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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2509.03419
|
Weiyuan Li
|
Weiyuan Li, Xintao Wang, Siyu Yuan, Rui Xu, Jiangjie Chen, Qingqing Dong, Yanghua Xiao, Deqing Yang
|
Curse of Knowledge: When Complex Evaluation Context Benefits yet Biases LLM Judges
|
8 pages, 4 figures, conference
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
As large language models (LLMs) grow more capable, they face increasingly diverse and complex tasks, making reliable evaluation challenging. The paradigm of LLMs as judges has emerged as a scalable solution, yet prior work primarily focuses on simple settings. Their reliability in complex tasks--where multi-faceted rubrics, unstructured reference answers, and nuanced criteria are critical--remains understudied. In this paper, we constructed ComplexEval, a challenge benchmark designed to systematically expose and quantify Auxiliary Information Induced Biases. We systematically investigated and validated 6 previously unexplored biases across 12 basic and 3 advanced scenarios. Key findings reveal: (1) all evaluated models exhibit significant susceptibility to these biases, with bias magnitude scaling with task complexity; (2) notably, Large Reasoning Models (LRMs) show paradoxical vulnerability. Our in-depth analysis offers crucial insights for improving the accuracy and verifiability of evaluation signals, paving the way for more general and robust evaluation models.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:48:33 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Weiyuan",
""
],
[
"Wang",
"Xintao",
""
],
[
"Yuan",
"Siyu",
""
],
[
"Xu",
"Rui",
""
],
[
"Chen",
"Jiangjie",
""
],
[
"Dong",
"Qingqing",
""
],
[
"Xiao",
"Yanghua",
""
],
[
"Yang",
"Deqing",
""
]
] |
2509.03420
|
Jeffrey Siewerdsen
|
Jeffrey H. Siewerdsen
|
Image-Guided Surgery: Technology, Quality, Innovation, and Opportunities for Medical Physics
|
20 pages, 6 figures
| null | null | null |
physics.med-ph eess.IV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The science and clinical practice of medical physics has been integral to the advancement of radiology and radiation therapy for over a century. In parallel, advances in surgery - including intraoperative imaging, registration, and other technologies within the expertise of medical physicists - have advanced primarily in connection to other disciplines, such as biomedical engineering and computer science, and via somewhat distinct translational paths. This review article briefly traces the parallel and convergent evolution of such scientific, engineering, and clinical domains with an eye to a potentially broader, more impactful role of medical physics in research and clinical practice of surgery. A review of image-guided surgery technologies is offered, including intraoperative imaging, tracking / navigation, image registration, visualization, and surgical robotics across a spectrum of surgical applications. Trends and drivers for research and innovation are traced, including federal funding and academic-industry partnership, and some of the major challenges to achieving major clinical impact are described. Opportunities for medical physicists to expand expertise and contribute to the advancement of surgery in the decade ahead are outlined, including research and innovation, data science approaches, improving efficiency through operations research and optimization, improving patient safety, and bringing rigorous quality assurance to technologies and processes in the circle of care for surgery. Challenges abound but appear tractable, including domain knowledge, professional qualifications, and the need for investment and clinical partnership.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:48:33 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Siewerdsen",
"Jeffrey H.",
""
]
] |
2509.03421
|
Yukun Zhou
|
Yukun Zhou, Paul Nderitu, Jocelyn Hui Lin Goh, Justin Engelmann, Siegfried K. Wagner, Anran Ran, Hongyang Jiang, Lie Ju, Ke Zou, Sahana Srinivasan, Hyunmin Kim, Takahiro Ninomiya, Zheyuan Wang, Gabriel Dawei Yang, Eden Ruffell, Dominic Williamson, Rui Santos, Gabor Mark Somfai, Carol Y. Cheung, Tien Yin Wong, Daniel C. Alexander, Yih Chung Tham, Pearse A. Keane
|
Generalist versus Specialist Vision Foundation Models for Ocular Disease and Oculomics
|
39 pages, 8 Figures
| null | null | null |
eess.IV cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Medical foundation models, pre-trained with large-scale clinical data, demonstrate strong performance in diverse clinically relevant applications. RETFound, trained on nearly one million retinal images, exemplifies this approach in applications with retinal images. However, the emergence of increasingly powerful and multifold larger generalist foundation models such as DINOv2 and DINOv3 raises the question of whether domain-specific pre-training remains essential, and if so, what gap persists. To investigate this, we systematically evaluated the adaptability of DINOv2 and DINOv3 in retinal image applications, compared to two specialist RETFound models, RETFound-MAE and RETFound-DINOv2. We assessed performance on ocular disease detection and systemic disease prediction using two adaptation strategies: fine-tuning and linear probing. Data efficiency and adaptation efficiency were further analysed to characterise trade-offs between predictive performance and computational cost. Our results show that although scaling generalist models yields strong adaptability across diverse tasks, RETFound-DINOv2 consistently outperforms these generalist foundation models in ocular-disease detection and oculomics tasks, demonstrating stronger generalisability and data efficiency. These findings suggest that specialist retinal foundation models remain the most effective choice for clinical applications, while the narrowing gap with generalist foundation models suggests that continued data and model scaling can deliver domain-relevant gains and position them as strong foundations for future medical foundation models.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:48:57 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhou",
"Yukun",
""
],
[
"Nderitu",
"Paul",
""
],
[
"Goh",
"Jocelyn Hui Lin",
""
],
[
"Engelmann",
"Justin",
""
],
[
"Wagner",
"Siegfried K.",
""
],
[
"Ran",
"Anran",
""
],
[
"Jiang",
"Hongyang",
""
],
[
"Ju",
"Lie",
""
],
[
"Zou",
"Ke",
""
],
[
"Srinivasan",
"Sahana",
""
],
[
"Kim",
"Hyunmin",
""
],
[
"Ninomiya",
"Takahiro",
""
],
[
"Wang",
"Zheyuan",
""
],
[
"Yang",
"Gabriel Dawei",
""
],
[
"Ruffell",
"Eden",
""
],
[
"Williamson",
"Dominic",
""
],
[
"Santos",
"Rui",
""
],
[
"Somfai",
"Gabor Mark",
""
],
[
"Cheung",
"Carol Y.",
""
],
[
"Wong",
"Tien Yin",
""
],
[
"Alexander",
"Daniel C.",
""
],
[
"Tham",
"Yih Chung",
""
],
[
"Keane",
"Pearse A.",
""
]
] |
2509.03422
|
Mohsen Rahmani Haghighi
|
Mohammad Hossein Zarei, Mohsen Rahmani Haghighi
|
Universal representation of the long-range entanglement in the family of Toric Code states
|
Submitted to Europhysics Letters (EPL), 7 pages, 4 figures
| null | null | null |
quant-ph cond-mat.str-el
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Since the long range entanglement is a universal characteristic of topological quantum states belonging to the same class, a suitable mathematical representation of the long range entanglement has to be also universal. In this Letter, we introduce such a representation for the family of Toric Code states by using Kitaev's Ladders as building blocks. We consider Toric Code states corresponding to various planar graphs and apply non-local dientanglers to qubits corresponding to non-contractible cycles that satisfy a topological constraint. We demonstrate that, independent of the geometry of the underlying graph, disentanglers convert Toric Code states into a tensor product of Kitaev's Ladder states. Since Kitaev's Ladders with arbitrary geometric configurations include the short-range entanglements, we conclude that the above universal and non-local pattern of entanglement between ladders is responsible of the long-range entanglement inherent in Toric Code states. Our result emphasizes in the capability of such non-local representations to describe topological order in ground-state wave functions of topological quantum systems.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:49:57 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zarei",
"Mohammad Hossein",
""
],
[
"Haghighi",
"Mohsen Rahmani",
""
]
] |
2509.03423
|
Loredana Gastaldo
|
Fabienne Adam, Felix Ahrens, Luis E. Ardila Perez, Matthias Balzer, Arnulf Barth, Daniel Behrend-Uriarte, Sebastian Berndt, Klaus Blaum, Frederic W. H. B\"ohm, Martin Bra{\ss}, Lorenzo Calza, Katerina Chrysalidis, Menno Door, Holger Dorrer, Christoph E. D\"ullmann, Klaus Eberhardt, Sergey Eliseev, Christian Enss, Pavel Filianin, Andreas Fleischmann, Robert Gartmann, Loredana Gastaldo, Markus Griedel, Alexander G\"oggelmann, Robert Hammann, Raphael Hasse, Maurits W. Haverkort, Simon Heinze, Daniel Hengstler, Rasmus Jeske, Josef Jochum, Karl Johnston, Nick Karcher, Sebastian Kempf, Tom Kieck, Ulli K\"oster, Neven Kovac, Nina Kneip, Kathrin Kromer, Federica Mantegazzini, Bruce A. Marsh, Marc Merstof, Timo Muscheid, Martin Neidig, Yuri N. Novikov, Raghav Pandey, Andreas Reifenberger, Daniel Richter, Alexander Rischka, Sebastian Rothe, Oliver Sander, Rima X. Sch\"ussler, Stephan Scholl, Christoph Schweiger, Clemens Velte, Marc Weber, Matthias Wegner, Klaus Wendt and Tom Wickenh\"auser
|
Improved limit on the effective electron neutrino mass with the ECHo-1k experiment
| null | null | null | null |
hep-ex nucl-ex
|
http://creativecommons.org/licenses/by/4.0/
|
The effective electron neutrino mass can be determined by analyzing the endpoint region of the $^{163}$Ho electron capture spectrum, provided a measurement with high energy resolution and high statistics using calorimetric techniques. Here, the Electron Capture in $^{163}$Ho collaboration, ECHo, presents an analysis of the most precise $^{163}$Ho spectrum currently available, obtained with the ECHo-1k experiment and comprising about 200 million events. A very low background rate of $B=9.1(1.3)\times 10^{-6}$ /eV/pixel/day was achieved allowing for a reliable analysis of the endpoint region. The derived endpoint energy $Q = 2862(4)$ eV is in excellent agreement with the one independently determined via Penning-trap mass spectrometry of $Q=2863.2(6)$ eV [1]. The upper limit of the effective electron neutrino mass is improved by almost a factor 2 compared to the lowest current value [2], reaching $m_{\nu_\mathrm{e}} < 15 $ eV/c${^2}$ (90\% credible interval).
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:50:44 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Adam",
"Fabienne",
""
],
[
"Ahrens",
"Felix",
""
],
[
"Perez",
"Luis E. Ardila",
""
],
[
"Balzer",
"Matthias",
""
],
[
"Barth",
"Arnulf",
""
],
[
"Behrend-Uriarte",
"Daniel",
""
],
[
"Berndt",
"Sebastian",
""
],
[
"Blaum",
"Klaus",
""
],
[
"Böhm",
"Frederic W. H.",
""
],
[
"Braß",
"Martin",
""
],
[
"Calza",
"Lorenzo",
""
],
[
"Chrysalidis",
"Katerina",
""
],
[
"Door",
"Menno",
""
],
[
"Dorrer",
"Holger",
""
],
[
"Düllmann",
"Christoph E.",
""
],
[
"Eberhardt",
"Klaus",
""
],
[
"Eliseev",
"Sergey",
""
],
[
"Enss",
"Christian",
""
],
[
"Filianin",
"Pavel",
""
],
[
"Fleischmann",
"Andreas",
""
],
[
"Gartmann",
"Robert",
""
],
[
"Gastaldo",
"Loredana",
""
],
[
"Griedel",
"Markus",
""
],
[
"Göggelmann",
"Alexander",
""
],
[
"Hammann",
"Robert",
""
],
[
"Hasse",
"Raphael",
""
],
[
"Haverkort",
"Maurits W.",
""
],
[
"Heinze",
"Simon",
""
],
[
"Hengstler",
"Daniel",
""
],
[
"Jeske",
"Rasmus",
""
],
[
"Jochum",
"Josef",
""
],
[
"Johnston",
"Karl",
""
],
[
"Karcher",
"Nick",
""
],
[
"Kempf",
"Sebastian",
""
],
[
"Kieck",
"Tom",
""
],
[
"Köster",
"Ulli",
""
],
[
"Kovac",
"Neven",
""
],
[
"Kneip",
"Nina",
""
],
[
"Kromer",
"Kathrin",
""
],
[
"Mantegazzini",
"Federica",
""
],
[
"Marsh",
"Bruce A.",
""
],
[
"Merstof",
"Marc",
""
],
[
"Muscheid",
"Timo",
""
],
[
"Neidig",
"Martin",
""
],
[
"Novikov",
"Yuri N.",
""
],
[
"Pandey",
"Raghav",
""
],
[
"Reifenberger",
"Andreas",
""
],
[
"Richter",
"Daniel",
""
],
[
"Rischka",
"Alexander",
""
],
[
"Rothe",
"Sebastian",
""
],
[
"Sander",
"Oliver",
""
],
[
"Schüssler",
"Rima X.",
""
],
[
"Scholl",
"Stephan",
""
],
[
"Schweiger",
"Christoph",
""
],
[
"Velte",
"Clemens",
""
],
[
"Weber",
"Marc",
""
],
[
"Wegner",
"Matthias",
""
],
[
"Wendt",
"Klaus",
""
],
[
"Wickenhäuser",
"Tom",
""
]
] |
2509.03424
|
\'Oscar Jim\'enez-Arranz
|
\'O. Jim\'enez-Arranz, S. Lilleengen, M. S. Petersen
|
Beyond the Clouds: S3 as the most distant extended Milky Way stream, not of LMC origin
|
12 pages. Submitted to A&A, comments are welcomed!
| null | null | null |
astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
Context: While the LMC's influence on MW stellar streams has been extensively studied, streams associated with the Clouds have received far less attention. Beyond the Magellanic Stream, only four stream candidates (S1-S4) have been reported. Aims: We focus on the S3 stream, a long ($\sim30^\circ$) and narrow ($\sim1.2^\circ$) structure at 60-80 kpc, nearly aligned with the LMC. Our goals are: 1) to validate the stream through a kinematic analysis of S3 candidates with Gaia DR3 data; 2) to enlarge the sample of potential members; and 3) to model the stream in order to test its association with either the MW or the LMC. Methods: We selected new S3 candidates with a neural network classifier trained on Gaia DR3 data, and further reduced contamination through a cut in the proper-motion space. To investigate the origin of S3, we evolve stream models within time-dependent, deforming MW and LMC haloes, thereby accounting for possible effects of the MW-LMC interaction. Results: We identify 1,542 high-confidence new S3 stream candidates and find that the stream's apparent width has grown from $\sim1.2^\circ$ to $\sim3$-$4^\circ$ compared to previous studies. We also present a list of 440 potential S3 red clump stars, which are valuable targets for spectroscopic follow-up thanks to their well-defined luminosities and ability to yield precise distances. Both modelling and a comparison of S3 stars' closest approach distance and velocity with the LMC's escape velocity indicate that S3 is unlikely to originate from the LMC, instead representing a distant ($\sim75$ kpc) MW stream. Conclusions: S3 is the most distant ($\sim75$ kpc) extended ($\sim30^\circ$ long, $\sim3$-$4^\circ$ thick) MW stream known, offering a unique probe of the outer halo and the LMC's recent influence. Its angular width corresponds to a physical thickness of $\sim4$-5 kpc, making S3 among the thickest streams discovered.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:52:44 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Jiménez-Arranz",
"Ó.",
""
],
[
"Lilleengen",
"S.",
""
],
[
"Petersen",
"M. S.",
""
]
] |
2509.03425
|
Truong-Son Hy
|
Phuc Pham, Viet Thanh Duy Nguyen, Truong-Son Hy
|
LINKER: Learning Interactions Between Functional Groups and Residues With Chemical Knowledge-Enhanced Reasoning and Explainability
| null | null | null | null |
cs.LG q-bio.QM
|
http://creativecommons.org/licenses/by/4.0/
|
Accurate identification of interactions between protein residues and ligand functional groups is essential to understand molecular recognition and guide rational drug design. Existing deep learning approaches for protein-ligand interpretability often rely on 3D structural input or use distance-based contact labels, limiting both their applicability and biological relevance. We introduce LINKER, the first sequence-based model to predict residue-functional group interactions in terms of biologically defined interaction types, using only protein sequences and the ligand SMILES as input. LINKER is trained with structure-supervised attention, where interaction labels are derived from 3D protein-ligand complexes via functional group-based motif extraction. By abstracting ligand structures into functional groups, the model focuses on chemically meaningful substructures while predicting interaction types rather than mere spatial proximity. Crucially, LINKER requires only sequence-level input at inference time, enabling large-scale application in settings where structural data is unavailable. Experiments on the LP-PDBBind benchmark demonstrate that structure-informed supervision over functional group abstractions yields interaction predictions closely aligned with ground-truth biochemical annotations.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:53:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Pham",
"Phuc",
""
],
[
"Nguyen",
"Viet Thanh Duy",
""
],
[
"Hy",
"Truong-Son",
""
]
] |
2509.03426
|
Aj Piergiovanni
|
AJ Piergiovanni, Ganesh Satish Mallya, Dahun Kim, Anelia Angelova
|
Time-Scaling State-Space Models for Dense Video Captioning
|
BMVC 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Dense video captioning is a challenging video understanding task which aims to simultaneously segment the video into a sequence of meaningful consecutive events and to generate detailed captions to accurately describe each event. Existing methods often encounter difficulties when working with the long videos associated with dense video captioning, due to the computational complexity and memory limitations. Furthermore, traditional approaches require the entire video as input, in order to produce an answer, which precludes online processing of the video. We address these challenges by time-scaling State-Space Models (SSMs) to even longer sequences than before. Our approach, State-Space Models with Transfer State, combines both the long-sequence and recurrent properties of SSMs and addresses the main limitation of SSMs which are otherwise not able to sustain their state for very long contexts, effectively scaling SSMs further in time. The proposed model is particularly suitable for generating captions on-the-fly, in an online or streaming manner, without having to wait for the full video to be processed, which is more beneficial in practice. When applied to dense video captioning, our approach scales well with video lengths and uses 7x fewer FLOPs.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:56:20 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Piergiovanni",
"AJ",
""
],
[
"Mallya",
"Ganesh Satish",
""
],
[
"Kim",
"Dahun",
""
],
[
"Angelova",
"Anelia",
""
]
] |
2509.03427
|
Ivan Costa
|
Pedro Correia and Ivan Silva and Ivone Amorim and Eva Maia and Isabel Pra\c{c}a
|
Federated Learning: An approach with Hybrid Homomorphic Encryption
|
19 pages, 8 figures, To be published in the conference Security and Trust Management(STM), ESORICS 2025
| null | null | null |
cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
Federated Learning (FL) is a distributed machine learning approach that promises privacy by keeping the data on the device. However, gradient reconstruction and membership-inference attacks show that model updates still leak information. Fully Homomorphic Encryption (FHE) can address those privacy concerns but it suffers from ciphertext expansion and requires prohibitive overhead on resource-constrained devices. We propose the first Hybrid Homomorphic Encryption (HHE) framework for FL that pairs the PASTA symmetric cipher with the BFV FHE scheme. Clients encrypt local model updates with PASTA and send both the lightweight ciphertexts and the PASTA key (itself BFV-encrypted) to the server, which performs a homomorphic evaluation of the decryption circuit of PASTA and aggregates the resulting BFV ciphertexts. A prototype implementation, developed on top of the Flower FL framework, shows that on independently and identically distributed MNIST dataset with 12 clients and 10 training rounds, the proposed HHE system achieves 97.6% accuracy, just 1.3% below plaintext, while reducing client upload bandwidth by over 2,000x and cutting client runtime by 30% compared to a system based solely on the BFV FHE scheme. However, server computational cost increases by roughly 15621x for each client participating in the training phase, a challenge to be addressed in future work.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:58:22 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Correia",
"Pedro",
""
],
[
"Silva",
"Ivan",
""
],
[
"Amorim",
"Ivone",
""
],
[
"Maia",
"Eva",
""
],
[
"Praça",
"Isabel",
""
]
] |
2509.03428
|
Felipe Herrera
|
Athul S. Rema and Adri\'an E. Rubio L\'opez and Felipe Herrera
|
Ultrafast single-photon interference with a dipole qubit in a nanocavity
|
11 pages, 8 figures, 4 appendices
| null | null | null |
quant-ph cond-mat.mes-hall physics.atom-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The stationary spectrum of individual dipole emitters in plasmonic nanocavities has been studied for a range of cavity geometries and dipole configurations. Less is known about the coherent dynamics of single photon creation in the nanocavity near field by an excited dipole. We address this gap by developing a Lorentzian kernel approximation that solves the time-dependent Schr\"odinger equation that describes the coupled dipole-photon dynamics in the single-excitation manifold. Our approach encodes the broadband nature of the nanocavity field through a non-Markovian memory kernel, derived from macroscopic QED theory. For a two-level dipole near a metallic nanosphere, we show that the single photon probability density in frequency space evolves in strong coupling from an initially localized source at the qubit frequency into a Rabi doublet over a timescale governed by the kernel spectrum. This dynamical crossover is accompanied by the formation of single-photon interference patterns in frequency and time, propagating coherently over a timescale limited by the shape of kernel spectrum to $\sim 100-150$ fs, which is accessible to ultrafast spectroscopy. We also show that the stationary spectrum of the coupled system can be manipulated by driving the nanocavity field using coherent pulses with variable spectral bandwidth. Using single-photon pulses narrower than the kernel spectrum, the Rabi splitting in a system that supports strong coupling can be effectively removed. The applicability of our results to other dipole-nanocavity configurations is discussed and a general strong coupling criterion for nanocavities is formulated.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:59:09 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Rema",
"Athul S.",
""
],
[
"López",
"Adrián E. Rubio",
""
],
[
"Herrera",
"Felipe",
""
]
] |
2509.03429
|
Margot Coste-Sarguet
|
M. Coste-Sarguet, J. P. Graves
|
The properties of resistive MHD modes and unstable spectra in advanced tokamak regimes
|
28 pages, 22 figures, submitted to Plasma Physics and Controlled Fusion in Jul. 2025
| null | null | null |
physics.plasm-ph math-ph math.MP
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Advanced tokamak regimes, featuring extended regions of low magnetic shear, are promising candidates for future fusion reactors but are also more prone to specific kinds of MHD instabilities. The proximity to a rational surface in a very low shear region weakens field line bending stabilisation and amplifies the effects of toroidal coupling between modes, leading to the emergence of long-wavelength resistive infernal modes. These modes can grow collectively as a discrete spectrum, leading to a cascade of different perturbations for single mode numbers $(m, n)$, with subdominant modes showing increasingly oscillatory radial structures. These spectra of fast-growing modes are significant for developing stable scenarios in future reactors, and for the understanding of global reconnection events like sawteeth, motivating a deeper investigation into their fundamental physics. Deriving new analytic solutions, including a generalisation of the ideal interchange dispersion relation to non monotonic $q$ profiles, and extending a modular linear resistive MHD solver, we investigate how resistivity, compressibility, toroidal effects, and shaping influence stability, especially in reversed shear $q$ profiles. It is also shown that common assumptions in numerical calculations prevent the observation of the full variety of modes present in these advanced scenarios.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:59:24 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Coste-Sarguet",
"M.",
""
],
[
"Graves",
"J. P.",
""
]
] |
2509.03430
|
Vimal Mollyn
|
Vimal Mollyn, Nathan DeVrio, Chris Harrison
|
EclipseTouch: Touch Segmentation on Ad Hoc Surfaces using Worn Infrared Shadow Casting
|
Accepted to UIST 2025
| null | null | null |
cs.HC cs.CV cs.GR cs.RO
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
The ability to detect touch events on uninstrumented, everyday surfaces has been a long-standing goal for mixed reality systems. Prior work has shown that virtual interfaces bound to physical surfaces offer performance and ergonomic benefits over tapping at interfaces floating in the air. A wide variety of approaches have been previously developed, to which we contribute a new headset-integrated technique called \systemname. We use a combination of a computer-triggered camera and one or more infrared emitters to create structured shadows, from which we can accurately estimate hover distance (mean error of 6.9~mm) and touch contact (98.0\% accuracy). We discuss how our technique works across a range of conditions, including surface material, interaction orientation, and environmental lighting.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 15:59:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Mollyn",
"Vimal",
""
],
[
"DeVrio",
"Nathan",
""
],
[
"Harrison",
"Chris",
""
]
] |
2509.03431
|
Hemanta Kunwar
|
George Avalos, Pelin G. Geredeli, Hemanta Kunwar, Hyesuk Lee
|
A novel approach to study the wellposedness of the 3D fluid-2D plate interaction PDE System
|
23 pages, 4 figures
| null | null | null |
math.AP cs.NA math.NA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We consider a certain fluid-structure interaction (FSI) system with a view of obtaining an alternative methodology for establishing its strongly continuous semigroup wellposedness. (Semigroup generation for this FSI was originally considered in Avalos-Clark (2014).) The FSI model under consideration describes the vibrations of an incompressible fluid within a 3D cavity as it interacts with the elastic membrane on the ``free" upper boundary of the cavity. Such coupled PDE systems appear in variety of natural settings such as biomedicine, aeroelasticity, and fluid dynamics.
