id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.01103
|
Stephen Watt
|
Arthur C. Norman, Stephen M. Watt
|
Semi-Centennial REDUCE
| null | null | null | null |
cs.SC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a version of the REDUCE computer algebra system as it was in the early 1970s. We show how this historical version of REDUCE may be built and run in very modest present-day environments and outline some of its capabilities.
|
[
{
"version": "v1",
"created": "Fri, 02 May 2025 08:31:28 GMT"
},
{
"version": "v2",
"created": "Fri, 08 Aug 2025 15:58:06 GMT"
},
{
"version": "v3",
"created": "Tue, 02 Sep 2025 19:19:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Norman",
"Arthur C.",
""
],
[
"Watt",
"Stephen M.",
""
]
] |
2505.01530
|
Muhammad Tayyab Khan
|
Muhammad Tayyab Khan, Zane Yong, Lequn Chen, Jun Ming Tan, Wenhe Feng, and Seung Ki Moon
|
Automated Parsing of Engineering Drawings for Structured Information Extraction Using a Fine-tuned Document Understanding Transformer
|
This manuscript has been accepted for publication at IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)
| null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Accurate extraction of key information from 2D engineering drawings is crucial for high-precision manufacturing. Manual extraction is slow and labor-intensive, while traditional Optical Character Recognition (OCR) techniques often struggle with complex layouts and overlapping symbols, resulting in unstructured outputs. To address these challenges, this paper proposes a novel hybrid deep learning framework for structured information extraction by integrating an Oriented Bounding Box (OBB) detection model with a transformer-based document parsing model (Donut). An in-house annotated dataset is used to train YOLOv11 for detecting nine key categories: Geometric Dimensioning and Tolerancing (GD&T), General Tolerances, Measures, Materials, Notes, Radii, Surface Roughness, Threads, and Title Blocks. Detected OBBs are cropped into images and labeled to fine-tune Donut for structured JSON output. Fine-tuning strategies include a single model trained across all categories and category-specific models. Results show that the single model consistently outperforms category-specific ones across all evaluation metrics, achieving higher precision (94.77% for GD&T), recall (100% for most categories), and F1 score (97.3%), while reducing hallucinations (5.23%). The proposed framework improves accuracy, reduces manual effort, and supports scalable deployment in precision-driven industries.
|
[
{
"version": "v1",
"created": "Fri, 02 May 2025 18:33:21 GMT"
},
{
"version": "v2",
"created": "Sat, 30 Aug 2025 04:30:08 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 03:04:00 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Khan",
"Muhammad Tayyab",
""
],
[
"Yong",
"Zane",
""
],
[
"Chen",
"Lequn",
""
],
[
"Tan",
"Jun Ming",
""
],
[
"Feng",
"Wenhe",
""
],
[
"Moon",
"Seung Ki",
""
]
] |
2505.01579
|
Herbert Neuberger
|
Herbert Neuberger
|
On the momentum of gluons in Lattice Gauge Theory (LGT)
|
5 pages
| null |
10.1142/S0217751X25501234
| null |
hep-lat hep-ph hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Attempts to improve LGT simulation algorithms by Fourier space preconditioning have been handicapped by the gauge dependence of momenta, familiar from perturbation theory. The continuum theory has a gauge invariant energy-momentum density, indicating this to be a fake obstacle. While perturbation-theory momenta are evidently wrong for the task, momentum-transfer carried by gluons is physical. Simulations using these momenta may become practical in the near future thanks to recent progress in generative models, stochastic and/or deterministic.
|
[
{
"version": "v1",
"created": "Fri, 02 May 2025 20:45:27 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Neuberger",
"Herbert",
""
]
] |
2505.02273
|
Rimon Melamed
|
Rimon Melamed and Lucas H. McCabe and H. Howie Huang
|
Demystifying optimized prompts in language models
|
EMNLP 2025 Main
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Modern language models (LMs) are not robust to out-of-distribution inputs. Machine generated (``optimized'') prompts can be used to modulate LM outputs and induce specific behaviors while appearing completely uninterpretable. In this work, we investigate the composition of optimized prompts, as well as the mechanisms by which LMs parse and build predictions from optimized prompts. We find that optimized prompts primarily consist of punctuation and noun tokens which are more rare in the training data. Internally, optimized prompts are clearly distinguishable from natural language counterparts based on sparse subsets of the model's activations. Across various families of instruction-tuned models, optimized prompts follow a similar path in how their representations form through the network.
