally, the previously conducted model parameter identification studies are extended, and an additional model parameter identification study is provided in an attempt to validate the ability of the augmented models to represent the effects of external forces.
|
[
{
"version": "v1",
"created": "Thu, 10 Jul 2025 17:38:52 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Jul 2025 05:13:40 GMT"
},
{
"version": "v3",
"created": "Sun, 27 Jul 2025 18:08:16 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 06:21:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Milehins",
"Mihails",
""
],
[
"Marghitu",
"Dan B.",
""
]
] |
2507.08295
|
Dachun Yang
|
Jun Cao, Dachun Yang and Qishun Zhang
|
Equivalent Characterizations and Applications of Fractional Sobolev Spaces with Partially Vanishing Traces on $(\epsilon,\delta,D)$-Domains Supporting $D$-Adapted Fractional Hardy Inequalities
|
42 pages; Submitted
| null | null | null |
math.AP math.CA math.FA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Let $\Omega\subset\mathbb{R}^n$ be an $(\epsilon,\delta,D)$-domain, with $\epsilon\in(0,1]$, $\delta\in(0,\infty]$, and $D\subset \partial \Omega$ being a closed part of $\partial \Omega$, which is a general open connected set when $D=\partial \Omega$ and an $(\epsilon,\delta)$-domain when $D=\emptyset$. Let $s\in(0,1)$ and $p\in[1,\infty)$. If ${W}^{s,p}(\Omega)$, ${\mathcal W}^{s,p}(\Omega)$, and $\mathring{W}_D^{s,p}(\Omega)$ are the fractional Sobolev spaces on $\Omega$ that are defined respectively via the restriction of $W^{s,p}(\mathbb{R}^n)$ to $\Omega$, the intrinsic Gagliardo norm, and the completion of all $C^\infty(\Omega)$ functions with compact support away from $D$, in this article we prove their equivalences [that is, ${W}^{s,p}(\Omega)={\mathcal{W}}^{s,p}(\Omega) =\mathring{W}_D^{s,p}(\Omega)$] if $\Omega$ supports a $D$-adapted fractional Hardy inequality and, moreover, when $sp\ne 1$ such a fractional Hardy inequality is shown to be necessary to guarantee these equivalences under some mild geometric conditions on $\Omega$. Using the aforementioned equivalences, we show that the real interpolation space $(L^p(\Omega), \mathring{W}_D^{1,p}(\Omega))_{s,p}$ equals to some weighted fractional order Sobolev space $\mathcal{W}^{s,p}_{d_D^s}(\Omega)$ when $p\in (1,\infty)$. Applying this to the elliptic operator $\mathcal{L}_D$ in $\Omega$ with mixed boundary condition, we characterize both the domain of its fractional power and the parabolic maximal regularity of its Cauchy initial problem by means of $\mathcal{W}^{s,p}_{d_D^s}(\Omega)$.
|
[
{
"version": "v1",
"created": "Fri, 11 Jul 2025 04:01:49 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:54:42 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cao",
"Jun",
""
],
[
"Yang",
"Dachun",
""
],
[
"Zhang",
"Qishun",
""
]
] |
2507.08391
|
Davide Breoni PhD
|
Davide Breoni, Emanuele Locatelli and Luca Tubiana
|
Effects of knotting on the collapse of active ring polymers
|
34 pages, 15 figures
| null | null | null |
cond-mat.soft cond-mat.stat-mech
|
http://creativecommons.org/licenses/by/4.0/
|
We use numerical simulations to study tangentially active flexible ring polymers with different knot topologies. Simple, unknotted active rings display a transition from an extended phase to a collapsed one upon increasing the degree of polymerization. We find that topology has a significant effect on the polymer size at which the collapse takes place, with twist knots collapsing earlier than torus knots. Increasing knot complexity further accentuates this difference, as the collapse point of torus knots grows linearly with the minimum crossing number of the knot while that of twist knots shrinks, eventually canceling the actively stretched regime altogether. This behavior is a consequence of the ordered configuration of torus knots in their stretched active state, featuring an effective alignment for non-neighboring bonds which increases with the minimal crossing number. Twist knots do not feature ordered configurations or bond alignment, increasing the likelihood of collisions, leading to collapse. These results show that topology yields a degree of control on the properties of active ring polymers, and can be used to tune them. At the same time, they suggest that activity might introduce a bias for torus knots, as complex twist knots cannot be formed in extended active polymers.