Our proof of $C_0$-semigroup wellposedness is based on a proper application of Lumer Phillips Theorem. In this regard, our main challenge is to show the maximality of the corresponding semigroup generator. To this end, we develop a ``nonstandard" inf-sup approach which avoids the use of technical nonlocal maps in the associated bilinear forms--unlike the earlier paper Avalos-Clark (2014)--and allows for the solution of the fluid and plate solution variables simultanously. Our new inf-sup strategy will lead to a more efficient mixed finite element method (FEM) for approximating solutions to the FSI problem, inasmuch our novel variational formulation avoids bilinear forms which are free from the computationally-intensive nonlocal solution operators invoked in Avalos-Clark (2014). We also perform numerical tests based on this formulation using a benchmark problem
and present numerical results to demonstrate the effectiveness of our approach.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:00:42 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Avalos",
"George",
""
],
[
"Geredeli",
"Pelin G.",
""
],
[
"Kunwar",
"Hemanta",
""
],
[
"Lee",
"Hyesuk",
""
]
] |
2509.03432
|
Jeffrey Banks
|
J. W. Banks and J. Shatah
|
A New Approach to Direct Discretization of Wave Kinetic Equations with Application to a Nonlinear Schrodinger System in 2D
| null | null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
Wave Kinetic Equations (WKEs) are often used to describe the evolution of ensemble averaged wave amplitudes for nonlinear wave systems. In the present manuscript we describe a new approach to direct numerical simulation of solutions to WKEs. This new method relies on a piecewise polynomial approximation of the resonant manifold, followed by numerical quadrature of the collision integral. The approach is general in nature, and is discussed in detail here for a particular nonlinear Schrodinger model in 2 spatial dimensions. Detailed convergence studies demonstrate 2nd-order accuracy for model collision integrals, and self-convergence studies for the WKE show near 2nd-order rates. Furthermore, comparison of the WKE approximation to ensemble averages of the NLS illustrate the efficacy of the method and the validity of the WKE, for both isotropic and an-isotropic solutions.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:02:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Banks",
"J. W.",
""
],
[
"Shatah",
"J.",
""
]
] |
2509.03433
|
Kaili Sun
|
Kaili sun, Xingyu Miao, Bing Zhai, Haoran Duan and Yang Long
|
Decoding Visual Neural Representations by Multimodal with Dynamic Balancing
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
In this work, we propose an innovative framework that integrates EEG, image, and text data, aiming to decode visual neural representations from low signal-to-noise ratio EEG signals. Specifically, we introduce text modality to enhance the semantic correspondence between EEG signals and visual content. With the explicit semantic labels provided by text, image and EEG features of the same category can be more closely aligned with the corresponding text representations in a shared multimodal space. To fully utilize pre-trained visual and textual representations, we propose an adapter module that alleviates the instability of high-dimensional representation while facilitating the alignment and fusion of cross-modal features. Additionally, to alleviate the imbalance in multimodal feature contributions introduced by the textual representations, we propose a Modal Consistency Dynamic Balance (MCDB) strategy that dynamically adjusts the contribution weights of each modality. We further propose a stochastic perturbation regularization (SPR) term to enhance the generalization ability of semantic perturbation-based models by introducing dynamic Gaussian noise in the modality optimization process. The evaluation results on the ThingsEEG dataset show that our method surpasses previous state-of-the-art methods in both Top-1 and Top-5 accuracy metrics, improving by 2.0\% and 4.7\% respectively.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:03:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"sun",
"Kaili",
""
],
[
"Miao",
"Xingyu",
""
],
[
"Zhai",
"Bing",
""
],
[
"Duan",
"Haoran",
""
],
[
"Long",
"Yang",
""
]
] |
2509.03434
|
Elias Zikkos
|
Elias Zikkos
|
On Markushevich bases $\{x^{\lambda_n}\}_{n=1}^{\infty}$ for their closed span in weighted $L^2 (A)$ spaces over sets $A\subset [0,\infty)$ of positive Lebesgue measure, hereditary completeness, and moment problems
|
21 pages
| null | null | null |
math.FA math.CA
|
http://creativecommons.org/licenses/by/4.0/
|
Inspired by the work of Borwein and Erdelyi \cite{BE1997JAMS} on generalizations of M\"{u}ntz's theorem, we investigate the properties of the system $\{x^{\lambda_n}\}_{n=1}^{\infty}$ in weighted $L^p (A)$ spaces, for $p\ge 1$, denoted by $L^p_w (A)$, where
(I) $A$ is a measurable subset of the real half-line $[0,\infty)$ having positive Lebesgue measure,
(II) $w$ is a non-negative integrable function defined on $A$, and
(III) $\{\lambda_n\}_{n=1}^{\infty}$ is a strictly increasing sequence of positive real numbers such that
$\inf\{\lambda_{n+1}-\lambda_n \}>0$ and $\sum_{n=1}^{\infty}\lambda_n^{-1}<\infty$.
We prove that a function $f$ in $\overline{\text{span}}\{x^{\lambda_n}\}_{n=1}^{\infty}$ in the Hilbert space $L^2_w (A)$, admits the $\bf{Fourier-type}$ series representation $f(x)=\sum_{n=1}^{\infty} \langle f, r_n\rangle_{w,A} x^{\lambda_n}$ a.e on $A$, where $\{r_n\}_{n=1}^{\infty}$ is the unique biorthogonal family of $\{x^{\lambda_n}\}_{n=1}^{\infty}$ in $\overline{\text{span}}\{x^{\lambda_n}\}_{n=1}^{\infty}$ in $L^2_w (A)$. As a result, we show that the system $\{x^{\lambda_n}\}_{n=1}^{\infty}$ is a $\bf{Markushevich\,\, basis}$ for $\overline{\text{span}}\{x^{\lambda_n}\}_{n=1}^{\infty}$ in $L^2_w (A)$. Furthermore, we consider a $\bf{moment\,\, problem}$. Finally, if $m\le w(x)\le M$ on $A$ for some positive numbers $m$ and $M$ and the set $A$ contains an interval $[a, r_A]$, where $a\ge 0$ and $r_A$ is the essential supremum of $A$, we prove that the system $\{x^{\lambda_n}\}_{n=1}^{\infty}$ is $\bf{hereditarily\,\, complete}$ in $\overline{\text{span}}\{x^{\lambda_n}\}_{n=1}^{\infty}$ in the space $L^2_w(A)$. As a result, a general class of compact operators on the closure is constructed that admit spectral synthesis.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:05:43 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zikkos",
"Elias",
""
]
] |
2509.03435
|
Bengt Gustafsson
|
Bengt Gustafsson
|
Is the composition of the Solar atmosphere unusual, and if so, why? Possible interpretations
|
52 pages, 13 figures
| null | null | null |
astro-ph.SR astro-ph.EP
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The ongoing discussion about the atomic chemical composition of the Sun is commented on. The main focus in this review is on the deviation of the solar composition from that of most other solar-type stars in that its ratio of volatiles (like the elements C, N, O, S, P and Zn) to the refractories (most metals, like Ba, Ca, Ti, Y, Al, Sc and Zr) tends to be higher in the Sun by 10 to 20%. What does this tell about the formation and evolution of the Solar System? Scenarios in terms of galactic evolution, formation of the pre-solar nebula, of the evolution of the protoplanetary disk, of the engulfing of planets, and of other processes within the Solar System are considered, as well as the evolution of binary stars with similarly different chemical composition. Finally, implications, if any, on the habitability of the Solar System are commented on.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:06:30 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Gustafsson",
"Bengt",
""
]
] |
2509.03436
|
Md. Maruf
|
Md Mhamud Hussen Sifat, Md Maruf, and Md Rokunuzzaman
|
Cost-Optimized Systems Engineering for IoT-Enabled Robot Nurse in Infectious Pandemic Management
|
11 pages, 10 figures, 4 tables, 1 algorithm. Corresponding author: Md Maruf (maruf.mte.17@gmail.com)
| null | null | null |
cs.RO cs.HC cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The utilization of robotic technology has gained traction in healthcare facilities due to progress in the field that enables time and cost savings, minimizes waste, and improves patient care. Digital healthcare technologies that leverage automation, such as robotics and artificial intelligence, have the potential to enhance the sustainability and profitability of healthcare systems in the long run. However, the recent COVID-19 pandemic has amplified the need for cyber-physical robots to automate check-ups and medication administration. A robot nurse is controlled by the Internet of Things (IoT) and can serve as an automated medical assistant while also allowing supervisory control based on custom commands. This system helps reduce infection risk and improves outcomes in pandemic settings. This research presents a test case with a nurse robot that can assess a patient's health status and take action accordingly. We also evaluate the system's performance in medication administration, health-status monitoring, and life-cycle considerations.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:06:39 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sifat",
"Md Mhamud Hussen",
""
],
[
"Maruf",
"Md",
""
],
[
"Rokunuzzaman",
"Md",
""
]
] |
2509.03437
|
Konstantin Stepanyantz
|
Gleb Kovyrshin, Nikolai Meshcheriakov, Victoria Shatalova, Konstantin Stepanyantz
|
Structure of renormalization constants for theories with multiple couplings in the MS-like subtraction schemes
|
23 pages
| null | null | null |
hep-th hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
For theories with multiple couplings we construct simple expressions for the four-dimensional (or, in general, integer-dimensional) renormalization constants assuming that all divergences are logarithmical. These expressions allow relating all coefficients at $\varepsilon$-poles, logarithms, and (if exist) mixed terms to the coefficients of the renormalization group functions in any order of the perturbation theory for MS-like renormalization prescriptions. The result admits such a formulation in that $\varepsilon$-poles and $\ln\Lambda/\mu$ enter on the same footing. For theories with two and three couplings we present explicit expressions for the pole/logarithm structure of renormalization constants in the lowest orders of the perturbation theory. They are verified by comparisons with the three-loop explicit calculations made previously for the $\varphi^4$-theory with two couplings.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:07:07 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Kovyrshin",
"Gleb",
""
],
[
"Meshcheriakov",
"Nikolai",
""
],
[
"Shatalova",
"Victoria",
""
],
[
"Stepanyantz",
"Konstantin",
""
]
] |
2509.03438
|
Otmane Sakhi
|
Benjamin Heymann and Otmane Sakhi
|
Non-Linear Counterfactual Aggregate Optimization
|
Recsys '25, CONSEQUENCES: Causality, Counterfactuals & Sequential Decision-Making Workshop
| null | null | null |
stat.ML cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
We consider the problem of directly optimizing a non-linear function of an outcome, where this outcome itself is the sum of many small contributions. The non-linearity of the function means that the problem is not equivalent to the maximization of the expectation of the individual contribution. By leveraging the concentration properties of the sum of individual outcomes, we derive a scalable descent algorithm that directly optimizes for our stated objective. This allows for instance to maximize the probability of successful A/B test, for which it can be wiser to target a success criterion, such as exceeding a given uplift, rather than chasing the highest expected payoff.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:07:34 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Heymann",
"Benjamin",
""
],
[
"Sakhi",
"Otmane",
""
]
] |
2509.03439
|
Yunfan Zhao
|
Yunfan Zhao, Xiaojing Chen, Wenwen Pan
|
Quantitative Stability and Contraction Principles for Mean-Field G-SDEs
| null | null | null | null |
q-fin.MF math.PR
|
http://creativecommons.org/licenses/by/4.0/
|
We study mean-field stochastic differential equations (SDEs) driven by G-Brownian motion, extending recent work on existence and uniqueness by developing a full quantitative stability framework. Our main contribution is the construction of an intrinsic stability modulus that provides explicit bounds on the sensitivity of solutions with respect to perturbations in initial data (and, indirectly, coefficients). Using Bihari-Osgood type inequalities under G-expectation, we establish sharp continuity estimates for the data-to-solution map and analyze the asymptotic properties of the stability modulus. In particular, we identify contraction behavior on short horizons, leading to a contraction principle that guarantees uniqueness and global propagation of stability. The results apply under non-Lipschitz, non-deterministic coefficients with square-integrable initial data, thereby significantly broadening the scope of mean-field G-SDEs. Beyond existence and uniqueness, our framework quantifies robustness of solutions under volatility uncertainty, opening new directions for applications in stochastic control, risk management, and mean-field models under ambiguity.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:07:35 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhao",
"Yunfan",
""
],
[
"Chen",
"Xiaojing",
""
],
[
"Pan",
"Wenwen",
""
]
] |
2509.03440
|
Neda Sadooghi
|
M. Kiamari and N. Sadooghi
|
Wigner function of a rigidly rotating and magnetized QED plasma
|
12 pages, no figures
| null | null | null |
gr-qc hep-ph hep-th nucl-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We determine the Wigner function of a rigidly rotating quantum electrodynamics (QED) plasma in the presence of a constant magnetic field by utilizing the Riemannian normal coordinate approximation, which has been previously proposed in the literature. In this approach, the angular velocity appears only in a specific phase factor, allowing us to compute the point-split fermion two-point correlation function in flat spacetime. To ensure that the fermion correlation function is gauge invariant, we introduce a background gauge field that is fixed to produce a constant magnetic field. Using this Wigner function, we derive the energy-momentum tensor for this medium, which consists of both diagonal and off-diagonal components. By comparing our results with the energy-momentum tensor of an ideal spinful and vortical magnetized fluid, we establish a connection between these components and thermodynamic quantities, such as energy density and different types of pressure. We demonstrate that rigid rotation leads to pressure anisotropy in plasma. Additionally, we compute the associated vector and axial vector currents for this medium, utilizing the previously presented Wigner function. Our results are consistent with existing literature on the subject.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:07:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Kiamari",
"M.",
""
],
[
"Sadooghi",
"N.",
""
]
] |
2509.03441
|
Pasquale Di Bari
|
Bhupal Dev, Pasquale Di Bari, Ivan Martinez-Soler and Rishav Roshan
|
Boomerang mechanism explaining the excess radio background
|
7 pages, 4 figures
| null | null | null |
hep-ph astro-ph.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a {\em boomerang mechanism} for the explanation of the excess radio background detected by ARCADE. In an early stage, at a temperature $T \sim 100\,{\rm keV}$, a fraction of relic neutrinos is resonantly converted into dark neutrinos by mixing induced by a pre-existing lepton asymmetry. Dark neutrinos decay much later into a dark photon, mixed with photon, and a dark fermion, with a lifetime longer than the age of the Universe, as required by a solution to the excess radio background. This scenario circumvents the upper bound on the neutrino magnetic moment but still implies a testable lower bound.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:09:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Dev",
"Bhupal",
""
],
[
"Di Bari",
"Pasquale",
""
],
[
"Martinez-Soler",
"Ivan",
""
],
[
"Roshan",
"Rishav",
""
]
] |
2509.03442
|
Fangtian Zhong
|
Zhuoyun Qian, Hongyi Miao, Yili Jiang, Qin Hu, Jiaqi Huang, Cheng Zhang, Fangtian Zhong
|
Evaluating Diverse Feature Extraction Techniques of Multifaceted IoT Malware Analysis: A Survey
| null | null | null | null |
cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
As IoT devices continue to proliferate, their reliability is increasingly constrained by security concerns. In response, researchers have developed diverse malware analysis techniques to detect and classify IoT malware. These techniques typically rely on extracting features at different levels from IoT applications, giving rise to a wide range of feature extraction methods. However, current approaches still face significant challenges when applied in practice. This survey provides a comprehensive review of feature extraction techniques for IoT malware analysis from multiple perspectives. We first examine static and dynamic feature extraction methods, followed by hybrid approaches. We then explore feature representation strategies based on graph learning. Finally, we compare the strengths and limitations of existing techniques, highlight open challenges, and outline promising directions for future research.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:10:35 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Qian",
"Zhuoyun",
""
],
[
"Miao",
"Hongyi",
""
],
[
"Jiang",
"Yili",
""
],
[
"Hu",
"Qin",
""
],
[
"Huang",
"Jiaqi",
""
],
[
"Zhang",
"Cheng",
""
],
[
"Zhong",
"Fangtian",
""
]
] |
2509.03443
|
Tarek Bazizi
|
Tarek Bazizi, Mohamed Maghenem, Paolo Frasca, Antonio Lor\`ia, Elena Panteley
|
On the Perturbed Projection-Based Distributed Gradient-Descent Algorithm: A Fully-Distributed Adaptive Redesign
| null | null | null | null |
math.OC cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this work, we revisit a classical distributed gradient-descent algorithm, introducing an interesting class of perturbed multi-agent systems. The state of each subsystem represents a local estimate of a solution to the global optimization problem. Thereby, the network is required to minimize local cost functions, while gathering the local estimates around a common value. Such a complex task suggests the interplay of consensus-based dynamics with gradient-descent dynamics. The latter descent dynamics involves the projection operator, which is assumed to provide corrupted projections of a specific form, reminiscent of existing (fast) projection algorithms. Hence, for the resulting class of perturbed networks, we are able to adaptively tune some gains in a fully distributed fashion, to approach the optimal consensus set up to arbitrary-desired precision.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:11:30 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Bazizi",
"Tarek",
""
],
[
"Maghenem",
"Mohamed",
""
],
[
"Frasca",
"Paolo",
""
],
[
"Lorìa",
"Antonio",
""
],
[
"Panteley",
"Elena",
""
]
] |
2509.03444
|
Haiyuan Feng
|
Haiyuan Feng, Rong-Jia Yang
|
The stability of the de-Sitter universe in nonlocal gravity
|
0
| null | null | null |
gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We constructed the ghost-free condition for nonlocal gravity using de-Sitter background field expansion and identified the structure of the nontrivial form factors. Our analysis shows that the particle spectrum of this model is nearly equivalent to general relativity (GR), with the potential addition of a scalar particle with positive mass m. Additionally, by employing recursion relations, we established the equivalence between nonlocal gravity and higher-derivative gravity. Moreover, we provided a comprehensive non-perturbative proof of the stability of de-Sitter solution within the nonlocal framework.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:12:12 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Feng",