|
[
{
"version": "v1",
"created": "Sun, 04 May 2025 22:04:14 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:25:16 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Melamed",
"Rimon",
""
],
[
"McCabe",
"Lucas H.",
""
],
[
"Huang",
"H. Howie",
""
]
] |
2505.02476
|
Hubert Padusinski
|
Hubert Padusinski, Christian Steinhauser, Christian Scherl, Julian Gaal, Jacob Langner
|
Point Cloud Recombination: Systematic Real Data Augmentation Using Robotic Targets for LiDAR Perception Validation
|
Pre-print for IEEE IAVVC 2025
| null | null | null |
cs.RO cs.CV eess.IV
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The validation of LiDAR-based perception of intelligent mobile systems operating in open-world applications remains a challenge due to the variability of real environmental conditions. Virtual simulations allow the generation of arbitrary scenes under controlled conditions but lack physical sensor characteristics, such as intensity responses or material-dependent effects. In contrast, real-world data offers true sensor realism but provides less control over influencing factors, hindering sufficient validation. Existing approaches address this problem with augmentation of real-world point cloud data by transferring objects between scenes. However, these methods do not consider validation and remain limited in controllability because they rely on empirical data. We solve these limitations by proposing Point Cloud Recombination, which systematically augments captured point cloud scenes by integrating point clouds acquired from physical target objects measured in controlled laboratory environments. Thus enabling the creation of vast amounts and varieties of repeatable, physically accurate test scenes with respect to phenomena-aware occlusions with registered 3D meshes. Using the Ouster OS1-128 Rev7 sensor, we demonstrate the augmentation of real-world urban and rural scenes with humanoid targets featuring varied clothing and poses, for repeatable positioning. We show that the recombined scenes closely match real sensor outputs, enabling targeted testing, scalable failure analysis, and improved system safety. By providing controlled yet sensor-realistic data, our method enables trustworthy conclusions about the limitations of specific sensors in compound with their algorithms, e.g., object detection.
|
[
{
"version": "v1",
"created": "Mon, 05 May 2025 09:00:16 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 13:01:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Padusinski",
"Hubert",
""
],
[
"Steinhauser",
"Christian",
""
],
[
"Scherl",
"Christian",
""
],
[
"Gaal",
"Julian",
""
],
[
"Langner",
"Jacob",
""
]
] |
2505.03854
|
Ayumu Inoue
|
Ayumu Inoue
|
A counter example of Yashiro's theorem on pseudo-cycles of a surface knot
|
The contents of this paper was integrated into the paper "The lengths of 3-cocycles of the 7-dihedral and the octahedral quandles" (arXiv:2508.14612) as Appendix A
| null | null | null |
math.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We point out that the claim in Theorem 1.1 of Yashiro's paper ``Pseudo-cycles of surface-knots'' is not true, giving a counter example.
|
[
{
"version": "v1",
"created": "Tue, 06 May 2025 07:25:24 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 04:20:15 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Inoue",
"Ayumu",
""
]
] |
2505.03911
|
Hans Hohenfeld
|
Hans Hohenfeld, Marius Beuerle, Elie Mounzer
|
Explaining Anomalies with Tensor Networks
|
6 pages, 3 figures, Accepted for publication at IEEE QAI 2025
| null | null | null |
cs.LG quant-ph
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Tensor networks, a class of variational quantum many-body wave functions have attracted considerable research interest across many disciplines, including classical machine learning. Recently, Aizpurua et al. demonstrated explainable anomaly detection with matrix product states on a discrete-valued cyber-security task, using quantum-inspired methods to gain insight into the learned model and detected anomalies. Here, we extend this framework to real-valued data domains. We furthermore introduce tree tensor networks for the task of explainable anomaly detection. We demonstrate these methods with three benchmark problems, show adequate predictive performance compared to several baseline models and both tensor network architectures' ability to explain anomalous samples. We thereby extend the application of tensor networks to a broader class of potential problems and open a pathway for future extensions to more complex tensor network architectures.