|
[
{
"version": "v1",
"created": "Fri, 11 Jul 2025 08:02:56 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:58:58 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Breoni",
"Davide",
""
],
[
"Locatelli",
"Emanuele",
""
],
[
"Tubiana",
"Luca",
""
]
] |
2507.08540
|
Ioannis Lamprou
|
Ioannis Lamprou, Alexander Shevtsov, Ioannis Arapakis, Sotiris Ioannidis
|
White-Basilisk: A Hybrid Model for Code Vulnerability Detection
| null | null | null | null |
cs.CR cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The proliferation of software vulnerabilities presents a significant challenge to cybersecurity, necessitating more effective detection methodologies. We introduce White-Basilisk, a novel approach to vulnerability detection that demonstrates superior performance while challenging prevailing assumptions in AI model scaling. Utilizing an innovative architecture that integrates Mamba layers, linear self-attention, and a Mixture of Experts framework, White-Basilisk achieves state-of-the-art results in vulnerability detection tasks with a parameter count of only 200M. The model's capacity to process sequences of unprecedented length enables comprehensive analysis of extensive codebases in a single pass, surpassing the context limitations of current Large Language Models (LLMs). White-Basilisk exhibits robust performance on imbalanced, real-world datasets, while maintaining computational efficiency that facilitates deployment across diverse organizational scales. This research not only establishes new benchmarks in code security but also provides empirical evidence that compact, efficiently designed models can outperform larger counterparts in specialized tasks, potentially redefining optimization strategies in AI development for domain-specific applications.
|
[
{
"version": "v1",
"created": "Fri, 11 Jul 2025 12:39:25 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:49:27 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lamprou",
"Ioannis",
""
],
[
"Shevtsov",
"Alexander",
""
],
[
"Arapakis",
"Ioannis",
""
],
[
"Ioannidis",
"Sotiris",
""
]
] |
2507.08559
|
Jaewon Shim
|
Jaewon Shim (Seoul National University), Manhee Lee (Chungbuk National University), Wonho Jhe (Seoul National University)
|
Interfacial Entanglement-Induced Time-Dependent Solidification of Polymeric Fluids
|
This version has been accepted for publication in *Physics of Fluids* (Special Topic: Microrheology), July 2025. No content changes
|
Phys. Fluids 37 (8), (2005)
|
10.1063/5.0278221
| null |
physics.flu-dyn cond-mat.soft
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The structure of polymers at solid interfaces evolves over time, but the corresponding changes in their rheological properties remain poorly understood. Here, using a home-built quartz tuning fork atomic force microscope-based nano-rheometer, we directly measure the time-dependent viscoelasticity of the interfacial fluid. The bottommost layer, closest to the substrate, undergoes solidification over 10 hours, exhibiting an approximately five-fold increase in storage modulus and a two-fold increase in loss modulus. This arises from interfacial entanglement due to the strong binding of polymers to the solid surface driven by solid-wall attractive interactions. In contrast, within the second and third layers, the storage modulus remains nearly constant over time, while the loss modulus shows approximately two-fold increase. In this region, unlike the strongly bound first layer, entropic repulsion dominates, allowing the material to behave fluid-like while becoming increasingly viscous. Notably, as the first layer, where interfacial entanglement occurs, undergoes solidification, the flow boundary for interfacial fluid flow shifts upward away from the substrate, resulting in a negative slip length. This highlights the critical role of nanoscale interfacial structure and properties in governing macroscopic flow behavior.
|
[
{
"version": "v1",
"created": "Fri, 11 Jul 2025 12:59:26 GMT"
},
{
"version": "v2",
"created": "Sat, 26 Jul 2025 14:21:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Shim",
"Jaewon",
"",
"Seoul National University"
],
[
"Lee",
"Manhee",
"",
"Chungbuk National University"
],
[
"Jhe",
"Wonho",
"",
"Seoul National University"
]
] |
2507.08606
|
Benno Uthayasooriyar
|
Benno Uthayasooriyar, Antoine Ly, Franck Vermet, Caio Corro
|
DocPolarBERT: A Pre-trained Model for Document Understanding with Relative Polar Coordinate Encoding of Layout Structures
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
We introduce DocPolarBERT, a layout-aware BERT model for document understanding that eliminates the need for absolute 2D positional embeddings. We extend self-attention to take into account text block positions in relative polar coordinate system rather than the Cartesian one. Despite being pre-trained on a dataset more than six times smaller than the widely used IIT-CDIP corpus, DocPolarBERT achieves state-of-the-art results. These results demonstrate that a carefully designed attention mechanism can compensate for reduced pre-training data, offering an efficient and effective alternative for document understanding.