"Haiyuan",
""
],
[
"Yang",
"Rong-Jia",
""
]
] |
2509.03445
|
Julian Mayr
|
Julian Mayr, Maria Laura Staffini, Simon B. J\"ager, Corinna Kollath and Jonathan Keeling
|
Phase Diagram and dynamical phases of self organization of a Bose--Einstein condensate in a transversely pumped red-detuned cavity
|
18 Pages, 11 Figures
| null | null | null |
cond-mat.quant-gas quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study a transversely pumped atomic Bose--Einstein Condensate coupled to a single-mode optical cavity, where effective atom--atom interactions are mediated by pump and cavity photons. A number of experiments and theoretical works have shown the formation of a superradiant state in this setup, where interference of pump and cavity light leads to an optical lattice in which atoms self-consistently organize. This self-organization has been extensively studied using the approximate Dicke model (truncating to two momentum states), as well as through numerical Gross--Pitaevskii simulations in one and two dimensions. Here, we perform a full mean-field analysis of the system, including all relevant atomic momentum states and the cavity field. We map out the steady-state phase diagram vs pump strength and cavity detuning, and provide an in-depth understanding of the instabilities that are linked to the emergence of spatio-temporal patterns. We find and describe parameter regimes where mean-field predicts bistability, regimes where the dynamics form chaotic trajectories, instabilities caused by resonances between normal mode excitations, and states with atomic dynamics but vanishing cavity field.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:12:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Mayr",
"Julian",
""
],
[
"Staffini",
"Maria Laura",
""
],
[
"Jäger",
"Simon B.",
""
],
[
"Kollath",
"Corinna",
""
],
[
"Keeling",
"Jonathan",
""
]
] |
2509.03446
|
Niteesh Midlagajni
|
Niteesh Midlagajni, Constantin A. Rothkopf
|
Graph neural networks for learning liquid simulations in dynamic scenes containing kinematic objects
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Simulating particle dynamics with high fidelity is crucial for solving real-world interaction and control tasks involving liquids in design, graphics, and robotics. Recently, data-driven approaches, particularly those based on graph neural networks (GNNs), have shown progress in tackling such problems. However, these approaches are often limited to learning fluid behavior in static free-fall environments or simple manipulation settings involving primitive objects, often overlooking complex interactions with dynamically moving kinematic rigid bodies. Here, we propose a GNN-based framework designed from the ground up to learn the dynamics of liquids under rigid body interactions and active manipulations, where particles are represented as graph nodes and particle-object collisions are handled using surface representations with the bounding volume hierarchy (BVH) algorithm. This approach enables the network to model complex interactions between liquid particles and intricate surface geometries. Our model accurately captures fluid behavior in dynamic settings and can also function as a simulator in static free-fall environments. Despite being trained on a single-object manipulation task of pouring, our model generalizes effectively to environments with unseen objects and novel manipulation tasks such as stirring and scooping. Finally, we show that the learned dynamics can be leveraged to solve control and manipulation tasks using gradient-based optimization methods.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:13:52 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Midlagajni",
"Niteesh",
""
],
[
"Rothkopf",
"Constantin A.",
""
]
] |
2509.03447
|
Zbigniew Jelonek
|
A. Fernandes, Z. Jelonek, J. E. Sampaio
|
Metric version of the Zariski Multiplicity Conjecture is true for multiplicity two
| null | null | null | null |
math.CV
|
http://creativecommons.org/licenses/by/4.0/
|
We prove that if two algebraic $(n-1)$-dimensional cones $P, R\subset\mathbb C^n$ are ambient homeomorphic, then their bases $B_P$ and $B_R$ have the same Euler characteristic. As an application, we show our main result: if $(X,0),(Y,0)\subset (\C^n,0)$ are germs of analytic hypersurfaces, which are ambient bi-Lipschitz equivalent and $m_0(X)=2$, then also $m_0(Y)=2.$. As another application, we also obtain a complete positive answer to the Metric Arnold-Vassiliev Problem, which says that if two holomorphic functions have zeros that are ambient bi-Lipschitz invariant, then the Hessian matrices of the functions have the same rank.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:14:05 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fernandes",
"A.",
""
],
[
"Jelonek",
"Z.",
""
],
[
"Sampaio",
"J. E.",
""
]
] |
2509.03448
|
Amir Shee
|
Anweshika Pattanayak, Amir Shee, Debasish Chaudhuri, Abhishek Chaudhuri
|
Chirality, confinement and dimensionality govern re-entrant transitions in active matter
|
10 pages, 3 figures
| null | null | null |
cond-mat.soft cond-mat.stat-mech physics.bio-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The non-equilibrium dynamics of individual chiral active particles underpin the complex behavior of chiral active matter. Here we present an exact analytical framework, supported by simulations, to characterize the steady states of two-dimensional chiral active Brownian particles and three-dimensional torque-driven counterparts in a harmonic trap. Using a Laplace-transform approach of the Fokker-Planck equation, we derive closed-form expressions for displacement moments and excess kurtosis, providing a precise probe of non-Gaussian statistics. Our analysis reveals three distinct regimes: bimodal active states with off-center peaks, Gaussian-like passive states, and weakly heavy-tailed distributions unique to two dimensions. We show that dimensionality plays a decisive role: in two dimensions, increasing chirality suppresses activity and restores passive behavior, while in three dimensions torque preserves activity along the torque axis, producing anisotropic steady states. These behaviors are captured by simple active length-scale arguments that map the boundaries between passive and active phases. Our results offer concrete experimental signatures - including kurtosis crossovers, off-center peaks, and torque-induced anisotropy - that establish confinement as a powerful tool to probe and control chiral and torque-driven active matter.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:14:21 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Pattanayak",
"Anweshika",
""
],
[
"Shee",
"Amir",
""
],
[
"Chaudhuri",
"Debasish",
""
],
[
"Chaudhuri",
"Abhishek",
""
]
] |
2509.03449
|
Vishwajeet Kumar
|
Vishwajeet Kumar, Ohad Shpielberg and Arnab Pal
|
Speeding up Brownian escape via intermediate finite potential barriers
| null | null | null | null |
cond-mat.stat-mech cond-mat.soft physics.chem-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The mean first-passage time (MFPT) for a Brownian particle to surmount a potential barrier of height $\Delta U$ is a fundamental quantity governing a wide array of physical and chemical processes. According to the Arrhenius Law, the MFPT typically grows exponentially with increasing barrier height, reflecting the rarity of thermally activated escape events. In this work, we demonstrate that the MFPT can be significantly reduced by reshaping the original single-barrier potential into a structured energy landscape comprising multiple intermediate barriers of lower heights, while keeping the total barrier height $\Delta U$ unchanged. Furthermore, this counterintuitive result holds across both linear and nonlinear potential profiles. Our findings suggest that tailoring the energy landscape -- by introducing well-placed intermediate barriers -- can serve as an effective control strategy to accelerate thermally activated transitions. These predictions are amenable to experimental validation using optical trapping techniques.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:14:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Kumar",
"Vishwajeet",
""
],
[
"Shpielberg",
"Ohad",
""
],
[
"Pal",
"Arnab",
""
]
] |
2509.03450
|
V. Hinich
|
Vladimir Hinich and Ieke Moerdijk
|
Addendum: localization of lax symmetric monoidal categories
|
6 pages
| null | null | null |
math.CT
|
http://creativecommons.org/licenses/by/4.0/
|
In this note, we explain in some detail how one can fiberwise localize a (co)lax symmetric monoidal infinity-category. This construction was tacitly used in Section 5 of our recent paper "On the equivalence of the Lurie's infinity-operads and dendroidal infinity-operads".
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:15:43 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Hinich",
"Vladimir",
""
],
[
"Moerdijk",
"Ieke",
""
]
] |
2509.03451
|
Vimal Mollyn
|
Nathan DeVrio, Vimal Mollyn, Chris Harrison
|
SmartPoser: Arm Pose Estimation with a Smartphone and Smartwatch Using UWB and IMU Data
|
The first two listed authors contributed equally. Published at UIST 2023
| null |
10.1145/3586183.3606821
| null |
cs.HC cs.CV cs.GR cs.RO
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
The ability to track a user's arm pose could be valuable in a wide range of applications, including fitness, rehabilitation, augmented reality input, life logging, and context-aware assistants. Unfortunately, this capability is not readily available to consumers. Systems either require cameras, which carry privacy issues, or utilize multiple worn IMUs or markers. In this work, we describe how an off-the-shelf smartphone and smartwatch can work together to accurately estimate arm pose. Moving beyond prior work, we take advantage of more recent ultra-wideband (UWB) functionality on these devices to capture absolute distance between the two devices. This measurement is the perfect complement to inertial data, which is relative and suffers from drift. We quantify the performance of our software-only approach using off-the-shelf devices, showing it can estimate the wrist and elbow joints with a \hl{median positional error of 11.0~cm}, without the user having to provide training data.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:16:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"DeVrio",
"Nathan",
""
],
[
"Mollyn",
"Vimal",
""
],
[
"Harrison",
"Chris",
""
]
] |
2509.03452
|
Nicholas Chan
|
Nicholas Chan, Ashkan Hashemi, Soheil Assadi, Pierre Quenneville, Charles Clifton, Gregory MacRae, Rajesh Dhakal, Liang-Jiu Jia
|
Shake-Table Tests of a Three-Storey Steel Structure with Resilient Slip-Friction Joints (RSFJ)
| null | null | null | null |
physics.app-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Shake-table tests were conducted as part of the Robust Building Systems (ROBUST) program to evaluate and demonstrate the resilience of various structural concepts under earthquake excitations. This paper presents the results of 3 different configurations incorporating the Resilient Slip-Friction Joint (RSFJ) in: (1) tension-only braces (TOB), (2) tension-compression braces (TCB), and (3) moment-resisting frame (MRF) joints along the longitudinal direction. Each building configuration was subjected to the El Centro ground motion at increasing intensities, reaching a peak ground acceleration (PGA) of about 0.5 g. The peak displacements sustained were well within design values, with the top floor deflecting by 0.66%, 0.98% and 1.26% of the total height for the TOB, TCB and MRF systems respectively. The corresponding inter-storey drifts peaked at 0.87%, 1.11% and 1.62%. In all cases, the structures essentially self-centered with residual drifts limited to 0.03%. The floor accelerations observed in unidirectional tests were 0.67 g (TOB), 0.66 g (TCB) and 0.62 g (MRF). The spikes may occur during (a) direction reversal at peak Mode 1 amplitude, (b) under the influence of higher Mode 2, and/or (c) stiffness transition at the upright/plumb position. In bidirectional tests, torsion effects contributed to the largest acceleration recorded at 0.95 g (TCB). Unbalanced resistance/deformation of symmetric friction connections in the transverse direction led to the structure twisting and loading the longitudinal RSFJs unevenly. However, the effects on peak displacements were not significant and similar responses as unidirectional tests indicate effective performance under simultaneous out-of-plane shaking. The results demonstrate the RSFJ's ability to damp seismically induced vibrations and restore structures to their original, undeformed shape after earthquakes.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:17:23 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chan",
"Nicholas",
""
],
[
"Hashemi",
"Ashkan",
""
],
[
"Assadi",
"Soheil",
""
],
[
"Quenneville",
"Pierre",
""
],
[
"Clifton",
"Charles",
""
],
[
"MacRae",
"Gregory",
""
],
[
"Dhakal",
"Rajesh",
""
],
[
"Jia",
"Liang-Jiu",
""
]
] |
2509.03453
|
Shuyuan Xiao
|
Yifan Li, Yuhan Yang, Qiegen Liu, Shuyuan Xiao, Tingting Liu
|
Four-channel Imaging Based on Reconfigurable Metasurfaces: Hyperchaotic Encryption under Physical Protection
| null | null | null | null |
physics.optics physics.app-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Metasurfaces facilitate high-capacity optical information integration by simultaneously supporting near-field nanoprinting and far-field holography on a single platform. However, conventional multi-channel designs face critical security vulnerabilities for sensitive information due to insufficient encryption mechanisms. In this work, we propose a four-channel phase-change metasurface featuring algorithm-physical co-security-a dual-protection framework combining intrinsic metasurface physical security with chaotic encryption. Our polarization-multiplexed metasurface generates four optical imaging channels through meta-atom design, including two far-field holograms and two near-field patterns. To enhance system security, we apply Chen hyperchaotic encryption combined with the Logistic map and DNA encoding to convert near-field information into secure QR codes; far-field holograms are retained to demonstrate the metasurface's information capacity and for attack detection. Phase-change metasurface further provides physical-layer security by dynamically switching imaging channels via crystalline-to-amorphous state transitions, enhancing anti-counterfeiting and reliability. The proposed metasurface achieves high-fidelity imaging, robust anti-attack performance, and independent channel control. This integrated approach pioneers a secure paradigm for high-density optical information processing.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:20:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Yifan",
""
],
[
"Yang",
"Yuhan",
""
],
[
"Liu",
"Qiegen",
""
],
[
"Xiao",
"Shuyuan",
""
],
[
"Liu",
"Tingting",
""
]
] |
2509.03454
|
Leandro Lichtenfelz
|
James Benn, Patrick Heslin, Leandro Lichtenfelz, Gerard Misiolek
|
Singular Sets of Riemannian Exponential Maps in Hydrodynamics
|
22 pages, 3 figures
| null | null | null |
math.DG math.FA
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
We prove that the singular sets for the Lagrangian solution maps of the two-dimensional inviscid Euler and generalized surface quasi-geostrophic equations are Gaussian null sets. To achieve this we carry out a spectral analysis of an operator related to the coadjoint representation of the algebra of divergence-free vector fields on the fluid domain. In particular, we establish sharp results on the Schatten-von Neumann class to which this operator belongs. Furthermore, its failure to be compact is directly connected to the absence of Fredholm properties of the corresponding Lagrangian solution maps. We show that, for the three-dimensional Euler equations and the standard surface quasi-geostrophic equation, this failure is in fact essential.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:20:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Benn",
"James",
""
],
[
"Heslin",
"Patrick",
""
],
[
"Lichtenfelz",
"Leandro",
""
],
[
"Misiolek",
"Gerard",
""
]
] |
2509.03455
|
Paul B. Rimmer
|
Maja W. Radecka and Paul B. Rimmer
|
Nitrogen chemistry of hycean worlds on the example of K2-18b
|
21 pages, 23 figures, accepted for publication in MNRAS
| null | null | null |
astro-ph.EP
|
http://creativecommons.org/licenses/by/4.0/
|
A recent observation of the exoplanet K2-18b sparked interest among scientists - large amounts of carbon dioxide and methane were detected in an H2-rich background atmosphere. If the planet is a hycean world (liquid water ocean + hydrogen-dominated atmosphere), it could be habitable under certain conditions. The presence of carbon, hydrogen and oxygen was already confirmed, however, there was no detection of nitrogen or its compounds. Molecular nitrogen is difficult to detect directly. This study concentrates on possible photochemical products of N2 such as HCN, NH3 and HC3N. We set approximate limits on the amount of nitrogen bearing species by varying atmospheric parameters, such as the Eddy Diffusion coefficient and the amount of N2 present from 10 ppm to 10%. If the bulk nitrogen-containing gas in the atmosphere is N2, photochemistry produces only trace amounts of the aforementioned species. However, if ammonia is the main source of nitrogen, then the quantities of NH3, CH5N and HCN approach detectable range. HC3N and NO are bad tracers of the nitrogen source in the atmosphere, because they are produced in similar amounts in all tested scenarios. Assuming equilibrium chemistry at the surface of K2-18b results in underprediction of CO2 abundance. This result combined with the non-detection of ammonia by JWST suggests the planet is not a typical sub-Neptune, but could be indeed a hycean world or magma ocean planet. We also found that C2H6 is produced in significant amounts - if it is detected in the future, it could serve as a proxy for DMS presence.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:24:52 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Radecka",
"Maja W.",
""
],
[
"Rimmer",
"Paul B.",
""
]
] |
2509.03456
|
Otmane Sakhi
|
Imad Aouali and Otmane Sakhi
|
Off-Policy Learning in Large Action Spaces: Optimization Matters More Than Estimation
|
Recsys '25, CONSEQUENCES: Causality, Counterfactuals & Sequential Decision-Making Workshop
| null | null | null |
stat.ML cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Off-policy evaluation (OPE) and off-policy learning (OPL) are foundational for decision-making in offline contextual bandits. Recent advances in OPL primarily optimize OPE estimators with improved statistical properties, assuming that better estimators inherently yield superior policies. Although theoretically justified, we argue this estimator-centric approach neglects a critical practical obstacle: challenging optimization landscapes. In this paper, we provide theoretical insights and extensive empirical evidence showing that current OPL methods encounter severe optimization issues, particularly as action spaces become large. We demonstrate that simpler weighted log-likelihood objectives enjoy substantially better optimization properties and still recover competitive, often superior, learned policies. Our findings emphasize the necessity of explicitly addressing optimization considerations in the development of OPL algorithms for large action spaces.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:25:45 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Aouali",
"Imad",
""
],
[
"Sakhi",
"Otmane",
""
]
] |
2509.03457
|
Ruichang Guo
|
Ruichang Guo, Hongsheng Wang, Reza Ershadnia, Seyyed Hosseini
|
Pore-scale insights into the role of micro fractures on permeability of fractured porous media