|
[
{
"version": "v1",
"created": "Tue, 06 May 2025 18:35:05 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 01:04:35 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Hohenfeld",
"Hans",
""
],
[
"Beuerle",
"Marius",
""
],
[
"Mounzer",
"Elie",
""
]
] |
2505.04255
|
Nay Klaimi
|
Nay Klaimi (IETR, INSA Rennes), Amira Bedoui (IETR, INSA Rennes), Cl\'ement Elvira (IETR), Philippe Mary (INSA Rennes, IETR), Luc Le Magoarou (INSA Rennes, IETR)
|
Model-based learning for joint channel estimationand hybrid MIMO precoding
| null | null | null | null |
eess.SP cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Hybrid precoding is a key ingredient of cost-effective massive multiple-input multiple-output transceivers. However, setting jointly digital and analog precoders to optimally serve multiple users is a difficult optimization problem. Moreover, it relies heavily on precise knowledge of the channels, which is difficult to obtain, especially when considering realistic systems comprising hardware impairments. In this paper, a joint channel estimation and hybrid precoding method is proposed, which consists in an end-to-end architecture taking received pilots as inputs and outputting pre-coders. The resulting neural network is fully model-based, making it lightweight and interpretable with very few learnable parameters. The channel estimation step is performed using the unfolded matching pursuit algorithm, accounting for imperfect knowledge of the antenna system, while the precoding step is done via unfolded projected gradient ascent. The great potential of the proposed method is empirically demonstrated on realistic synthetic channels.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 09:00:34 GMT"
},
{
"version": "v2",
"created": "Tue, 10 Jun 2025 09:03:16 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 07:16:01 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Klaimi",
"Nay",
"",
"IETR, INSA Rennes"
],
[
"Bedoui",
"Amira",
"",
"IETR, INSA Rennes"
],
[
"Elvira",
"Clément",
"",
"IETR"
],
[
"Mary",
"Philippe",
"",
"INSA Rennes, IETR"
],
[
"Magoarou",
"Luc Le",
"",
"INSA Rennes, IETR"
]
] |
2505.04283
|
Dingyuan Liu
|
Felix Christian Clemen, Adrian Dumitrescu, Dingyuan Liu
|
On multiplicities of interpoint distances
|
11 pages, 4 figures
| null | null | null |
math.CO cs.DM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Given a set $X\subseteq\mathbb{R}^2$ of $n$ points and a distance $d>0$, the multiplicity of $d$ is the number of times the distance $d$ appears between points in $X$. Let $a_1(X) \geq a_2(X) \geq \cdots \geq a_m(X)$ denote the multiplicities of the $m$ distances determined by $X$ and let $a(X)=\left(a_1(X),\dots,a_m(X)\right)$. In this paper, we study several questions from Erd\H{o}s's time regarding distance multiplicities. Among other results, we show that:
(1) If $X$ is convex or ``not too convex'', then there exists a distance other than the diameter that has multiplicity at most $n$.
(2) There exists a set $X \subseteq \mathbb{R}^2$ of $n$ points, such that many distances occur with high multiplicity. In particular, at least $n^{\Omega(1/\log\log{n})}$ distances have superlinear multiplicity in $n$.
(3) For any (not necessarily fixed) integer $1\leq k\leq\log{n}$, there exists $X\subseteq\mathbb{R}^2$ of $n$ points, such that the difference between the $k^{\text{th}}$ and $(k+1)^{\text{th}}$ largest multiplicities is at least $\Omega(\frac{n\log{n}}{k})$. Moreover, the distances in $X$ with the largest $k$ multiplicities can be prescribed.