|
[
{
"version": "v1",
"created": "Fri, 11 Jul 2025 14:00:56 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Jul 2025 07:51:41 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 15:39:10 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Uthayasooriyar",
"Benno",
""
],
[
"Ly",
"Antoine",
""
],
[
"Vermet",
"Franck",
""
],
[
"Corro",
"Caio",
""
]
] |
2507.09252
|
Shukai Gong
|
Shukai Gong, Yiyang Fu, Fengyuan Ran, Quyu Kong, Feng Zhou
|
TPP-SD: Accelerating Transformer Point Process Sampling with Speculative Decoding
| null | null | null | null |
cs.LG stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose TPP-SD, a novel approach that accelerates Transformer temporal point process (TPP) sampling by adapting speculative decoding (SD) techniques from language models. By identifying the structural similarities between thinning algorithms for TPPs and speculative decoding for language models, we develop an efficient sampling framework that leverages a smaller draft model to generate multiple candidate events, which are then verified by the larger target model in parallel. TPP-SD maintains the same output distribution as autoregressive sampling while achieving significant acceleration. Experiments on both synthetic and real datasets demonstrate that our approach produces samples from identical distributions as standard methods, but with 2-6$\times$ speedup. Our ablation studies analyze the impact of hyperparameters such as draft length and draft model size on sampling efficiency. TPP-SD bridges the gap between powerful Transformer TPP models and the practical need for rapid sequence sampling.
|
[
{
"version": "v1",
"created": "Sat, 12 Jul 2025 11:18:07 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:50:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gong",
"Shukai",
""
],
[
"Fu",
"Yiyang",
""
],
[
"Ran",
"Fengyuan",
""
],
[
"Kong",
"Quyu",
""
],
[
"Zhou",
"Feng",
""
]
] |
2507.10073
|
Simon M\"unker
|
Simon M\"unker
|
Cultural Bias in Large Language Models: Evaluating AI Agents through Moral Questionnaires
|
15pages, 1 figure, 2 tables
|
Proceedings of 0th Symposium on Moral and Legal AI Alignment of the IACAP/AISB Conference, 2025
| null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Are AI systems truly representing human values, or merely averaging across them? Our study suggests a concerning reality: Large Language Models (LLMs) fail to represent diverse cultural moral frameworks despite their linguistic capabilities. We expose significant gaps between AI-generated and human moral intuitions by applying the Moral Foundations Questionnaire across 19 cultural contexts. Comparing multiple state-of-the-art LLMs' origins against human baseline data, we find these models systematically homogenize moral diversity. Surprisingly, increased model size doesn't consistently improve cultural representation fidelity. Our findings challenge the growing use of LLMs as synthetic populations in social science research and highlight a fundamental limitation in current AI alignment approaches. Without data-driven alignment beyond prompting, these systems cannot capture the nuanced, culturally-specific moral intuitions. Our results call for more grounded alignment objectives and evaluation metrics to ensure AI systems represent diverse human values rather than flattening the moral landscape.
|
[
{
"version": "v1",
"created": "Mon, 14 Jul 2025 08:59:26 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:13:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Münker",
"Simon",
""
]
] |
2507.10521
|
Takahiro Morishita
|
Takahiro Morishita, Zhaoran Liu, Massimo Stiavelli, Tommaso Treu, Pietro Bergamini, Yechi Zhang
|
Pristine Massive Star Formation Caught at the Break of Cosmic Dawn
|
Updated after addressing some comments. A new figure S6 was added showing Lya and Hb lines; Fig.3 was updated to reflect the updated Oxygen abundance measurement (12+log (O/H) < 5.8, at 2sigma); update to the physical size of the source; Hb EW0 measurement is quoted. None of the main conclusions changed
| null | null | null |
astro-ph.CO
|
http://creativecommons.org/licenses/by/4.0/
|
The existence of galaxies with no heavy elements is a key prediction of cosmological models. So far no "zero-metallicity", or Population III, galaxies have been identified. Here, we report the identification of an extremely metal-poor galaxy at redshift z = 5.725 ("AMORE6"), multiply imaged by a foreground galaxy cluster. JWST spectra consistently detect H$\beta$ at both positions, but [OIII]${\lambda}{\lambda}$4960,5008 remains undetected. This places a firm upper limit on its oxygen abundance, < 0.12 % of solar metallicity at 2${\sigma}$, establishing itself as the most pristine object by far. AMORE6 exhibits exceptional properties that indicate the presence of pristine massive star formation. The finding of such an example at a relatively late cosmic time is surprising, but it also validates the basic ideas behind the Big Bang model.