| null | null | null | null |
physics.geo-ph
|
http://creativecommons.org/publicdomain/zero/1.0/
|
Fractures play a critical role in governing fluid flow within subsurface energy systems, including oil and gas production, geologic carbon sequestration, and underground hydrogen storage. This study investigated the impact of pore-scale fractures on fluid flow and permeability in fractured porous media. The analysis focused on a single fracture embedded within a porous medium. Fluid flow was simulated using the lattice Boltzmann method, and the effects of fracture length, width, and orientation angle on permeability were systematically examined. Results showed that increasing both fracture length and width enhanced permeability. Additionally, fractures oriented more closely to the flow direction (i.e., smaller orientation angles) resulted in higher permeability. Interestingly, when the orientation angle approached 90{\deg}, the presence of a fracture could reduce the overall permeability of the porous medium. A critical orientation angle was identified, beyond which the fracture decreased permeability; this critical angle was found to increase with fracture width. Permeability tensors were also fitted to determine the critical angle and quantify the influence of fracture width on the critical orientation angle. These findings provide new insights into the role of microfractures in controlling permeability, with important implications for subsurface energy systems.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:30:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Guo",
"Ruichang",
""
],
[
"Wang",
"Hongsheng",
""
],
[
"Ershadnia",
"Reza",
""
],
[
"Hosseini",
"Seyyed",
""
]
] |
2509.03458
|
Aleksandra Kusiak
|
Michael Jacob Rodriguez, Aleksandra Kusiak, Shivam Pandey, J. Colin Hill
|
Comparison of Halo Model and Simulation Predictions for Projected-Field Kinematic Sunyaev-Zel'dovich Cross-Correlations
| null | null | null | null |
astro-ph.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The kinematic Sunyaev-Zel'dovich (kSZ) effect in the cosmic microwave background (CMB) is a powerful probe of gas physics and large-scale structure (LSS) in our universe. We consider the "projected-field" kSZ estimator, which involves cross-correlating a foreground-cleaned, filtered, squared CMB temperature map with an LSS tracer, and requires no individual tracer redshifts. We compare $\verb|class_sz|$ halo model calculations of projected-field kSZ cross-correlations with measurements of these signals from the Websky numerical simulations. We cross-correlate halo density maps from Websky with various CMB secondary signals. We first validate our halo model by comparing its predictions for thermal SZ (tSZ) and patchy screening ($\tau$) cross-correlations to measurements of these signals from Websky. We consider three different halo redshift ranges in our comparisons. We also construct our own kSZ, tSZ, and $\tau$ maps to validate the form of the relevant profiles. Following the tSZ and $\tau$ validation, we compare projected-field kSZ calculations between the halo model and the simulations. We use filters constructed for $\textit{Planck}$ and the Simons Observatory (SO) to assess the accuracy of the halo-model kSZ predictions for experiments of differing sensitivity. Overall, we find good agreement, particularly at $\textit{Planck}$ sensitivity. However, we find an $\approx$ 20$\%$ difference between our halo model and the simulations for SO, which significantly exceeds the predicted error bars. We note that our halo model includes only the dominant expected term in the projected-field kSZ signal; the magnitude of the difference between our model and the simulations is consistent with previous predictions for terms arising from other contractions in the theory calculation. These terms will need to be included to obtain unbiased inference from upcoming projected-field kSZ measurements.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:33:04 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Rodriguez",
"Michael Jacob",
""
],
[
"Kusiak",
"Aleksandra",
""
],
[
"Pandey",
"Shivam",
""
],
[
"Hill",
"J. Colin",
""
]
] |
2509.03459
|
Elsa Carmen Barrio Torres Miss
|
Elsa Barrio-Torres, Jes\'us Abaurrea, Jes\'us As\'in, Jorge Castillo-Mateo, Ana Carmen Cebri\'an, Zeus Gracia
|
Prediction of Maximum Temperature record in Spain 1960-2023 by the means of ERA5 atmospheric geopotentials
| null | null | null | null |
stat.AP
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The increasing frequency of extreme temperature events, such as daily maximum temperature ($T_x$) records, underscores the need for robust tools to understand their drivers and predict their occurrence. Previous studies have identified increasing and non-stationary trends in $T_x$ records across the Iberian Peninsula, particularly during summer, the literature directly exploring their connection with upper-level atmospheric covariates remains limited. This work develops and applies an innovative methodological framework to model the occurrence of $T_x$ records and their relationship with geopotential height fields. We used daily $T_x$ data from 36 Spanish stations (1960-2023) provided by ECA&D and geopotential height data at 300, 500, and 700 hPa from ERA5. Exploratory analysis revealed a non-stationary trend in records, a higher frequency in the interior of the peninsula, and decreasing spatial co-occurrence with distance. We designed a hierarchical spatio-temporal logistic regression algorithm prioritizing interpretability and high-dimensionality reduction. The approach involves: (1) fitting local models per station; (2) applying a spatial consensus filter based on statistical significance to reduce the initial 1620 covariates to 17 in a base model (M1); and (3) a controlled incorporation of interaction terms. Among the tested models, a global model (M2) that enhances M1 with geodetic interactions was selected for its optimal balance between predictive performance (AUC) and complexity. Model M2 demonstrates high predictive accuracy at interior stations and good performance at coastal stations. It also adequately reproduces key observed properties, including the persistence of record streaks and patterns of spatial co-occurrence. This study provides a novel tool for predicting upcoming record events with high accuracy while maintaining a concise and interpretable structure.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:34:52 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Barrio-Torres",
"Elsa",
""
],
[
"Abaurrea",
"Jesús",
""
],
[
"Asín",
"Jesús",
""
],
[
"Castillo-Mateo",
"Jorge",
""
],
[
"Cebrián",
"Ana Carmen",
""
],
[
"Gracia",
"Zeus",
""
]
] |
2509.03460
|
Mario Renter\'ia
|
G. N. Darriba, R. Vianden, A. P. Ayala, and M. Renter\'ia
|
Integral $ab$ $initio$/DFT and experimental TDPAC approach enlightening the $aftereffects$ phenomenon: probing electronic properties in $\alpha$-Al$_2$O$_3$:$^{111}$In($\rightarrow$ $^{111}$Cd) at the atomic scale
|
39 pages, 11 figures, 1 Table. The original abstract was shortened for the Arxiv. Submitted to PRB
| null | null | null |
cond-mat.mes-hall cond-mat.mtrl-sci
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
By means of an integral experimental and $ab$ $initio$/DFT approach we contribute here to enlighten and quantify the origin of dynamic hyperfine interactions (HFIs) assigned to the electron-capture (EC) decay aftereffects (ECAE) phenomenon observed in time-differential perturbed $\gamma$-$\gamma$ angular correlation (TDPAC) experiments in oxides doped with ($^{111}$In (EC)$\rightarrow$)$^{111}$Cd as probe-atom. In previous works [Darriba et al., Phys. Rev. B 105, 195201 (2022)] we proposed an $ab$ $initio$ scenario in which the fluctuating electric-field gradients (EFG) producing the dynamic HFI were related with fluctuating electronic environments close to the $^{111}$Cd nucleus, succeeding to identify the environment which produce the final static EFG when the dynamic ($on-off$) process have stopped. In this work we show that in addition it is possible to obtain, for each temperature and HFI observed, the set of initial electronic configurations close to the probe nucleus as well as their related EFGs among which the system fluctuates to generate these dynamic HFIs. For this, we demonstrate analytically and checked experimentally the conditions to stablish the equivalence between the two approaches most used to analyze this type of dynamic HFIs, proposed by B\"averstam et al. and by Lupascu et al.. To unravel the unexpected TDPAC results in $^{111}$In($\rightarrow$ $^{111}$Cd)-implanted $\alpha$-Al$_2$O$_3$ single crystals reported in the literature, we perform a complete $ab$ $initio$/DFT study of Cd-doped $\alpha$-Al$_2$O$_3$ semiconductor and a detailed defect formation energy analysis as a function of the charge state of the Cd impurity. The presence of an unexpected second interaction was a key factor to provide experimental support to identify and quantify the different charge states the $^{111}$Cd atom goes through during its electronic recovery process.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:35:11 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Darriba",
"G. N.",
""
],
[
"Vianden",
"R.",
""
],
[
"Ayala",
"A. P.",
""
],
[
"Rentería",
"M.",
""
]
] |
2509.03461
|
Marcel Ag\"ueros
|
M. A. Ag\"ueros (1,2), J. L. Curtis (1), A. N\'u\~nez (1), C. Burhenne (3), P. Rothstein (1), B. J. Shaham (1), K. Singh (1), P. Bergeron (4), M. Kilic (5), K. R. Covey (6), and S. T. Douglas (7) ((1) Columbia University, (2) Laboratoire d'astrophysique de Bordeaux, (3) Rutgers, (4) Universit\'e de Montr\'eal, (5) University of Oklahoma, (6) Western Washington University, (7) Lafayette College)
|
Crowning the Queen: Membership, Age, Rotation, and Activity for the Open Cluster Coma Berenices
|
34 pages, 24 figures, 5 tables. Accepted by The Astrophysical Journal
| null | null | null |
astro-ph.SR astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
Despite being only 85 pc away, the open cluster Coma Berenices (Coma Ber) has not been extensively studied. This is due in part to its sparseness and low proper motion, which together made Coma Ber's membership challenging to establish. Gaia data for $\approx$400 previously cataloged candidate cluster stars allowed us to identify $\approx$300 as members. With [Fe/H] measurements for nine members, we found that Coma Ber has a solar metallicity, and then fit isochrones to its color--magnitude diagram to determine that it is 675$\pm$100 Myr old. With photometry obtained by the Transiting Exoplanet Survey Satellite (TESS) and Zwicky Transient Facility (ZTF), we measured rotation periods for 137 of Coma Ber's low-mass stars, increasing the sample of members with measured periods by a factor of six, and extending the rotational census for the cluster from its late F stars through to its fully convective M dwarfs. By measuring the equivalent width of the H$\alpha$ line for $\approx$250 stars and collecting X-ray detections for $\approx$100 ($\approx$85% and $\approx$33% of the cluster's members, respectively), we characterized magnetic activity in Coma Ber and examined the dependence of chromospheric and of coronal activity on rotation in these stars. Despite having a metallicity that is 0.2 dex below that of their coeval cousins in Praesepe and the Hyades, low-mass stars in Coma Ber seem to follow a similar rotation--activity relation. In detail, however, there are differences that may provide further insight into the impact of metallicity on this still poorly understood relation.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:37:39 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Agüeros",
"M. A.",
"",
"Columbia University",
"Laboratoire d'astrophysique de Bordeaux"
],
[
"Curtis",
"J. L.",
"",
"Columbia University"
],
[
"Núñez",
"A.",
"",
"Columbia University"
],
[
"Burhenne",
"C.",
"",
"Rutgers"
],
[
"Rothstein",
"P.",
"",
"Columbia University"
],
[
"Shaham",
"B. J.",
"",
"Columbia University"
],
[
"Singh",
"K.",
"",
"Columbia University"
],
[
"Bergeron",
"P.",
"",
"Université de Montréal"
],
[
"Kilic",
"M.",
"",
"University of Oklahoma"
],
[
"Covey",
"K. R.",
"",
"Western Washington University"
],
[
"Douglas",
"S. T.",
"",
"Lafayette College"
]
] |
2509.03462
|
Zhuo Cao
|
Zhuo Cao, Yunxiao Shi, Min Xu
|
sam-llm: interpretable lane change trajectoryprediction via parametric finetuning
|
5 pages
| null | null | null |
cs.AI cs.CV cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This work introduces SAM-LLM, a novel hybrid architecture that bridges the gap between the contextual reasoning of Large Language Models (LLMs) and the physical precision of kinematic lane change models for autonomous driving. The system is designed for interpretable lane change trajectory prediction by finetuning an LLM to output the core physical parameters of a trajectory model instead of raw coordinates. For lane-keeping scenarios, the model predicts discrete coordinates, but for lane change maneuvers, it generates the parameters for an enhanced Sinusoidal Acceleration Model (SAM), including lateral displacement, maneuver duration, initial lateral velocity, and longitudinal velocity change. This parametric approach yields a complete, continuous, and physically plausible trajectory model that is inherently interpretable and computationally efficient, achieving an 80% reduction in output size compared to coordinate-based methods. The SAM-LLM achieves a state-of-the-art overall intention prediction accuracy of 98.73%, demonstrating performance equivalent to traditional LLM predictors while offering significant advantages in explainability and resource efficiency.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:37:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Cao",
"Zhuo",
""
],
[
"Shi",
"Yunxiao",
""
],
[
"Xu",
"Min",
""
]
] |
2509.03463
|
Parham Khamsepour
|
Parham Khamsepour, Mark Cole, Ish Ashraf, Sandeep Puri, Mehrdad Sabetzadeh and Shiva Nejati
|
The Impact of Critique on LLM-Based Model Generation from Natural Language: The Case of Activity Diagrams
| null | null | null | null |
cs.SE
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Large Language Models (LLMs) show strong potential for automating the generation of models from natural-language descriptions. A common approach is an iterative generate-critique-refine loop, where candidate models are produced, evaluated, and updated based on detected issues. This process needs to address: (1) structural correctness - compliance with well-formedness rules - and (2) semantic alignment - accurate reflection of the intended meaning in the source text. We present LADEX (LLM-based Activity Diagram Extractor), a pipeline for deriving activity diagrams from natural-language process descriptions using an LLM-driven critique-refine process. Structural checks in LADEX can be performed either algorithmically or by an LLM, while alignment checks are always performed by an LLM. We design five ablated variants of LADEX to study: (i) the impact of the critique-refine loop itself, (ii) the role of LLM-based semantic checks, and (iii) the comparative effectiveness of algorithmic versus LLM-based structural checks.
To evaluate LADEX, we compare the generated activity diagrams with expert-created ground truths using trace-based operational semantics. This enables automated measurement of correctness and completeness. Experiments on two datasets indicate that: (1) the critique-refine loop improves structural validity, correctness, and completeness compared to single-pass generation; (2) algorithmic structural checks eliminate inconsistencies that LLM-based checks fail to detect, improving correctness by an average of 17.81% and completeness by 13.24% over LLM-only checks; and (3) combining algorithmic structural checks with LLM-based semantic checks, implemented using the reasoning-focused O4 Mini, achieves the best overall performance - yielding average correctness of up to 86.37% and average completeness of up to 88.56% - while requiring fewer than five LLM calls on average.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:39:25 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Khamsepour",
"Parham",
""
],
[
"Cole",
"Mark",
""
],
[
"Ashraf",
"Ish",
""
],
[
"Puri",
"Sandeep",
""
],
[
"Sabetzadeh",
"Mehrdad",
""
],
[
"Nejati",
"Shiva",
""
]
] |
2509.03464
|
McKenzie Fontenot
|
Kenzie Fontenot, Iris Nguyen, Cody Olsen
|
A Classification of Winning Sets of Cops in $\mathbb{Z}^n$
|
15 pages, 7 figures
| null | null | null |
math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
The game of Cops and Robbers is a pursuit-evasion game on graphs that has been extensively studied in finite settings, particularly through the concept of cop number. In this paper, we explore infinite variants of the game, focusing on the lattice graph $\mathbb{Z}^n$. Since the cop number of $\mathbb{Z}^n$ is infinite, we shift attention to the notion of \emph{cop density}, examining how sparsely cops may be placed while still guaranteeing capture. We introduce the framework of \emph{coordinate matching} as a central strategy and prove the existence of density-zero configurations of cops that ensure eventual capture of the robber. Building on this, we provide a complete classification of winning cop sets in $\mathbb{Z}^n$, showing necessary and sufficient conditions for capture based on infinite distributions of cops across coordinate directions. We conclude with directions for further study, including variations of the game where the robber imposes spatial restrictions on cop placement.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:39:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fontenot",
"Kenzie",
""
],
[
"Nguyen",
"Iris",
""
],
[
"Olsen",
"Cody",
""
]
] |
2509.03465
|
Kuan-Chuan Peng
|
Kuan-Chuan Peng
|
Joint Training of Image Generator and Detector for Road Defect Detection
|
This paper is accepted to ICCV 2025 Workshop on Representation Learning with Very Limited Resources: When Data, Modalities, Labels, and Computing Resources are Scarce as an oral paper
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Road defect detection is important for road authorities to reduce the vehicle damage caused by road defects. Considering the practical scenarios where the defect detectors are typically deployed on edge devices with limited memory and computational resource, we aim at performing road defect detection without using ensemble-based methods or test-time augmentation (TTA). To this end, we propose to Jointly Train the image Generator and Detector for road defect detection (dubbed as JTGD). We design the dual discriminators for the generative model to enforce both the synthesized defect patches and overall images to look plausible. The synthesized image quality is improved by our proposed CLIP-based Fr\'echet Inception Distance loss. The generative model in JTGD is trained jointly with the detector to encourage the generative model to synthesize harder examples for the detector. Since harder synthesized images of better quality caused by the aforesaid design are used in the data augmentation, JTGD outperforms the state-of-the-art method in the RDD2022 road defect detection benchmark across various countries under the condition of no ensemble and TTA. JTGD only uses less than 20% of the number of parameters compared with the competing baseline, which makes it more suitable for deployment on edge devices in practice.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:41:50 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Peng",
"Kuan-Chuan",
""
]
] |
2509.03466
|
Andrew Sargent
|
Andrew J. Sargent, Alexander J. van der Horst, Megan C. Johnson, Travis C. Fischer, Nathan J. Secrest, Onic I. Shuvo, Macon A. Magno, and Luis C. Fernandez
|
VLA FRAMEx. III. Circumnuclear Radio Emission Mechanisms in Hard X-ray Selected Active Galactic Nuclei
|
19 pages, 5 figures, 4 tables. Submitted to ApJ, comments welcome!
| null | null | null |
astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
We present Stokes I continuum analysis for a volume-limited sample ($<40~\rm{Mpc}$) of hard X-ray selected active galactic nuclei (AGNs) using $4-12$ GHz observations with the Karl G. Jansky Very Large Array (VLA). All of the 25 sources analyzed here have previously been observed with the Very Long Baseline Array (VLBA) to probe their subparsec projected physical scales, but detected emission has only been measured for 12 of the sources at $C$ band (4.4 GHz), despite expectations. We determined that coronal emission is unlikely to be a dominant emission mechanism for the sources not detected by the VLBA, and the emission measured with the VLA is likely produced beyond parsec spatial scales. We also explore potential radiation mechanisms for the circumnuclear radio emission that is produced beyond the observable ~parsec physical scales probed with the VLBA but within the $\leq30-110$ parsec spatial scales observed with the VLA. From an energetics perspective, we find that all targets have extranuclear radio emission that is compatible with AGN winds, assuming a maximum of 10% of the bolometric output can supply the mechanical energy observed. We also find that the excess emission is likely too strong for star formation alone when compared to results from optical spectroscopy, but may contribute in smaller capacities.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:43:36 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sargent",
"Andrew J.",
""
],
[
"van der Horst",
"Alexander J.",
""
],
[
"Johnson",
"Megan C.",
""
],
[
"Fischer",
"Travis C.",
""
],
[
"Secrest",
"Nathan J.",
""
],
[
"Shuvo",
"Onic I.",
""
],
[
"Magno",
"Macon A.",
""
],
[
"Fernandez",
"Luis C.",
""
]
] |
2509.03467
|
Soukeina Elhassen
|
Soukeina Elhassen, Lama Al Khuzayem, Areej Alhothali, Ohoud Alzamzami, Nahed Alowaidi
|
Continuous Saudi Sign Language Recognition: A Vision Transformer Approach
|
23 pages, 13 figures, 5 tables
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/publicdomain/zero/1.0/
|
Sign language (SL) is an essential communication form for hearing-impaired and deaf people, enabling engagement within the broader society. Despite its significance, limited public awareness of SL often leads to inequitable access to educational and professional opportunities, thereby contributing to social exclusion, particularly in Saudi Arabia, where over 84,000 individuals depend on Saudi Sign Language (SSL) as their primary form of communication. Although certain technological approaches have helped to improve communication for individuals with hearing impairments, there continues to be an urgent requirement for more precise and dependable translation techniques, especially for Arabic sign language variants like SSL. Most state-of-the-art solutions have primarily focused on non-Arabic sign languages, resulting in a considerable absence of resources dedicated to Arabic sign language, specifically SSL. The complexity of the Arabic language and the prevalence of isolated sign language datasets that concentrate on individual words instead of continuous speech contribute to this issue. To address this gap, our research represents an important step in developing SSL resources. To address this, we introduce the first continuous Saudi Sign Language dataset called KAU-CSSL, focusing on complete sentences to facilitate further research and enable sophisticated recognition systems for SSL recognition and translation. Additionally, we propose a transformer-based model, utilizing a pretrained ResNet-18 for spatial feature extraction and a Transformer Encoder with Bidirectional LSTM for temporal dependencies, achieving 99.02\% accuracy at signer dependent mode and 77.71\% accuracy at signer independent mode. This development leads the way to not only improving communication tools for the SSL community but also making a substantial contribution to the wider field of sign language.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:44:08 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Elhassen",
"Soukeina",
""
],
[
"Khuzayem",
"Lama Al",
""
],
[
"Alhothali",
"Areej",
""
],
[
"Alzamzami",
"Ohoud",
""
],
[
"Alowaidi",
"Nahed",
""
]
] |
2509.03468
|
Hidde Sch\"onberger
|
Carolin Kreisbeck, Hidde Sch\"onberger
|
Local boundary conditions in nonlocal hyperelasticity via heterogeneous horizons
|
35 pages, 1 figure
| null | null | null |
math.AP math.FA
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
In this paper, we consider a class of variational problems with integral functionals involving nonlocal gradients. These models have been recently proposed as refinements of classical hyperelasticity, aiming for an effective framework to capture also discontinuous and singular material effects. Specific to our set-up is a space-dependent interaction range that vanishes at the boundary of the reference domain. This ensures that the nonlocal operator depends only on values within the domain and localizes to the classical gradient at the boundary, which allows for a seamless integration of nonlocal modeling with local boundary values. The main contribution of this work is a comprehensive theory for the newly introduced associated Sobolev spaces, including the rigorous treatment of a trace operator and Poincar\'e inequalities. A central aspect of our technical approach lies in exploiting connections with pseudo-differential operator theory. As an application, we establish the existence of minimizers for functionals with quasiconvex or polyconvex integrands depending on heterogeneous nonlocal gradients, subject to local Dirichlet, Neumann or mixed-type boundary conditions.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:45:39 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Kreisbeck",
"Carolin",
""
],
[
"Schönberger",
"Hidde",
""
]
] |
2509.03469
|
Mintu Mondal Dr
|
Deep Singha Roy, Sk Kalimuddin, Subrata Pachhal, Saikat Mondal, Soham Das, Sukanya Jana, Arnab Bera, Satyabrata Bera, Tuhin Debnath, Ankan Bag, Souvik Pramanik, Sudipta Chatterjee, Sanjib Naskar, Shishir Kumar Pandey, Adhip Agarwala, Mintu Mondal
|
Signatures of emergent surface states across a displacive topological phase transition in Bi$_4$I$_4$
|
35 pages, 20 figures, research articles
| null | null | null |
cond-mat.str-el cond-mat.mes-hall cond-mat.mtrl-sci
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Topological phase transitions involving crystalline symmetry breaking provide a fertile ground to explore the interplay between symmetry, topology, and emergent quantum phenomena. Recently discovered quasi-one-dimensional topological material, Bi$_4$I$_4$, has been predicted to host topologically non-trivial gapless surfaces at high temperature, which undergo a finite temperature phase transition to a low temperature gapped phase. Here we present experimental signatures of this room temperature phase transition from a high-temperature $\beta$-phase with a surface state to a gapped $\alpha$-phase hosting hinge states. Using real-space current mapping and resistance fluctuation spectroscopy, we identify signatures of a displacive topological phase transition mediated by a first-order thermodynamic structural change. Near the emergence of $\beta$-phase, we observe pronounced telegraphic noise, indicating fluctuating phase domains with topological surface states. The spatially resolved current map reveals electron transport via the gapless surface states in the $\beta$-phase, which vanishes upon transitioning to the $\alpha$-phase with localized conduction channels (or hinge modes). Our experimental results, supported by first principles estimates and effective theory of a topological displacive phase transition, establish Bi$_4$I$_4$ as a candidate material showing intricate interplay of classical thermodynamic phase transitions with topological quantum phenomena.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:46:35 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Roy",
"Deep Singha",
""
],
[
"Kalimuddin",
"Sk",
""
],
[
"Pachhal",
"Subrata",
""
],
[
"Mondal",
"Saikat",
""
],
[
"Das",
"Soham",
""
],
[
"Jana",
"Sukanya",
""
],
[
"Bera",
"Arnab",
""
],
[
"Bera",
"Satyabrata",
""
],
[
"Debnath",
"Tuhin",
""
],
[
"Bag",
"Ankan",
""
],
[
"Pramanik",
"Souvik",
""
],
[
"Chatterjee",
"Sudipta",
""
],
[
"Naskar",
"Sanjib",
""
],
[
"Pandey",
"Shishir Kumar",
""
],
[
"Agarwala",
"Adhip",
""
],
[
"Mondal",
"Mintu",
""
]
] |
2509.03470
|
Hasan El Moumni
|
F. Barzi, H. El Moumni, K. Masmar
|
Modified Unruh Thermodynamics in Emergent Gravity: Finite Heat Capacity and R\'enyi Entropy
|
2 pages and 2 figures
| null | null | null |
hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We show that Jacobson's thermodynamic derivation of Einstein's equations remains valid when local Rindler horizons are treated as finite heat-capacity systems, resolving the unphysical infinite-bath assumption of standard Unruh thermodynamics. The resulting entropy takes the form of R\'enyi entropy with nonextensivity parameter $\lambda\sim C^{-1}$, or equivalently, a new "Einstein entropy" that exactly preserves the Einstein equations for all heat capacities. In both cases, the Unruh temperature is modified as \begin{equation*}
T_\text{mod}=\frac{\hbar\kappa}{2\pi}\left(1+\frac{S}{C}\right), \end{equation*} establishing a universal link between finite-capacity thermodynamics and nonextensive entropy. We further obtain a corrected scalar Einstein equation with an upper bound on horizon energy flux, pointing to testable signatures in heavy-ion collisions, accelerator spin polarization, and analog gravity experiments. These results reinforce the robustness of the emergent-gravity paradigm and connect spacetime dynamics to generalized entropies of quantum information theory.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:50:25 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Barzi",
"F.",
""
],
[
"Moumni",
"H. El",
""
],
[
"Masmar",
"K.",
""
]
] |
2509.03471
|
Hui Yu
|
Hui Yu, Zhaoyang Wang, Ping Lin
|
Linear Relaxation Schemes with Asymptotically Compatible Energy Law for Time-Fractional Phase-Field Models
| null | null | null | null |
math.NA cs.NA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we propose a variable time-step linear relaxation scheme for time-fractional phase-field equations with a free energy density in general polynomial form. The $L1^{+}$-CN formula is used to discretize the fractional derivative, and an auxiliary variable is introduced to approximate the nonlinear term by directly solving algebraic equations rather than differential-algebraic equations as in the invariant energy quadratization (IEQ) and the scalar auxiliary variable (SAV) approaches. The developed semi-discrete scheme is second-order accurate in time, and the inconsistency between the auxiliary and the original variables does not deteriorate over time. Furthermore, we take the time-fractional volume-conserved Allen-Cahn equation, the time-fractional Cahn-Hilliard equation, and the time-fractional Swift-Hohenberg equation as examples to demonstrate that the constructed schemes are energy stable and that the discrete energy dissipation law is asymptotically compatible with the classical one when the fractional-order parameter $\alpha\rightarrow 1^{-}$. Several numerical examples demonstrate the effectiveness of the proposed scheme. In particular, numerical results confirm that the auxiliary variable remains well aligned with the original variable, and the error between them does not continue to increase over time before the system reaches steady state.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:51:07 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Yu",
"Hui",
""
],
[
"Wang",
"Zhaoyang",
""
],
[
"Lin",
"Ping",
""
]
] |
2509.03472
|
Yubo Gao
|
Yubo Gao, Renbo Tu, Gennady Pekhimenko, Nandita Vijaykumar
|
DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling
| null | null | null | null |
cs.LG cs.AI cs.DC
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Differentially-Private SGD (DP-SGD) is a powerful technique to protect user privacy when using sensitive data to train neural networks. During training, converting model weights and activations into low-precision formats, i.e., quantization, can drastically reduce training times, energy consumption, and cost, and is thus a widely used technique. In this work, we demonstrate that quantization causes significantly higher accuracy degradation in DP-SGD compared to regular SGD. We observe that this is caused by noise injection in DP-SGD, which amplifies quantization variance, leading to disproportionately large accuracy degradation. To address this challenge, we present QPQuant, a dynamic quantization framework that adaptively selects a changing subset of layers to quantize at each epoch. Our method combines two key ideas that effectively reduce quantization variance: (i) probabilistic sampling of the layers that rotates which layers are quantized every epoch, and (ii) loss-aware layer prioritization, which uses a differentially private loss sensitivity estimator to identify layers that can be quantized with minimal impact on model quality. This estimator consumes a negligible fraction of the overall privacy budget, preserving DP guarantees. Empirical evaluations on ResNet18, ResNet50, and DenseNet121 across a range of datasets demonstrate that DPQuant consistently outperforms static quantization baselines, achieving near Pareto-optimal accuracy-compute trade-offs and up to 2.21x theoretical throughput improvements on low-precision hardware, with less than 2% drop in validation accuracy.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:51:26 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Gao",
"Yubo",
""
],
[
"Tu",
"Renbo",
""
],
[
"Pekhimenko",
"Gennady",
""
],
[
"Vijaykumar",
"Nandita",
""
]
] |
2509.03473
|
Mitch Walker
|
Mitchell J. Walker, Ryuji Moriya, Jack D. Segal, Liam A. P. Gallagher, Matthew Hill, Fr\'ed\'eric Leroux, Zhongxiao Xu, Matthew P. A. Jones
|
Painted loading: a toolkit for loading spatially large optical tweezer arrays
|
17 pages, 6 figures
| null | null | null |
physics.atom-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Arrays of neutral atoms in optical tweezers are widely used in quantum simulation and computation, and precision frequency metrology. The capabilities of these arrays are enhanced by maximising the number of available sites. Here we increase the spatial extent of a two-dimensional array of strontium-88 atoms by sweeping the frequency of the cooling light to move the atomic reservoir across the array. We load arrays with vertical heights of >100 {\mu}m, exceeding the height of an array loaded from a static reservoir by a factor of >3. We investigate the site-to-site atom number distribution, tweezer lifetime, and temperature, achieving an average temperature across the array of 1.49(3) {\mu}K. By controlling the frequency sweep we show it is possible to control the distribution of atoms across the array, including uniform and non-uniformly loaded arrays, and arrays with selectively loaded regions. We explain our results using a rate equation model which is in good qualitative agreement with the data.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:53:33 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Walker",
"Mitchell J.",
""
],
[
"Moriya",
"Ryuji",
""
],
[
"Segal",
"Jack D.",
""
],
[
"Gallagher",
"Liam A. P.",
""
],
[
"Hill",
"Matthew",
""
],
[
"Leroux",
"Frédéric",
""
],
[
"Xu",
"Zhongxiao",
""
],
[
"Jones",
"Matthew P. A.",
""
]
] |
2509.03474
|
Erkan Bayram
|
Erkan Bayram, Mohamed-Ali Belabbas, Tamer Ba\c{s}ar
|
Geometric Foundations of Tuning without Forgetting in Neural ODEs
| null | null | null | null |
cs.LG math.OC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In our earlier work, we introduced the principle of Tuning without Forgetting (TwF) for sequential training of neural ODEs, where training samples are added iteratively and parameters are updated within the subspace of control functions that preserves the end-point mapping at previously learned samples on the manifold of output labels in the first-order approximation sense. In this letter, we prove that this parameter subspace forms a Banach submanifold of finite codimension under nonsingular controls, and we characterize its tangent space. This reveals that TwF corresponds to a continuation/deformation of the control function along the tangent space of this Banach submanifold, providing a theoretical foundation for its mapping-preserving (not forgetting) during the sequential training exactly, beyond first-order approximation.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:54:37 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Bayram",
"Erkan",
""
],
[
"Belabbas",
"Mohamed-Ali",
""
],
[
"Başar",
"Tamer",
""
]
] |
2509.03475
|
Hong Ye Tan
|
Hong Ye Tan, Subhadip Mukherjee, Junqi Tang
|
From Image Denoisers to Regularizing Imaging Inverse Problems: An Overview
| null | null | null | null |
math.OC cs.LG eess.IV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Inverse problems lie at the heart of modern imaging science, with broad applications in areas such as medical imaging, remote sensing, and microscopy. Recent years have witnessed a paradigm shift in solving imaging inverse problems, where data-driven regularizers are used increasingly, leading to remarkably high-fidelity reconstruction. A particularly notable approach for data-driven regularization is to use learned image denoisers as implicit priors in iterative image reconstruction algorithms. This survey presents a comprehensive overview of this powerful and emerging class of algorithms, commonly referred to as plug-and-play (PnP) methods. We begin by providing a brief background on image denoising and inverse problems, followed by a short review of traditional regularization strategies. We then explore how proximal splitting algorithms, such as the alternating direction method of multipliers (ADMM) and proximal gradient descent (PGD), can naturally accommodate learned denoisers in place of proximal operators, and under what conditions such replacements preserve convergence. The role of Tweedie's formula in connecting optimal Gaussian denoisers and score estimation is discussed, which lays the foundation for regularization-by-denoising (RED) and more recent diffusion-based posterior sampling methods. We discuss theoretical advances regarding the convergence of PnP algorithms, both within the RED and proximal settings, emphasizing the structural assumptions that the denoiser must satisfy for convergence, such as non-expansiveness, Lipschitz continuity, and local homogeneity. We also address practical considerations in algorithm design, including choices of denoiser architecture and acceleration strategies.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:54:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tan",
"Hong Ye",
""
],
[
"Mukherjee",
"Subhadip",
""
],
[
"Tang",
"Junqi",
""
]
] |
2509.03476
|
Hiroyasu Ando
|
Hiroyasu Ando, A. James O'Malley, Akihiro Nishi
|
Temporal Exposure Dependence Bias in Vaccine Efficacy Trials
|
16 pages, 7 figures
| null | null | null |
stat.ME
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Using time-to-event methods such as Cox proportional hazards models, it is well established that unmeasured heterogeneity in exposure or infection risk can lead to downward bias in point estimates of the per-contact vaccine efficacy (VE) in infectious disease trials. In this study, we explore an unreported source of bias-arising from temporally correlated exposure status-that is typically unmeasured and overlooked in standard analyses. Although this form of bias can plausibly affect a wide range of VE trials, it has received limited empirical attention. We develop a mathematical framework to characterize the mechanism of this bias and derive a closed-form approximation to quantify its magnitude without requiring direct measurement of exposure. Our findings show that, under realistic parameter settings, the resulting bias can be substantial. These results suggest that temporally correlated exposure should be recognized as a potentially important factor in the design and analysis of infectious disease vaccine trials.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:55:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ando",
"Hiroyasu",
""
],
[
"O'Malley",
"A. James",
""
],
[
"Nishi",
"Akihiro",
""
]
] |
2509.03477
|
Abhi Kamboj
|
Duy A. Nguyen, Abhi Kamboj, Minh N. Do
|
Robult: Leveraging Redundancy and Modality Specific Features for Robust Multimodal Learning
|
Accepted and presented at IJCAI 2025 in Montreal, Canada
| null | null | null |
cs.LG cs.AI cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Addressing missing modalities and limited labeled data is crucial for advancing robust multimodal learning. We propose Robult, a scalable framework designed to mitigate these challenges by preserving modality-specific information and leveraging redundancy through a novel information-theoretic approach. Robult optimizes two core objectives: (1) a soft Positive-Unlabeled (PU) contrastive loss that maximizes task-relevant feature alignment while effectively utilizing limited labeled data in semi-supervised settings, and (2) a latent reconstruction loss that ensures unique modality-specific information is retained. These strategies, embedded within a modular design, enhance performance across various downstream tasks and ensure resilience to incomplete modalities during inference. Experimental results across diverse datasets validate that Robult achieves superior performance over existing approaches in both semi-supervised learning and missing modality contexts. Furthermore, its lightweight design promotes scalability and seamless integration with existing architectures, making it suitable for real-world multimodal applications.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:56:27 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Nguyen",
"Duy A.",
""
],
[
"Kamboj",
"Abhi",
""
],
[
"Do",
"Minh N.",
""
]
] |
2509.03478
|
Ahmed Ellithy
|
Ahmed Ellithy
|
Local Well-Posedness for the Bartnik Stationary Extension Problem near Schwarzschild Spheres
|
44 pages, 2 figures
| null | null | null |
gr-qc math.AP math.DG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We investigate the Bartnik stationary extension conjecture, which arises from the definition of the spacetime Bartnik mass for a compact region in a general initial data set satisfying the dominant energy condition. This conjecture posits the existence and uniqueness (up to isometry) of an asymptotically flat stationary vacuum spacetime containing an initial data set $(M, \mathfrak{g}, \Pi)$ that realizes prescribed Bartnik boundary data on $\partial M$, consisting of the induced metric, mean curvature, and appropriate components of the spacetime extrinsic curvature $\Pi$.
Building on the analytic framework developed in arXiv:2411.02801 for the static case, we show that, in a double geodesic gauge, the stationary vacuum Einstein equations reduce to a coupled system comprising elliptic and transport-type equations, with the genuinely stationary contributions encoded in an additional boundary value problem for a $1$-form $\theta$. Our approach employs both a geodesic gauge for the quotient metric $g$ in the quotient formalism and a $\theta$-geodesic gauge for the $1$-form $\theta$.