|
[
{
"version": "v1",
"created": "Mon, 14 Jul 2025 17:47:04 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:05:53 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Morishita",
"Takahiro",
""
],
[
"Liu",
"Zhaoran",
""
],
[
"Stiavelli",
"Massimo",
""
],
[
"Treu",
"Tommaso",
""
],
[
"Bergamini",
"Pietro",
""
],
[
"Zhang",
"Yechi",
""
]
] |
2507.10665
|
Stefano Palmisano
|
Stefano Palmisano, Diego Redigolo, Michele Tammaro, Andrea Tesi
|
Exploring ultra-high energy neutrino experiments through the lens of the transport equation
|
47 pages, 16 figures, 3 tables
| null | null | null |
hep-ph astro-ph.HE hep-ex
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We develop a first-principles formalism, based on the transport equation in the line-of-sight approximation, to link the expected number of muons at neutrino telescopes to the flux of neutrinos at the Earth's surface. We compute the distribution of muons inside Earth, arising from the up-scattering of neutrinos close to the detector, as well as from the decay of taus produced farther away. This framework allows one to account for systematic uncertainties, as well as to clarify the assumptions behind definitions commonly used in the literature, such as the effective area. We apply this formalism to analyze the high-energy muon event recorded by KM3NeT, with a reconstructed energy of $ 120^{+110}_{-60} \, \mathrm{PeV}$ and an elevation angle of $\left(0.54\pm 2.4\right)^\circ$, in comparison with the non-observation of similar events by IceCube. We find a $3.1\,\sigma$ tension between the two experiments, assuming a diffuse neutrino source with a power-law energy dependence. Combining both datasets leads to a preference for a very low number of expected events at KM3NeT, in stark contrast to the observed data. The tension increases both in the case of a diffuse source peaking at the KM3NeT energy and of a steady point source, whereas a transient source may reduce the tension down to $1.6\,\sigma$. The formalism allows one to treat potential beyond-the-Standard-Model sources of muons, and we speculate on this possibility to explain the tension.
|
[
{
"version": "v1",
"created": "Mon, 14 Jul 2025 18:00:05 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:50:10 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Palmisano",
"Stefano",
""
],
[
"Redigolo",
"Diego",
""
],
[
"Tammaro",
"Michele",
""
],
[
"Tesi",
"Andrea",
""
]
] |
2507.10769
|
Ferenc Kun
|
Csan\'ad Szuszik and Ferenc Kun
|
Control of fragment sizes of exploding rings
|
9 pages, 9 figures
|
International Journal of Solids and Structures 321, 113530 (2025)
|
10.1016/j.ijsolstr.2025.113530
| null |
cond-mat.dis-nn cond-mat.mtrl-sci cond-mat.stat-mech
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
We investigate the fragmentation of ring-like brittle structures under explosive loading using a discrete element model. By systematically varying ring thickness and strain rate, we uncover a transition from one-dimensional (1D) segmentation to two-dimensional (2D) planar fragmentation and, ultimately, to complete shattering. This transition is driven by the effective dimensionality of the crack pattern, which evolves with increasing strain rate. We identify a critical ring thickness beyond which segmentation ceases, and fragmentation directly follows a power-law mass distribution characteristic of 2D systems. In the crossover regime, spanning and non-spanning fragments coexist, enabling control over the power-law exponent of the mass distribution. At very high strain rates, we observe a transition to complete shattering, where the system follows a novel scaling law relating the shattering strain rate to ring thickness. Our results provide fundamental insights into fragmentation universality classes and offer potential applications in space debris prediction, controlled detonation technologies, and materials engineering.
|
[
{
"version": "v1",
"created": "Mon, 14 Jul 2025 19:47:51 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Szuszik",
"Csanád",
""
],
[
"Kun",
"Ferenc",
""
]
] |
2507.10817
|
Domenic Di Francesco
|
Domenic Di Francesco, Alan Forrest, Fiona McGarry, Nicholas Hall, Adam Sobey
|
Is Your Model Risk ALARP? Evaluating Prospective Safety-Critical Applications of Complex Models
| null | null | null | null |
stat.AP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The increasing availability of advanced computational modelling offers new opportunities to improve safety, efficacy, and emissions reductions. Application of complex models to support engineering decisions has been slow in comparison to other sectors, reflecting the higher consequence of unsafe applications. Adopting a complex model introduces a \emph{model risk}, namely the expected consequence of incorrect or otherwise unhelpful outputs. This should be weighed against the prospective benefits that the more sophisticated model can provide, also accounting for the non-zero risk of existing practice. Demonstrating when the model risk of a proposed machine learning application is As Low As Reasonably Practicable (ALARP) can help ensure that safety-critical industries benefit from complex models where appropriate while avoiding their misuse. An example of automated weld radiograph classification is presented to demonstrate how this can be achieved by combining statistical decision analysis, uncertainty quantification, and value of information.
|
[
{
"version": "v1",
"created": "Mon, 14 Jul 2025 21:34:29 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 14:18:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Di Francesco",
"Domenic",
""
],
[
"Forrest",
"Alan",
""
],
[
"McGarry",
"Fiona",
""
],
[
"Hall",
"Nicholas",
""
],
[
"Sobey",
"Adam",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.