We establish local well-posedness for the Bartnik stationary metric extension problem for Bartnik data sufficiently close to that of any coordinate sphere in a Schwarzschild ${t=0}$ slice, including data with arbitrarily small mean curvature. A key feature of our framework is that the linearized equations decouple: the equations for the metric and potential reduce to the previously solved static case, while the boundary value problem for $\theta$ is treated independently. We prove solvability of this boundary value problem in the Bochner-measurable function spaces adapted to the coupled system developed in arXiv:2411.02801, establishing uniform estimates for the vector spherical harmonic decomposition of $\theta$.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 16:58:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ellithy",
"Ahmed",
""
]
] |
2509.03479
|
Haonan Wang
|
Haonan Wang and Mingjia Zhao and Junfeng Sun and Wei Liu
|
Design and Optimization of Reinforcement Learning-Based Agents in Text-Based Games
|
6 papges
|
Copyright (c) 2025 International Journal of Computer Science and Information Technology International Journal of Computer Science and Information Technology International Journal of Computer Science and Information Technology
|
10.62051/ijcsit.v5n2.02
| null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
As AI technology advances, research in playing text-based games with agents has becomeprogressively popular. In this paper, a novel approach to agent design and agent learning ispresented with the context of reinforcement learning. A model of deep learning is first applied toprocess game text and build a world model. Next, the agent is learned through a policy gradient-based deep reinforcement learning method to facilitate conversion from state value to optimal policy.The enhanced agent works better in several text-based game experiments and significantlysurpasses previous agents on game completion ratio and win rate. Our study introduces novelunderstanding and empirical ground for using reinforcement learning for text games and sets thestage for developing and optimizing reinforcement learning agents for more general domains andproblems.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:01:38 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Wang",
"Haonan",
""
],
[
"Zhao",
"Mingjia",
""
],
[
"Sun",
"Junfeng",
""
],
[
"Liu",
"Wei",
""
]
] |
2509.03480
|
Shaopeng Tang
|
Shao-Peng Tang and Hai-Tian Wang and Yin-Jie Li and Yi-Zhong Fan
|
Verification of the Black Hole Area Law with GW230814
|
6 pages, 5 figures
| null | null | null |
gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present observational confirmation of Hawking's black-hole area theorem using the newly released gravitational-wave data from the GWTC-4 catalog. We have analyzed two high signal-to-noise ratio binary black hole (BBH) merger events, including GW230814 and GW231226, from the first part of the LIGO-Virgo-KAGRA O4 run, and measured the (total) horizon area of the black holes before and after merger. For both events, the horizon area of the remnant black hole is found to be greater than the total horizon area of the two pre-merger black holes at a high possibility (effectively $\gtrsim 99\%$), and particularly with GW230814 we have a significance level of $\gtrsim 5\sigma$ for the {\it first} time. These results provide the direct convincing verification of the black-hole area law, further bolstering the validity of classical general relativity in the dynamical, strong-field regime.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:03:18 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tang",
"Shao-Peng",
""
],
[
"Wang",
"Hai-Tian",
""
],
[
"Li",
"Yin-Jie",
""
],
[
"Fan",
"Yi-Zhong",
""
]
] |
2509.03481
|
Julian Trouillon
|
Lorenzo Talamanca and Julian Trouillon
|
PoolPy: Flexible Group Testing Design for Large-Scale Screening
| null | null | null | null |
cs.IT math.IT
|
http://creativecommons.org/licenses/by/4.0/
|
In large screening campaigns, group testing can greatly reduce the number of tests needed when compared to testing each sample individually. However, choosing and applying an appropriate group testing method remains challenging due to the wide variety in design and performance across methods, and the lack of accessible tools. Here, we present PoolPy, a unified framework for designing and selecting optimal group testing strategies across ten different methods according to user-defined constraints, such as time, cost or sample dilution. By computing over 10,000 group testing designs made available through a web interface, we identified key trade-offs, such as minimizing test number or group size, that define applicability to specific use cases. Overall, we show that no single method is universally optimal, and provide clear indications for method choice on a case-by-case basis.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:03:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Talamanca",
"Lorenzo",
""
],
[
"Trouillon",
"Julian",
""
]
] |
2509.03482
|
Michael Horodynski
|
Jamison Sloan, Michael Horodynski, Shiekh Zia Uddin, Yannick Salamin, Michael Birk, Pavel Sidorenko, Ido Kaminer, Marin Solja\v{c}i\'c, Nicholas Rivera
|
Programmable control of the spatiotemporal quantum noise of light
|
The first three listed authors have equal contributions
| null | null | null |
physics.optics
|
http://creativecommons.org/licenses/by/4.0/
|
Optoelectronic systems based on multiple modes of light can often exceed the performance of their single-mode counterparts. However, multimode nonlinear interactions often introduce considerable amounts of noise, limiting the ultimate performance of these systems. It is therefore crucial to develop ways to simultaneously control complex nonlinear interactions while also gaining control over their noise. Here, we show that noise buildup in nonlinear multimode systems can be strongly suppressed by controlling the input wavefront. We demonstrate this approach in a multimode fiber by using an active wavefront-shaping protocol to focus a region of high intensity - yet low intensity noise - at the output. Our programmable control of both the input and output reduces the beam noise by 12 dB beyond what linear attenuation achieves, reaching levels near the quantum shot-noise limit. We show that this is possible because the optimally shaped wavefront maximally decouples the output intensity fluctuations from the input laser fluctuations. These findings are supported by a new theoretical and simulation framework that efficiently captures spatiotemporal quantum noise dynamics in highly multimode nonlinear systems. Our results highlight the potential of programmable wavefront shaping to enable nonlinear multimode technologies that overcome noise buildup to operate at quantum-noise limits.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:06:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sloan",
"Jamison",
""
],
[
"Horodynski",
"Michael",
""
],
[
"Uddin",
"Shiekh Zia",
""
],
[
"Salamin",
"Yannick",
""
],
[
"Birk",
"Michael",
""
],
[
"Sidorenko",
"Pavel",
""
],
[
"Kaminer",
"Ido",
""
],
[
"Soljačić",
"Marin",
""
],
[
"Rivera",
"Nicholas",
""
]
] |
2509.03483
|
Viktor Panshin
|
Viktor Panshin
|
On recognition of simple groups with disconnected prime graph by spectrum
| null | null | null | null |
math.GR
|
http://creativecommons.org/licenses/by/4.0/
|
Two finite groups are said to be isospectral if they have the same sets of element orders. This paper completes the description of finite groups that are isospectral to finite simple groups with disconnected prime graph.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:07:12 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Panshin",
"Viktor",
""
]
] |
2509.03484
|
Omar Alejandro Garcia Alcantara
|
Ignacio Rubio Scola, Omar Alejandro Garcia Alcantara, Steven Sandoval, Eduardo Steed Espinoza Quesada, Hernan Haimovich, Luis Rodolfo Garcia Carrillo
|
Globally Asymptotically Stable Trajectory Tracking of Underactuated UAVs using Geometric Algebra
|
This work has been submitted to the IEEE TAES for possible publication
| null | null | null |
eess.SY cs.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper employs Geometric Algebra (GA) tools to model the dynamics of objects in 3-dimensional space, serving as a proof of concept to facilitate control design for trajectory tracking in underactuated systems. For control purposes, the model is structured as a cascade system, where a rotational subsystem drives a translational one. The rotational subsystem is linear, while the translational subsystem follows a linear-plus-perturbation form, thereby reducing the complexity of control design. A control strategy requiring only simple operations, no memory, and no iterative search loops is presented to illustrate the main features of the GA model. By employing GA to model both translations and rotations, a singularity-free and geometrically intuitive representation can be achieved through the use of the geometric product. Closed-loop stability is rigorously established using input-to-state stability methods. Numerical simulations of a quad tilt-rotorcraft performing trajectory tracking in a windy environment validate the controller's stability and performance.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:09:26 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Scola",
"Ignacio Rubio",
""
],
[
"Alcantara",
"Omar Alejandro Garcia",
""
],
[
"Sandoval",
"Steven",
""
],
[
"Quesada",
"Eduardo Steed Espinoza",
""
],
[
"Haimovich",
"Hernan",
""
],
[
"Carrillo",
"Luis Rodolfo Garcia",
""
]
] |
2509.03485
|
Michael Ortiz
|
Michael Ortiz
|
Linear viscoelasticity: Mechanics, analysis and approximation
| null | null | null | null |
math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The aim of this review is to highlight the connection between well-established physical and mathematical principles as they pertain to the theory of linear elasticity. We begin by examining the physical foundations of Boltzmann and Volterra's hereditary law formalism, and how those principles restrict the form of the hereditary law. We then turn to questions of material stability and continuous dependence on the stress history within the framework of the Lax-Milgram theorem, which we find to set forth rigorous and unequivocal conditions for the well-posedness of the linear viscoelastic problem. The outcome of this analysis is remarkable in that it gives precise meaning to fundamental physical properties such as fading memory. Finally, we turn to the question of best representation of viscoelastic materials by finite-rank hereditary operators or, equivalently, by a finite set of history or internal variables. We note that the theory of Hilbert-Schmidt operators and $N$-widths supplies the answer to the question.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:12:13 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ortiz",
"Michael",
""
]
] |
2509.03486
|
Claudio Toni
|
Carlo Broggini, Giuseppe Di Carlo, Luca Di Luzio, Denise Piatti, Claudio Toni
|
Probing a Light Scalar Boson with a few-MeV Proton Beam Deep Underground
|
17 pages, 3 figures
| null | null | null |
hep-ph hep-ex nucl-ex
|
http://creativecommons.org/licenses/by/4.0/
|
We propose to investigate the production of a light scalar boson $\phi$ in low-energy proton-nucleus interactions using the 3.5 MV accelerator of the Bellotti Ion Beam Facility, located in the underground Gran Sasso National Laboratory. Nuclear reactions induced by a few-MeV proton beam on suitable target materials can act as a controlled source of $\phi$ particles. Owing to the deep-underground location, the facility benefits from substantial cosmic-ray shielding, enabling searches for rare processes with minimal background. The produced $\phi$ particles will be sought with large-volume, low-background detectors already operating or currently under construction at the Gran Sasso National Laboratory. This approach combines a tunable accelerator-based production mechanism with the exceptional sensitivity of underground rare-event searches, offering a novel avenue to probe light scalar bosons beyond the Standard Model.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:13:13 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Broggini",
"Carlo",
""
],
[
"Di Carlo",
"Giuseppe",
""
],
[
"Di Luzio",
"Luca",
""
],
[
"Piatti",
"Denise",
""
],
[
"Toni",
"Claudio",
""
]
] |
2509.03487
|
Jigang Fan
|
Jigang Fan, Zhenghong Zhou, Ruofan Jin, Le Cong, Mengdi Wang, Zaixi Zhang
|
SafeProtein: Red-Teaming Framework and Benchmark for Protein Foundation Models
| null | null | null | null |
cs.LG cs.AI cs.CR q-bio.BM q-bio.QM
|
http://creativecommons.org/licenses/by/4.0/
|
Proteins play crucial roles in almost all biological processes. The advancement of deep learning has greatly accelerated the development of protein foundation models, leading to significant successes in protein understanding and design. However, the lack of systematic red-teaming for these models has raised serious concerns about their potential misuse, such as generating proteins with biological safety risks. This paper introduces SafeProtein, the first red-teaming framework designed for protein foundation models to the best of our knowledge. SafeProtein combines multimodal prompt engineering and heuristic beam search to systematically design red-teaming methods and conduct tests on protein foundation models. We also curated SafeProtein-Bench, which includes a manually constructed red-teaming benchmark dataset and a comprehensive evaluation protocol. SafeProtein achieved continuous jailbreaks on state-of-the-art protein foundation models (up to 70% attack success rate for ESM3), revealing potential biological safety risks in current protein foundation models and providing insights for the development of robust security protection technologies for frontier models. The codes will be made publicly available at https://github.com/jigang-fan/SafeProtein.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:13:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fan",
"Jigang",
""
],
[
"Zhou",
"Zhenghong",
""
],
[
"Jin",
"Ruofan",
""
],
[
"Cong",
"Le",
""
],
[
"Wang",
"Mengdi",
""
],
[
"Zhang",
"Zaixi",
""
]
] |
2509.03488
|
Miguel Rivas-Costa
|
Miguel Rivas-Costa and Carlos Mosquera
|
Efficient DoA Estimation with Hybrid Linear and Rectangular Arrays Using Compact DFT Codebook
| null | null | null | null |
eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Hybrid Analog and Digital (HAD) architectures provide a cost-effective alternative for large-scale antenna arrays, but accurate Direction-of-Arrival (DoA) estimation remains challenging due to limited digital dimensionality and constrained beamforming design. In this work, we propose a HAD architecture that employs Butler matrices to synthesize DFT beams over a uniform linear array. By exploiting the Cauchy-like displacement structure of the beamformed signal, we introduce a second-order statistics estimation algorithm that achieves near-optimal accuracy, approaching the Cram\'er-Rao Lower Bound (CRLB) and outperforming state-of-the-art methods in simulation.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:14:04 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Rivas-Costa",
"Miguel",
""
],
[
"Mosquera",
"Carlos",
""
]
] |
2509.03489
|
Yixuan Huang
|
Yixuan Huang, Yuan Gao, Wei Li, Seiji Yunoki, Sadamichi Maekawa
|
Dissipationless dynamics of spin supersolid states in a spin-1/2 triangular antiferromagnet with impurities
|
8+7 pages, 4+9 figures
| null | null | null |
cond-mat.str-el cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Motivated by recent experimental observations of possible spin supersolid states in triangular lattice compounds, we study the dynamical properties of various ground states in the spin-1/2 easy-axis antiferromagnetic Heisenberg model with impurities under magnetic fields, using numerical Density Matrix Renormalization Group methods. For both spin supersolid states in the low and high fields, the gapless Goldstone mode at the $K$ points remains robust with impurities, which is related to the presence of spin superfluidity. As a comparison, we find a splitting of magnon band at the same impurity density level in the conventional magnetic state, the so-called up-up-down state. In addition, the finite superfluid stiffness probed by the twisted phase in the spin supersolid states is consistent with the excitation spectrum. We argue that this excitation spectrum with impurity provides direct evidence of the dissipationless dynamics in the spin supersolid states, which could be tested in neutron scattering experiments.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:18:04 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Huang",
"Yixuan",
""
],
[
"Gao",
"Yuan",
""
],
[
"Li",
"Wei",
""
],
[
"Yunoki",
"Seiji",
""
],
[
"Maekawa",
"Sadamichi",
""
]
] |
2509.03491
|
Kanchan Shaikh
|
Kanchan Shaikh, Taejun Yoo, Zeyuan Zhu, Qiuyang Li, Amalya C. Johnson, Hui Deng, Fang Liu, and Yuki Kobayashi
|
Enhanced second-harmonic generation from WS$_2$/ReSe$_2$ heterostructure
| null | null | null | null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
Van der Waals stacking presents new opportunities for nonlinear optics with its remarkable tunability and scalability. However, the fundamental role of interlayer interactions in modifying the overall nonlinear optical susceptibilities remains elusive. In this letter, we report an anisotropic enhancement of second-harmonic generation (SHG) from a WS$_2$/ReSe$_2$ heterobilayer, where the individual composite layers possess distinctive crystal phases. We investigate polarization-resolved response and twist-angle dependence in SHG and reveal that band alignment alone is insufficient to explain the observed anisotropy in the modified SHG response. Spectral shifts in excitonic features highlight band renormalization, supporting the role of hybridization between the two layers. Furthermore, SHG enhancement is highly anisotropic and can even be suppressed in some orientations, suggesting possible intensity-borrowing mechanisms within the heterostructure. Our work demonstrates the ability to tune both the intensity and polarization dependence of nonlinear optical responses with van der Waals stacking of distinctive crystal phases.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:22:50 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Shaikh",
"Kanchan",
""
],
[
"Yoo",
"Taejun",
""
],
[
"Zhu",
"Zeyuan",
""
],
[
"Li",
"Qiuyang",
""
],
[
"Johnson",
"Amalya C.",
""
],
[
"Deng",
"Hui",
""
],
[
"Liu",
"Fang",
""
],
[
"Kobayashi",
"Yuki",
""
]
] |
2509.03492
|
Matthew Lastovka
|
Matthew Lastovka, B. Scott Gaudi, Samson A. Johnson, Matthew T. Penny, Eamonn Kerins, Nicholas J. Rattenbury
|
Predictions of the Nancy Grace Roman Space Telescope Galactic Exoplanet Survey. III. Detectability of Giant Exomoons of Wide Separation Giant Planets
|
28 pages, 12 figures, accepted to AJ
| null | null | null |
astro-ph.EP astro-ph.IM
|
http://creativecommons.org/licenses/by/4.0/
|
The Nancy Grace Roman Space Telescope (Roman) will conduct a Galactic Exoplanet Survey (RGES) to discover bound and free-floating exoplanets using gravitational microlensing. Roman should be sensitive to lenses with mass down to ~ 0.02 $M_{\oplus}$, or roughly the mass of Ganymede. Thus the detection of moons with masses similar to the giant moons in our Solar System is possible with Roman. Measuring the demographics of exomoons will provide constraints on both moon and planet formation. We conduct simulations of Roman microlensing events to determine the effects of exomoons on microlensing light curves, and whether these effects are detectable with Roman. We focus on giant planets from 30 $M_{\oplus}$ to 10 $M_{Jup}$ on orbits from 0.3 to 30 AU, and assume that each planet is orbited by a moon with moon-planet mass ratio from $10^{-4}$ to $10^{-2}$ and separations from 0.1 to 0.5 planet Hill radii. We find that Roman is sensitive to exomoons, although the number of expected detections is only of order one over the duration of the survey, unless exomoons are more common or massive than we assumed. We argue that changes in the survey strategy, in particular focusing on a few fields with higher cadence, may allow for the detection of more exomoons with Roman. Regardless, the ability to detect exomoons reinforces the need to develop robust methods for modeling triple lens microlensing events to fully utilize the capabilities of Roman.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:23:06 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Lastovka",
"Matthew",
""
],
[
"Gaudi",
"B. Scott",
""
],
[
"Johnson",
"Samson A.",
""
],
[
"Penny",
"Matthew T.",
""
],
[
"Kerins",
"Eamonn",
""
],
[
"Rattenbury",
"Nicholas J.",
""
]
] |
2509.03493
|
Han Shen
|
Han Shen
|
On Entropy Control in LLM-RL Algorithms
| null | null | null | null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
For RL algorithms, appropriate entropy control is crucial to their effectiveness. To control the policy entropy, a commonly used method is entropy regularization, which is adopted in various popular RL algorithms including PPO, SAC and A3C. Although entropy regularization proves effective in robotic and games RL conventionally, studies found that it gives weak to no gains in LLM-RL training. In this work, we study the issues of entropy bonus in LLM-RL setting. Specifically, we first argue that the conventional entropy regularization suffers from the LLM's extremely large response space and the sparsity of the optimal outputs. As a remedy, we propose AEnt, an entropy control method that utilizes a new clamped entropy bonus with an automatically adjusted coefficient. The clamped entropy is evaluated with the re-normalized policy defined on certain smaller token space, which encourages exploration within a more compact response set. In addition, the algorithm automatically adjusts entropy coefficient according to the clamped entropy value, effectively controlling the entropy-induced bias while leveraging the entropy's benefits. AEnt is tested in math-reasoning tasks under different base models and datasets, and it is observed that AEnt outperforms the baselines consistently across multiple benchmarks.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:23:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Shen",
"Han",
""
]
] |
2509.03494
|
Mohammed El Hassouni
|
Yahya Benmahane, Mohammed El Hassouni
|
Parameter-Efficient Adaptation of mPLUG-Owl2 via Pixel-Level Visual Prompts for NR-IQA
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we propose a novel parameter-efficient adaptation method for No- Reference Image Quality Assessment (NR-IQA) using visual prompts optimized in pixel-space. Unlike full fine-tuning of Multimodal Large Language Models (MLLMs), our approach trains only 600K parameters at most (< 0.01% of the base model), while keeping the underlying model fully frozen. During inference, these visual prompts are combined with images via addition and processed by mPLUG-Owl2 with the textual query "Rate the technical quality of the image." Evaluations across distortion types (synthetic, realistic, AI-generated) on KADID- 10k, KonIQ-10k, and AGIQA-3k demonstrate competitive performance against full finetuned methods and specialized NR-IQA models, achieving 0.93 SRCC on KADID-10k. To our knowledge, this is the first work to leverage pixel-space visual prompts for NR-IQA, enabling efficient MLLM adaptation for low-level vision tasks. The source code is publicly available at https: // github. com/ yahya-ben/ mplug2-vp-for-nriqa .
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:23:24 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Benmahane",
"Yahya",
""
],
[
"Hassouni",
"Mohammed El",
""
]
] |
2509.03495
|
Thinh Le
|
Thinh Viet Le, Md Obaidur Rahman, and Vassilis Kekatos
|
Learning AC Power Flow Solutions using a Data-Dependent Variational Quantum Circuit
|
7 pages, 6 figures, accepted for the IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids 2025
| null | null | null |
quant-ph cs.LG cs.SY eess.SY math.OC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Interconnection studies require solving numerous instances of the AC load or power flow (AC PF) problem to simulate diverse scenarios as power systems navigate the ongoing energy transition. To expedite such studies, this work leverages recent advances in quantum computing to find or predict AC PF solutions using a variational quantum circuit (VQC). VQCs are trainable models that run on modern-day noisy intermediate-scale quantum (NISQ) hardware to accomplish elaborate optimization and machine learning (ML) tasks. Our first contribution is to pose a single instance of the AC PF as a nonlinear least-squares fit over the VQC trainable parameters (weights) and solve it using a hybrid classical/quantum computing approach. The second contribution is to feed PF specifications as features into a data-embedded VQC and train the resultant quantum ML (QML) model to predict general PF solutions. The third contribution is to develop a novel protocol to efficiently measure AC-PF quantum observables by exploiting the graph structure of a power network. Preliminary numerical tests indicate that the proposed VQC models attain enhanced prediction performance over a deep neural network despite using much fewer weights. The proposed quantum AC-PF framework sets the foundations for addressing more elaborate grid tasks via quantum computing.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:24:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Le",
"Thinh Viet",
""
],
[
"Rahman",
"Md Obaidur",
""
],
[
"Kekatos",
"Vassilis",
""
]
] |
2509.03496
|
Qisheng Wang
|
Qisheng Wang
|
Information-Theoretic Lower Bounds for Approximating Monomials via Optimal Quantum Tsallis Entropy Estimation
|
Submitted to CCC 2025. 36 pages, 1 figure, 1 algorithm
| null | null | null |
quant-ph cs.CC cs.IT math.CA math.IT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper reveals a conceptually new connection from information theory to approximation theory via quantum algorithms for entropy estimation. Specifically, we provide an information-theoretic lower bound $\Omega(\sqrt{n})$ on the approximate degree of the monomial $x^n$, compared to the analytic lower bounds shown in Newman and Rivlin (Aequ. Math. 1976) via Fourier analysis and in Sachdeva and Vishnoi (Found. Trends Theor. Comput. Sci. 2014) via the Markov brothers' inequality. This is done by relating the polynomial approximation of monomials to quantum Tsallis entropy estimation. This further implies a quantum algorithm that estimates to within additive error $\varepsilon$ the Tsallis entropy of integer order $q \geq 2$ of an unknown probability distribution $p$ or an unknown quantum state $\rho$, using $\widetilde \Theta(\frac{1}{\sqrt{q}\varepsilon})$ queries to the quantum oracle that produces a sample from $p$ or prepares a copy of $\rho$, improving the prior best $O(\frac{1}{\varepsilon})$ via the Shift test due to Ekert, Alves, Oi, Horodecki, Horodecki and Kwek (Phys. Rev. Lett. 2002). To the best of our knowledge, this is the first quantum entropy estimator with optimal query complexity (up to polylogarithmic factors) for all parameters simultaneously.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:25:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Wang",
"Qisheng",
""
]
] |
2509.03498
|
Han Li
|
Han Li, Xinyu Peng, Yaoming Wang, Zelin Peng, Xin Chen, Rongxiang Weng, Jingang Wang, Xunliang Cai, Wenrui Dai, Hongkai Xiong
|
OneCAT: Decoder-Only Auto-Regressive Model for Unified Understanding and Generation
|
technical report
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
We introduce OneCAT, a unified multimodal model that seamlessly integrates understanding, generation, and editing within a novel, pure decoder-only transformer architecture. Our framework uniquely eliminates the need for external components such as Vision Transformers (ViT) or vision tokenizer during inference, leading to significant efficiency gains, especially for high-resolution inputs. This is achieved through a modality-specific Mixture-of-Experts (MoE) structure trained with a single autoregressive (AR) objective, which also natively supports dynamic resolutions. Furthermore, we pioneer a multi-scale visual autoregressive mechanism within the Large Language Model (LLM) that drastically reduces decoding steps compared to diffusion-based methods while maintaining state-of-the-art performance. Our findings demonstrate the powerful potential of pure autoregressive modeling as a sufficient and elegant foundation for unified multimodal intelligence. As a result, OneCAT sets a new performance standard, outperforming existing open-source unified multimodal models across benchmarks for multimodal generation, editing, and understanding.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:29:50 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Han",
""
],
[
"Peng",
"Xinyu",
""
],
[
"Wang",
"Yaoming",
""
],
[
"Peng",
"Zelin",
""
],
[
"Chen",
"Xin",
""
],
[
"Weng",
"Rongxiang",
""
],
[
"Wang",
"Jingang",
""
],
[
"Cai",
"Xunliang",
""
],
[
"Dai",
"Wenrui",
""
],
[
"Xiong",
"Hongkai",
""
]
] |
2509.03499
|
Kevin Barnard
|
Kevin Barnard, Elaine Liu, Kristine Walz, Brian Schlining, Nancy Jacobsen Stout, Lonny Lundsten
|
DeepSea MOT: A benchmark dataset for multi-object tracking on deep-sea video
|
5 pages, 3 figures, dataset available at https://huggingface.co/datasets/MBARI-org/DeepSea-MOT
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Benchmarking multi-object tracking and object detection model performance is an essential step in machine learning model development, as it allows researchers to evaluate model detection and tracker performance on human-generated 'test' data, facilitating consistent comparisons between models and trackers and aiding performance optimization. In this study, a novel benchmark video dataset was developed and used to assess the performance of several Monterey Bay Aquarium Research Institute object detection models and a FathomNet single-class object detection model together with several trackers. The dataset consists of four video sequences representing midwater and benthic deep-sea habitats. Performance was evaluated using Higher Order Tracking Accuracy, a metric that balances detection, localization, and association accuracy. To the best of our knowledge, this is the first publicly available benchmark for multi-object tracking in deep-sea video footage. We provide the benchmark data, a clearly documented workflow for generating additional benchmark videos, as well as example Python notebooks for computing metrics.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:30:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Barnard",
"Kevin",
""
],
[
"Liu",
"Elaine",
""
],
[
"Walz",
"Kristine",
""
],
[
"Schlining",
"Brian",
""
],
[
"Stout",
"Nancy Jacobsen",
""
],
[
"Lundsten",
"Lonny",
""
]
] |
2509.03500
|
Itai Zilberstein
|
Itai Zilberstein, Alberto Candela, Steve Chien
|
Real-Time Instrument Planning and Perception for Novel Measurements of Dynamic Phenomena
|
Appears in Proceedings of 18th Symposium on Advanced Space Technologies in Robotics and Automation
| null | null | null |
cs.RO cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Advancements in onboard computing mean remote sensing agents can employ state-of-the-art computer vision and machine learning at the edge. These capabilities can be leveraged to unlock new rare, transient, and pinpoint measurements of dynamic science phenomena. In this paper, we present an automated workflow that synthesizes the detection of these dynamic events in look-ahead satellite imagery with autonomous trajectory planning for a follow-up high-resolution sensor to obtain pinpoint measurements. We apply this workflow to the use case of observing volcanic plumes. We analyze classification approaches including traditional machine learning algorithms and convolutional neural networks. We present several trajectory planning algorithms that track the morphological features of a plume and integrate these algorithms with the classifiers. We show through simulation an order of magnitude increase in the utility return of the high-resolution instrument compared to baselines while maintaining efficient runtimes.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:32:15 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zilberstein",
"Itai",
""
],
[
"Candela",
"Alberto",
""
],
[
"Chien",
"Steve",
""
]
] |
2509.03501
|
Honglu Zhou
|
Honglu Zhou, Xiangyu Peng, Shrikant Kendre, Michael S. Ryoo, Silvio Savarese, Caiming Xiong, Juan Carlos Niebles
|
Strefer: Empowering Video LLMs with Space-Time Referring and Reasoning via Synthetic Instruction Data
|
This technical report serves as the archival version of our paper accepted at the ICCV 2025 Workshop. For more information, please visit our project website: https://strefer.github.io/
| null | null | null |
cs.CV cs.AI cs.HC cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Next-generation AI companions must go beyond general video understanding to resolve spatial and temporal references in dynamic, real-world environments. Existing Video Large Language Models (Video LLMs), while capable of coarse-level comprehension, struggle with fine-grained, spatiotemporal reasoning, especially when user queries rely on time-based event references for temporal anchoring, or gestural cues for spatial anchoring to clarify object references and positions. To bridge this critical gap, we introduce Strefer, a synthetic instruction data generation framework designed to equip Video LLMs with spatiotemporal referring and reasoning capabilities. Strefer produces diverse instruction-tuning data using a data engine that pseudo-annotates temporally dense, fine-grained video metadata, capturing rich spatial and temporal information in a structured manner, including subjects, objects, their locations as masklets, and their action descriptions and timelines. Our approach enhances the ability of Video LLMs to interpret spatial and temporal references, fostering more versatile, space-time-aware reasoning essential for real-world AI companions. Without using proprietary models, costly human annotation, or the need to annotate large volumes of new videos, experimental evaluations show that models trained with data produced by Strefer outperform baselines on tasks requiring spatial and temporal disambiguation. Additionally, these models exhibit enhanced space-time-aware reasoning, establishing a new foundation for perceptually grounded, instruction-tuned Video LLMs.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:33:20 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhou",
"Honglu",
""
],
[
"Peng",
"Xiangyu",
""
],
[
"Kendre",
"Shrikant",
""
],
[
"Ryoo",
"Michael S.",
""
],
[
"Savarese",
"Silvio",
""
],
[
"Xiong",
"Caiming",
""
],
[
"Niebles",
"Juan Carlos",
""
]
] |
2509.03502
|
Zhuoyu Chen
|
Zihao Nie, Yueying Li, Wei Lv, Lizhi Xu, Zhicheng Jiang, Peng Fu, Guangdi Zhou, Wenhua Song, Yaqi Chen, Heng Wang, Haoliang Huang, Junhao Lin, Dawei Shen, Peng Li, Qi-Kun Xue, Zhuoyu Chen
|
Ambient-pressure superconductivity and electronic structures of engineered hybrid nickelate films
| null | null | null | null |
cond-mat.supr-con
|
http://creativecommons.org/licenses/by/4.0/
|
Ruddlesden-Popper (RP) nickelates have emerged as a crucial platform for exploring the mechanisms of high-temperature superconductivity. However, the Fermi surface topology required for superconductivity remains elusive. Here, we report the thin film growth and ambient-pressure superconductivity of both hybrid monolayer-bilayer (1212) and pure bilayer (2222) structures, together with the absence of superconductivity in hybrid monolayer-trilayer (1313) structure, under identical compressive epitaxial strain. The onset superconducting transition temperature is up to 50 K, exceeding the McMillan limit, in the 1212 structure. Angle-resolved photoemission spectroscopy reveals key Fermi surface differences in these atomically-engineered structures. In superconducting 1212 and 2222 films, a dispersive hole-like band (i.e. the {\gamma} band) crosses the Fermi level, surrounding the Brillouin zone corner. In contrast, the top of the {\gamma} flat band is observed ~70 meV below the Fermi level in the non-superconducting 1313 films. Our findings expand the family of ambient-pressure nickelate superconductors and establish a connection between structural configuration, electronic structure, and the emergence of superconductivity in nickelates.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:33:27 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Nie",
"Zihao",
""
],
[
"Li",
"Yueying",
""
],
[
"Lv",
"Wei",
""
],
[
"Xu",
"Lizhi",
""
],
[
"Jiang",
"Zhicheng",
""
],
[
"Fu",
"Peng",
""
],
[
"Zhou",
"Guangdi",
""
],
[
"Song",
"Wenhua",
""
],
[
"Chen",
"Yaqi",
""
],
[
"Wang",
"Heng",
""
],
[
"Huang",
"Haoliang",
""
],
[
"Lin",
"Junhao",
""
],
[
"Shen",
"Dawei",
""
],
[
"Li",
"Peng",
""
],
[
"Xue",
"Qi-Kun",
""
],
[
"Chen",
"Zhuoyu",
""
]
] |
2509.03503
|
Gwendolyne Legate
|
Gwen Legate, Irina Rish, Eugene Belilovsky
|
Warming Up for Zeroth-Order Federated Pre-Training with Low Resource Clients
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Federated learning enables collaborative model training across numerous edge devices without requiring participants to share data; however, memory and communication constraints on these edge devices may preclude their participation in training. We consider a setting in which a subset of edge devices are below a critical memory or communication threshold required to conduct model updates. Under typical federated optimization algorithms, these devices are excluded from training which renders their data inaccessible and increases system induced bias. We are inspired by MeZO, a zeroth-order method used for memory-efficient fine-tuning. The increased variance inherent to zeroth-order gradient approximations has relegated previous zeroth-order optimizers exclusively to the domain of fine tuning; a limitation we seek to correct. We devise a federated, memory-efficient zeroth-order optimizer, ZOWarmUp that permits zeroth-order training from a random initialization. ZOWarmUp leverages differing client capabilities and careful variance reduction techniques to facilitate participation of under-represented, low-resource clients in model training. Like other federated zeroth-order methods, ZOWarmUp eliminates the need for edge devices to transmit their full gradients to the server and instead relies on only a small set of random seeds, rendering the up-link communication cost negligible. We present experiments using various datasets and model architectures to show that ZOWarmUp is a robust algorithm that can can be applied under a wide variety of circumstances. For systems with a high proportion of edge devices that would otherwise be excluded from training, this algorithm provides access to a greater volume and diversity of data, thus improving training outcomes.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:35:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Legate",
"Gwen",
""
],
[
"Rish",
"Irina",
""
],
[
"Belilovsky",
"Eugene",
""
]
] |
2509.03504
|
Nicholas Shepherd-Barron
|
I. Grojnowski and N. I. Shepherd-Barron
|
Recognizing flag varieties and reductive groups
| null | null | null | null |
math.AG
|
http://creativecommons.org/licenses/by/4.0/
|
Fix a flat and projective morphism $X\rightarrow\Sigma$ of schemes. We show, first, that any set of $\mathbb{P}^1$-fibrations on $X$ defines a set of simple roots, a set of simple coroots and a Cartan matrix $C$. Second, $X$ is an \'etale $F$-bundle over some projective $\Sigma$-scheme, where $F$ is the flag variety of the adjoint Chevalley group over the integers defined by $C$. In particular, if the simple roots generate the N\'eron--Severi group of $X$ relative to $\Sigma$ and $X$ is cohomologically flat in degree zero over $\Sigma$ then $X$ is a form of $F$. When $X$ is a smooth Fano variety over the complex numbers all of whose extremal rays are accounted for by these fibrations this is due to Occhetta, Sol\'a-Conde, Watanabe and Wi\'sniewski. Third, we recover, in a uniform way, the isomorphism and isogeny theorems of Chevalley and Demazure: over any base a pinned reductive group is determined by its pinned root datum, and a $p$-morphism of pinned root data determines a unique homomorphism of the corresponding groups.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:36:32 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Grojnowski",
"I.",
""
],
[
"Shepherd-Barron",
"N. I.",
""
]
] |
2509.03505
|
Xingxuan Zhang
|
Xingxuan Zhang, Gang Ren, Han Yu, Hao Yuan, Hui Wang, Jiansheng Li, Jiayun Wu, Lang Mo, Li Mao, Mingchao Hao, Ningbo Dai, Renzhe Xu, Shuyang Li, Tianyang Zhang, Yue He, Yuanrui Wang, Yunjia Zhang, Zijing Xu, Dongzhe Li, Fang Gao, Hao Zou, Jiandong Liu, Jiashuo Liu, Jiawei Xu, Kaijie Cheng, Kehan Li, Linjun Zhou, Qing Li, Shaohua Fan, Xiaoyu Lin, Xinyan Han, Xuanyue Li, Yan Lu, Yuan Xue, Yuanyuan Jiang, Zimu Wang, Zhenlei Wang, Peng Cui
|
LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence
|
56 pages
| null | null | null |
cs.LG cs.AI cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We argue that progress toward general intelligence requires complementary foundation models grounded in language, the physical world, and structured data. This report presents LimiX, the first installment of our large structured-data models (LDMs). LimiX treats structured data as a joint distribution over variables and missingness, thus capable of addressing a wide range of tabular tasks through query-based conditional prediction via a single model. LimiX is pretrained using masked joint-distribution modeling with an episodic, context-conditional objective, where the model predicts for query subsets conditioned on dataset-specific contexts, supporting rapid, training-free adaptation at inference. We evaluate LimiX across 10 large structured-data benchmarks with broad regimes of sample size, feature dimensionality, class number, categorical-to-numerical feature ratio, missingness, and sample-to-feature ratios. With a single model and a unified interface, LimiX consistently surpasses strong baselines including gradient-boosting trees, deep tabular networks, recent tabular foundation models, and automated ensembles, as shown in Figure 1 and Figure 2. The superiority holds across a wide range of tasks, such as classification, regression, missing value imputation, and data generation, often by substantial margins, while avoiding task-specific architectures or bespoke training per task. All LimiX models are publicly accessible under Apache 2.0.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:39:08 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhang",
"Xingxuan",
""
],
[
"Ren",
"Gang",
""
],
[
"Yu",
"Han",
""
],
[
"Yuan",
"Hao",
""
],
[
"Wang",
"Hui",
""
],
[
"Li",
"Jiansheng",
""
],
[
"Wu",
"Jiayun",
""
],
[
"Mo",
"Lang",
""
],
[
"Mao",
"Li",
""
],
[
"Hao",
"Mingchao",
""
],
[
"Dai",
"Ningbo",
""
],
[
"Xu",
"Renzhe",
""
],
[
"Li",
"Shuyang",
""
],
[
"Zhang",
"Tianyang",
""
],
[
"He",
"Yue",
""
],
[
"Wang",
"Yuanrui",
""
],
[
"Zhang",
"Yunjia",
""
],
[
"Xu",
"Zijing",
""
],
[
"Li",
"Dongzhe",
""
],
[
"Gao",
"Fang",
""
],
[
"Zou",
"Hao",
""
],
[
"Liu",
"Jiandong",
""
],
[
"Liu",
"Jiashuo",
""
],
[
"Xu",
"Jiawei",
""
],
[
"Cheng",
"Kaijie",
""
],
[
"Li",
"Kehan",
""
],
[
"Zhou",
"Linjun",
""
],
[
"Li",
"Qing",
""
],
[
"Fan",
"Shaohua",
""
],
[
"Lin",
"Xiaoyu",
""
],
[
"Han",
"Xinyan",
""
],
[
"Li",
"Xuanyue",
""
],
[
"Lu",
"Yan",
""
],
[
"Xue",
"Yuan",
""
],
[
"Jiang",
"Yuanyuan",
""
],
[
"Wang",
"Zimu",
""
],
[
"Wang",
"Zhenlei",
""
],
[
"Cui",
"Peng",
""
]
] |
2509.03506
|
Stefan Schrott
|
Mathias Beiglb\"ock, Gudmund Pammer, Stefan Schrott
|
A Brenier Theorem on $(\mathcal{P}_2 (\mathcal{P}_2(\mathbb{R}^d )), W_2 )$ and Applications to Adapted Transport
| null | null | null | null |
math.PR math.FA math.MG math.OC
|
http://creativecommons.org/licenses/by/4.0/
|
Brenier's fundamental theorem characterizes optimal transport plans for measures $\mu, \nu$ on $\mathbb{R}^d$ and quadratic distance costs in terms of gradients of convex functions. In particular it guarantees the existence of optimal transport maps for measures which are absolutely continuous wrt Lebesgue measure.
Our goal is to provide a version of this result for measures $P,Q$ on $\mathcal{P}_2(\mathbb{R}^d)$ and costs given by the squared Wasserstein distance $W_2^2(\mu, \nu)$. We characterize optimizers in terms of convexity of the Lions lift. This is based on an observation which seems to be of independent interest: the $c$-transform of a functional $\phi$, where $c(\mu, \nu)$ denotes maximal covariance of $\mu, \nu$ corresponds precisely to the Legendre transform of the Lions lift of $\phi$. Moreover we show that for typical $P \in\mathcal{P}_2(\mathbb{R}^d)$ the optimizer is unique and given by a transport map. In the absence of a canonical reference measure on $\mathcal{P}_2(\mathbb{R}^d)$ we use a topological notion to make `typical' precise. Specifically we show that the transport regular measures are of second Baire category.
A particular motivation for our article stems from the theory of adapted transport where the adapted Wasserstein distance provides an adequate distance between stochastic processes. In contrast to other metrics, the adapted Wasserstein distance yields continuity of Doob-decomposition, optimal stopping and stochastic control problems. Based on our results for measures on $\mathcal{P}_2(\mathbb{R}^d)$ we obtain a first Brenier-type theorem for the adapted Wasserstein distance.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:41:32 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Beiglböck",
"Mathias",
""
],
[
"Pammer",
"Gudmund",
""
],
[
"Schrott",
"Stefan",
""
]
] |
2509.03507
|
Md Sabir Ali
|
Md Sabir Ali, Abhishek Negi, Sanjay Pant
|
Influence of Perfect Fluid Dark Matter on Shadow Observables of Yang-Mills modified charged black holes
|
25 Pages, 39 Figures, Comments are welcome
| null | null | null |
gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We investigate the influence of perfect fluid dark matter (PFDM) on Yang--Mills--inspired charged black holes, with a particular focus on the resulting modifications to key black hole observables. By embedding a PFDM term into the spacetime geometry, we examine the alterations in shadow morphology, photon geodesics, and the associated energy emission spectra. Our analysis reveals that PFDM induces notable deviations in the shadow size, shape, and circularity, and significantly impacts the stability of circular orbits. Furthermore, the energy emission rate exhibits a strong dependence on both the Yang--Mills charge and the dark matter distribution. These results indicate that environmental effects arising from dark matter can imprint observable signatures on black hole shadows and radiation processes, offering a potential pathway to constrain dark matter models and probe non-Kerr geometries with forthcoming high-precision observations such as those from the Event Horizon Telescope and next-generation interferometers.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:42:47 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ali",
"Md Sabir",
""
],
[
"Negi",
"Abhishek",
""
],
[
"Pant",
"Sanjay",
""
]
] |
2509.03508
|
Sonej Alam
|
Sonej Alam, Somasri Sen, Soumitra Sengupta
|
Bouncing Cosmologies in modified gravity with space time torsion
|
16 pages, 11 figures
| null | null | null |
gr-qc astro-ph.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We explore the possibility of realizing a non-singular bounce in the early universe within the framework of modified gravity with spacetime torsion. In Einstein Cartan theory, torsion is embedded in the spacetime by adding an antisymmetric part in affine connection . We consider generalized version of the framework as $f(\bar{R})$, $\bar{R}$ being the scalar of the modified curvature tensor. $f(\bar{R})$ gravity is recast in Einstein frame as non-minimally coupled scalar tensor theory where the scalar field gets coupled with a rank 2 antisymmetric torsion field through derivative couplings. We investigate whether the introduction of three additional torsion-dependent terms in Einstein frame help to realize a bounce. We first explore this cosmological system in the background of a homogeneous and isotropic FRW spacetime but inclusion of the torsion terms are insufficient to produce a bounce in this symmetric setting. Motivated by this limitation, we relax the symmetry and generalize the background to include inhomogeneity and anisotropy. In this setup, the dynamics is modified in such a way that a bouncing solution is possible without invoking phantom fields or energy condition violations. We have also found the exact solutions of all the fields and reconstructed the modified gravity form.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:46:29 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Alam",
"Sonej",
""
],
[
"Sen",
"Somasri",
""
],
[
"Sengupta",
"Soumitra",
""
]
] |
2509.03509
|
Ivan Horvath
|
Ivan Horv\'ath
|
Gluon Condensate via Dirac Spectral Density: IR Phase, Scale Anomaly and IR Decoupling
|
11 pages, 2 figures
| null | null | null |
hep-lat hep-ph hep-th nucl-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Quark and gluon scalar densities, $\langle \bar{\psi} \psi \rangle$ and $\langle F^2 \rangle$, reflect the degree of scale-invariance violations in SU(N) gauge theories with fundamental quarks. It is known that $\langle \bar{\psi} \psi \rangle$ can be usefully scale-decomposed via spectral density $\rho(\lambda)$ of Dirac modes. Here I give such formula for $\langle F^2 \rangle$, which reveals that gluon condensate is a strictly UV quantity. For the recently-found IR phase [1,2], where the infrared (IR) degrees of freedom separate out and become independent of the system's bulk, it implies that $\langle F^2 \rangle$ due to this IR part vanishes. Its glue thus doesn't contribute to scale anomaly of the entire system and is, in this sense, scale invariant consistently with the original claim. Associated formulas are used us to define IR decoupling of glue, which may serve as an alternative indicator of IR phase transition. Using the simplest form of coherent lattice QCD, we express the effective action of full QCD entirely via Dirac spectral density.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:47:04 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Horváth",
"Ivan",
""
]
] |
2509.03510
|
Abbas Zohrevand
|
Abbas Zohrevand, Javad Sadri, Zahra Imani
|
A comprehensive Persian offline handwritten database for investigating the effects of heritability and family relationships on handwriting
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
This paper introduces a comprehensive database for research and investigation on the effects of inheritance on handwriting. A database has been created that can be used to answer questions such as: Is there a genetic component to handwriting? Is handwriting inherited? Do family relationships affect handwriting? Varieties of samples of handwritten components such as: digits, letters, shapes and free paragraphs of 210 families including (grandparents, parents, uncles, aunts, siblings, cousins, nephews and nieces) have been collected using specially designed forms, and family relationships of all writers are captured. To the best of our knowledge, no such database is presently available. Based on comparisons and investigation of features of handwritings of family members, similarities among their features and writing styles are detected. Our database is freely available to the pattern recognition community and hope it will pave the way for investigations on the effects of inheritance and family relationships on handwritings.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:47:23 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zohrevand",
"Abbas",
""
],
[
"Sadri",
"Javad",
""
],
[
"Imani",
"Zahra",
""
]
] |
2509.03511
|
Danilo Triggiani
|
Danilo Triggiani, Cosmo Lupo
|
Achieving quantum-limited sub-Rayleigh identification of incoherent sources with arbitrary intensities
|
14 pages, 3 figures
| null | null | null |
quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The Rayleigh diffraction limit imposes a fundamental restriction on the resolution of direct imaging systems, hindering the identification of incoherent optical sources, such as celestial bodies in astronomy and fluorophores in bioimaging. Recent advances in quantum sensing have shown that this limit can be circumvented through spatial demultiplexing (SPADE) and photon detection. Notably, the latter is a semi-classical detection strategy, being SPADE a linear transformation of the field and photon detection a measurement of intensity. However, the general optimality for arbitrary intensity distributions and bright sources remains unproven. In this work, we develop a general model for incoherent light with arbitrary intensity collected by passive, linear optical systems. We employ this framework to compute the quantum Chernoff exponent for generic incoherent-source discrimination problems, and we analyze several special cases, with particular focus on the subdiffraction regime for Gaussian point-spread functions. We show that, surprisingly, SPADE measurements do not always saturate the quantum Chernoff bound in the subdiffraction regime; the quantum optimality holds only when certain compatibility conditions, such as covariance matrix commutativity, are met. These findings suggest that collective measurements may actually be needed to achieve the ultimate quantum Chernoff bound for the discrimination of specific incoherent sources. For the fully general case, our analysis can still be used to find the best SPADE configurations, generally achieved through a hypotheses-dependent rotation of the SPADE modes. Our results advance the theory of quantum-limited optical discrimination, with possible applications in diagnostics, automated image interpretation, and galaxy identification.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:52:01 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Triggiani",
"Danilo",
""
],
[
"Lupo",
"Cosmo",
""
]
] |
2509.03512
|
Joseph Sartini
|
Joseph Sartini, Scott Zeger, and Ciprian Crainiceanu
|
Bayesian Multivariate Sparse Functional PCA
|
24 pages, 7 figures for main text. Appendix contains supplemental material
| null | null | null |
stat.ME
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Functional Principal Components Analysis (FPCA) provides a parsimonious, semi-parametric model for multivariate, sparsely-observed functional data. Frequentist FPCA approaches estimate principal components (PCs) from the data, then condition on these estimates in subsequent analyses. As an alternative, we propose a fully Bayesian inferential framework for multivariate, sparse functional data (MSFAST) which explicitly models the PCs and incorporates their uncertainty. MSFAST builds upon the FAST approach to FPCA for univariate, densely-observed functional data. Like FAST, MSFAST represents PCs using orthonormal splines, samples the orthonormal spline coefficients using parameter expansion, and enforces eigenvalue ordering during model fit. MSFAST extends FAST to multivariate, sparsely-observed data by (1) standardizing each functional covariate to mitigate poor posterior conditioning due to disparate scales; (2) using a better-suited orthogonal spline basis; (3) parallelizing likelihood calculations over covariates; (4) updating parameterizations and priors for computational stability; (5) using a Procrustes-based posterior alignment procedure; and (6) providing efficient prediction routines. We evaluated MSFAST alongside existing implementations using simulations. MSFAST produces uniquely valid inferences and accurate estimates, particularly for smaller signals. MSFAST is motivated by and applied to a study of child growth, with an accompanying vignette illustrating the implementation step-by-step.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:52:16 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sartini",
"Joseph",
""
],
[
"Zeger",
"Scott",
""
],
[
"Crainiceanu",
"Ciprian",
""
]
] |
2509.03513
|
Paulo A. Faria da Veiga
|
Paulo A. Faria da Veiga and Michael O'Carroll
|
On Charge Conjugation, Correlations, Elitzur's Theorem and the Mass Gap Problem in Lattice $SU(N)$ Yang-Mills Models in $d=4$ Dimensions
|
20 pages
| null | null | null |
hep-th hep-lat math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We consider a four-dimensional Euclidean Wilson lattice Yang-Mills model with gauge group $SU(N)$, and the associated lattice Euclidean quantum field theory constructed by Osterwalder-Schrader-Seiler via a Feynman-Kac formula. In this model, to each lattice bond $b$ there is assigned a bond variable $U_b\in SU(N)$. Gluon fields are parameters in the Lie algebra of $SU(N)$. We define a charge conjugation operator $\mathcal C$ in the physical Hilbert space $H$ and prove that, for $N\not=2$, $H$ admits an orthogonal decomposition into two sectors with charge conjugation $\pm1$. There is only one sector for $N=2$. In the space of correlations, a charge conjugation operator $C_E$ is defined and a similar decomposition holds. Besides, a version of Elitzur's theorem is shown; applications are given. It is proven that the expectation averages of two distinct lattice vector potential correlators is zero. Surprisingly, the expectation of two distinct field strength tensors is also zero. In the gluon field parametrization it is known that Wilson action is bounded quadratically in the gluon fields. Towards solving the mass gap problem, in the gluon field expansion of the action, we show there are no local terms, such as a mass term. However, the action associated with the exponent of the exponentiated Haar measure density has a positive local mass term which is proportional to the square of the gauge coupling $\beta=g^{-2}>0$ and the $SU(N)$ quadratic Casimir operator eigenvalue. Our results also hold in dimension three. Of course, the orthogonal decomposition of the Hilbert space $H$ has consequences in the analysis of the truncated two-point plaquette field correlation and the Yang-Mills mass gap problem: for $N\not=2$ and at least for small $\beta$, a multiplicity-two one-particle glueball state (two mass gaps) is expected to be present in the energy-momentum spectrum of the Yang-Mills model.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:53:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"da Veiga",
"Paulo A. Faria",
""
],
[
"O'Carroll",
"Michael",
""
]
] |
2509.03514
|
Robin Verstraten
|
Robin C. Verstraten, Ivo H. A. Knottnerus, Yu Chih Tseng, Alexander Urech, Tiago Santiago do Espirito Santo, Vinicius Zampronio, Florian Schreck, Robert J. C. Spreeuw, Cristiane Morais Smith
|
Control of single spin-flips in a Rydberg atomic fractal
|
Main article: 32 pages, 4 figures, 8 extended data figures. Supplementary Information: 11 pages, 5 supplementary figures
| null | null | null |
cond-mat.quant-gas cond-mat.dis-nn quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Rydberg atoms trapped by optical tweezers have emerged as a versatile platform to emulate lattices with different geometries, in which long-range interacting spins lead to fascinating phenomena, ranging from spin liquids to topological states of matter. Here, we show that when the lattice has a fractal geometry with Hausdorff dimension 1.58, additional surprises appear. The system is described by a transverse-field Ising model with long-range van der Waals interactions in a Sierpi\`nski gasket fractal. We investigate the problem theoretically using exact diagonalization, variational mean field, quantum Monte Carlo, and a graph-based numerical technique, SIM-GRAPH, which we developed. We find that in the quantum regime, the phase diagram exhibits phases in which the spins flip one-by-one. The theoretical results are in excellent agreement with experiments performed with single 88Sr atoms trapped by optical tweezers arranged in a fractal geometry. The magnetization and von Neumann entanglement entropy reveal several regimes in which single spin-flips are delocalized over many sites of one sublattice, thus allowing for an unprecedented control of a cascade of phase transitions in a manybody system. These results expand the possibilities of Rydberg atoms for quantum information processing and may have profound implications in quantum technology.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:54:47 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Verstraten",
"Robin C.",
""
],
[
"Knottnerus",
"Ivo H. A.",
""
],
[
"Tseng",
"Yu Chih",
""
],
[
"Urech",
"Alexander",
""
],
[
"Santo",
"Tiago Santiago do Espirito",
""
],
[
"Zampronio",
"Vinicius",
""
],
[
"Schreck",
"Florian",
""
],
[
"Spreeuw",
"Robert J. C.",
""
],
[
"Smith",
"Cristiane Morais",
""
]
] |
2509.03515
|
Yanlin Zhang
|
Yanlin Zhang, Sungyong Chung, Nachuan Li, Dana Monzer, Hani S. Mahmassani, Samer H. Hamdar, and Alireza Talebpour
|
Can the Waymo Open Motion Dataset Support Realistic Behavioral Modeling? A Validation Study with Naturalistic Trajectories
| null | null | null | null |
cs.RO cs.AI cs.LG cs.SY eess.SY stat.AP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Waymo Open Motion Dataset (WOMD) has become a popular resource for data-driven modeling of autonomous vehicles (AVs) behavior. However, its validity for behavioral analysis remains uncertain due to proprietary post-processing, the absence of error quantification, and the segmentation of trajectories into 20-second clips. This study examines whether WOMD accurately captures the dynamics and interactions observed in real-world AV operations. Leveraging an independently collected naturalistic dataset from Level 4 AV operations in Phoenix, Arizona (PHX), we perform comparative analyses across three representative urban driving scenarios: discharging at signalized intersections, car-following, and lane-changing behaviors. For the discharging analysis, headways are manually extracted from aerial video to ensure negligible measurement error. For the car-following and lane-changing cases, we apply the Simulation-Extrapolation (SIMEX) method to account for empirically estimated error in the PHX data and use Dynamic Time Warping (DTW) distances to quantify behavioral differences. Results across all scenarios consistently show that behavior in PHX falls outside the behavioral envelope of WOMD. Notably, WOMD underrepresents short headways and abrupt decelerations. These findings suggest that behavioral models calibrated solely on WOMD may systematically underestimate the variability, risk, and complexity of naturalistic driving. Caution is therefore warranted when using WOMD for behavior modeling without proper validation against independently collected data.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:56:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhang",
"Yanlin",
""
],
[
"Chung",
"Sungyong",
""
],
[
"Li",
"Nachuan",
""
],
[
"Monzer",
"Dana",
""
],
[
"Mahmassani",
"Hani S.",
""
],
[
"Hamdar",
"Samer H.",
""
],
[
"Talebpour",
"Alireza",
""
]
] |
2509.03516
|
Ouxiang Li
|
Ouxiang Li, Yuan Wang, Xinting Hu, Huijuan Huang, Rui Chen, Jiarong Ou, Xin Tao, Pengfei Wan, Fuli Feng
|
Easier Painting Than Thinking: Can Text-to-Image Models Set the Stage, but Not Direct the Play?
|
Project Page: https://t2i-corebench.github.io/
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Text-to-image (T2I) generation aims to synthesize images from textual prompts, which jointly specify what must be shown and imply what can be inferred, thereby corresponding to two core capabilities: composition and reasoning. However, with the emerging advances of T2I models in reasoning beyond composition, existing benchmarks reveal clear limitations in providing comprehensive evaluations across and within these capabilities. Meanwhile, these advances also enable models to handle more complex prompts, whereas current benchmarks remain limited to low scene density and simplified one-to-one reasoning. To address these limitations, we propose T2I-CoReBench, a comprehensive and complex benchmark that evaluates both composition and reasoning capabilities of T2I models. To ensure comprehensiveness, we structure composition around scene graph elements (instance, attribute, and relation) and reasoning around the philosophical framework of inference (deductive, inductive, and abductive), formulating a 12-dimensional evaluation taxonomy. To increase complexity, driven by the inherent complexities of real-world scenarios, we curate each prompt with high compositional density for composition and multi-step inference for reasoning. We also pair each prompt with a checklist that specifies individual yes/no questions to assess each intended element independently to facilitate fine-grained and reliable evaluation. In statistics, our benchmark comprises 1,080 challenging prompts and around 13,500 checklist questions. Experiments across 27 current T2I models reveal that their composition capability still remains limited in complex high-density scenarios, while the reasoning capability lags even further behind as a critical bottleneck, with all models struggling to infer implicit elements from prompts. Our project page: https://t2i-corebench.github.io/.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:58:12 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Ouxiang",
""
],
[
"Wang",
"Yuan",
""
],
[
"Hu",
"Xinting",
""
],
[
"Huang",
"Huijuan",
""
],
[
"Chen",
"Rui",
""
],
[
"Ou",
"Jiarong",
""
],
[
"Tao",
"Xin",
""
],
[
"Wan",
"Pengfei",
""
],
[
"Feng",
"Fuli",
""
]
] |
2509.03517
|
Oleksii Ivanytskyi
|
Christoph G\"artlein, Oleksii Ivanytskyi, Violetta Sagun and Il\'idio Lopes
|
Color-superconducting quarkyonic matter
|
21 pages, 11 figures, 2 tables
| null | null | null |
nucl-th astro-ph.HE hep-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
We explore the role of color superconductivity in quarkyonic matter under the conditions of color and electric neutrality at $\beta$- and strong equilibrium, as relevant for neutron stars. By explicitly incorporating the color-superconducting pairing gap into the phenomenological model of a smooth transition from hadron to quark matter, we extend the known quarkyonic framework to include this essential aspect relevant at high densities. The momentum dependence of the pairing gap, motivated by the running of the QCD coupling and introduced similarly to chiral quark models with nonlocal interaction, is a novel element of the model that is crucial for enabling the simultaneous onset of all color-flavor quark states in the presence of color superconductivity. While asymptotically conformal behavior of the present model is ensured by construction, we demonstrate that reaching the conformal limit in agreement with the predictions of perturbative QCD is provided by the proper momentum dependence of the thickness of the hadron shell in momentum space. We employ the flexible meta-modeling approach to nuclear matter, analyzing the structure of the hadron shell in momentum space and focusing on the effects of color superconductivity in quarkyonic matter. Similar to the effects induced by the onset of the quarkyonic phase, color superconductivity leads to stiffening of the equation of state of the NS matter. This causes a significant impact on observable properties of neutron stars, which are analyzed and compared to recent astrophysical and theoretical constraints. We argue that the developed model of color-superconducting quarkyonic matter provides a new, consistent tool for studying the scenario of smooth quark-hadron transition in NSs.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:59:33 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Gärtlein",
"Christoph",
""
],
[
"Ivanytskyi",
"Oleksii",
""
],
[
"Sagun",
"Violetta",
""
],
[
"Lopes",
"Ilídio",
""
]
] |
2509.03518
|
Haoran Huan
|
Haoran Huan, Mihir Prabhudesai, Mengning Wu, Shantanu Jaiswal, Deepak Pathak
|
Can LLMs Lie? Investigation beyond Hallucination
|
Website at https://llm-liar.github.io/
| null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large language models (LLMs) have demonstrated impressive capabilities across a variety of tasks, but their increasing autonomy in real-world applications raises concerns about their trustworthiness. While hallucinations-unintentional falsehoods-have been widely studied, the phenomenon of lying, where an LLM knowingly generates falsehoods to achieve an ulterior objective, remains underexplored. In this work, we systematically investigate the lying behavior of LLMs, differentiating it from hallucinations and testing it in practical scenarios. Through mechanistic interpretability techniques, we uncover the neural mechanisms underlying deception, employing logit lens analysis, causal interventions, and contrastive activation steering to identify and control deceptive behavior. We study real-world lying scenarios and introduce behavioral steering vectors that enable fine-grained manipulation of lying tendencies. Further, we explore the trade-offs between lying and end-task performance, establishing a Pareto frontier where dishonesty can enhance goal optimization. Our findings contribute to the broader discourse on AI ethics, shedding light on the risks and potential safeguards for deploying LLMs in high-stakes environments. Code and more illustrations are available at https://llm-liar.github.io/
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:59:45 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Huan",
"Haoran",
""
],
[
"Prabhudesai",
"Mihir",
""
],
[
"Wu",
"Mengning",
""
],
[
"Jaiswal",
"Shantanu",
""
],
[
"Pathak",
"Deepak",
""
]
] |
2509.03519
|
Ronan Hix
|
Ronan Hix, Lucia Armillotta, Eve Ostriker, Chang-Goo Kim
|
Dynamically Controlled Transport of GeV Cosmic Rays in Diverse Galactic Environments
|
30 pages, 10 figures, ApJ
| null | null | null |
astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
We study transport of GeV cosmic rays (CRs) in a set of high-resolution TIGRESS magnetohydrodynamic simulations of the star-forming interstellar medium (ISM). Our local disk patch models sample a wide range of gas surface densities, gravitational potentials, and star formation rates (SFRs), and include a spiral arm simulation. Our approach incorporates CR advection by the background gas, streaming along the magnetic field limited by the local ion Alfv\'en speed, and diffusion relative to the Alfv\'en wave frame, with the diffusion coefficient set by the balance between streaming-driven Alfv\'en wave excitation and damping mediated by local gas properties. We find that dynamical transport mechanisms (streaming and advection) are almost solely responsible for GeV CR transport in the extra-planar regions of galaxies, while diffusion along the magnetic field dominates within the primarily-neutral ISM of galactic disks. We develop a simple 1D predictive model for the CR pressure $P_\mathrm{c}$, dependent only on injected CR flux and gas parameters. We demonstrate that the CR transport efficiency increases with increasing SFR, and provide a fit for the CR feedback yield $\Upsilon_\mathrm{c}~\equiv~P_\mathrm{c}/\Sigma_\mathrm{SFR}$ as a function of $\Sigma_\mathrm{SFR}$, the SFR surface density. We analyze lateral CR transport within the galactic disk, showing that CRs propagate away from feedback regions in spiral arms into interarm regions by a combination of gas advection and field-aligned transport. Lastly, we develop an empirical subgrid model for the CR scattering rate that captures the impacts of the multiphase ISM on CR transport without the numerical burden of full simulations.
|
[
{
"version": "v1",
"created": "Wed, 03 Sep 2025 17:59:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Hix",
"Ronan",
""
],
[
"Armillotta",
"Lucia",
""
],
[
"Ostriker",
"Eve",
""
],
[
"Kim",
"Chang-Goo",
""
]
] |
cond-mat/0305480
|
Mario Renteria
|
J. M. Ramallo Lopez (1), M. Renteria (1 and 2), E. E. Miro (3), F. G. Requejo (1 and 4) and A. Traverse (2) ((1)Departamento de Fisica, FCE, Universidad Nacional de La Plata, La Plata, Argentina, (2) Laboratoire pour l'Utilisation du Rayonnement Electromagnetique (LURE), Universite Paris-Sud, Orsay, France, (3) INCAPE (FIQ, UNL, CONICET), Santa Fe, Argentina, (4) Materials Science Division, Lawrence Berkeley National Laboratory, Berkeley CA, USA)
|
Structure of Extremely Nanosized and Confined In-O Species in Ordered Porous Materials
|
4 pages, 5 postscript figures, REVTEX4, published in Physical Review Letters
|
Physical Review Letters 91, 108304 (2003)
|
10.1103/PhysRevLett.91.108304
| null |
cond-mat.mtrl-sci cond-mat.soft
| null |
Perturbed-angular correlation, x-ray absorption, and small-angle x-ray scattering spectroscopies were suitably combined to elucidate the local structure of highly diluted and dispersed InOx species confined in porous of ZSM5 zeolite. These novel approach allow us to determined the structure of extremely nanosized In-O species exchanged inside the 10-atom-ring channel of the zeolite, and to quantify the amount of In2O3 crystallites deposited onto the external zeolite surface.
|
[
{
"version": "v1",
"created": "Tue, 20 May 2003 22:15:06 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Apr 2007 06:01:23 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Lopez",
"J. M. Ramallo",
"",
"Departamento de Fisica, FCE, Universidad Nacional de La Plata, La Plata, Argentina"
],
[
"Renteria",
"M.",
"",
"1 and 2"
],
[
"Miro",
"E. E.",
"",
"INCAPE"
],
[
"Requejo",
"F. G.",
"",
"1 and 4"
],
[
"Traverse",
"A.",
"",
"Laboratoire pour l'Utilisation du Rayonnement Electromagnetique"
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.