(4) For every $n\in\mathbb{N}$, there exists $X\subseteq\mathbb{R}^2$ of $n$ points, not all collinear or cocircular, such that $a(X)= (n-1,n-2,\ldots,1)$. There also exists $Y\subseteq\mathbb{R}^2$ of $n$ points with pairwise distinct distance multiplicities and $a(Y) \neq (n-1,n-2,\ldots,1)$.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 09:37:27 GMT"
},
{
"version": "v2",
"created": "Fri, 09 May 2025 09:30:27 GMT"
},
{
"version": "v3",
"created": "Sat, 26 Jul 2025 07:33:27 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 09:30:27 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Clemen",
"Felix Christian",
""
],
[
"Dumitrescu",
"Adrian",
""
],
[
"Liu",
"Dingyuan",
""
]
] |
2505.04447
|
Dibya Chakraborty
|
Dibya Chakraborty, Rudnei O. Ramos
|
Warming up the Fibres
|
30 pages, 4 figure, 3 tables; This matches the published version in JHEP
| null |
10.1007/jhep09(2025)020
| null |
hep-th astro-ph.CO gr-qc hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Fibre inflationary models are constructed in type-IIB string flux compactification. These models have been shown to be in agreement with the cosmological observations under appropriate choices of parameters, which originate from their string theory construction. In the present work, we embed such models, originally studied in the cold inflation picture, in the context of warm inflation. We study the viability of different fibre inflation potentials in both strong and weak dissipative regime of warm inflation. In fibre inflation, the inflaton is a four-dimensional complex manifold -- a fibre of a $K3$ fibred Calabi-Yau. The potential in this case is generated by an interplay of various perturbative and non-perturbative corrections. The former type of corrections consists of leading-order $\alpha'^3-$ term, higher derivative $F^4-$ correction, and various string loop corrections of KK, log-loop and winding type. Depending on the balance between several corrections, we present four different fibre inflationary potentials and show that the warm inflationary pictures for all of them can successfully fall in the viable window from both Planck and recent Atacama Cosmology Telescope (ACT) data. We show that warm inflation makes it possible to extend the range of parameters of applicability of these models. Our results also indicate that with the help of a large dissipation ratio, one can achieve a sub-Planckian field excursion, although, this runs on the possibility of moving away from the perturbative control of low-energy four-dimensional supergravity theory.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 14:17:31 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 07:54:32 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chakraborty",
"Dibya",
""
],
[
"Ramos",
"Rudnei O.",
""
]
] |
2505.04593
|
William Blair
|
William L. Blair
|
Bicomplex Hardy Classes of Solutions to Higher-Order Vekua Equations
|
Changes from last version: 1. Many typographical and notational errors were corrected. 2. The definition of the integral operator in Theorem 2.17 was corrected. 3. Aspects of the referenced background material have been condensed and/or clarified. Other minor improvements throughout. The results of all new theorems are unchanged outside of corrections necessary due to typographical errors
| null | null | null |
math.CV math.AP math.CA math.FA
|
http://creativecommons.org/licenses/by/4.0/
|
We extend representation formulas for functions in Hardy classes of solutions to higher-order iterated Vekua equations to Hardy classes of bicomplex-valued functions that solve a bicomplex version of the Vekua equation or its higher-order generalizations. Using these representations, we show that functions in these bicomplex-valued Hardy classes have nontangential boundary values and boundary values in the sense of distributions.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 17:35:47 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 20:45:41 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Blair",
"William L.",
""
]
] |
2505.04692
|
Lorenzo Stefanuto
|
Valentina De Romeri, Fiorenza Donato, David Maurin, Lorenzo Stefanuto, Agnese Tolino
|
Antinuclei from Primordial Black Holes
|
22 pages, 6 figures, minor corrections in version 2
|
Phys. Rev. D 112 (2025) 2, 023003
|
10.1103/r6w9-vr6b
| null |
hep-ph astro-ph.HE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Light primordial black holes (PBHs) may have originated in the early Universe, and could contribute to the dark matter in the Universe. Their Hawking evaporation into particles could eventually lead to the production of antinuclei, which propagate and arrive at Earth as cosmic rays with a flux peaked at GeV energies. We revisit here the antiproton and antideuteron signatures from PBH evaporation, relying on a lognormal PBH mass distribution, state-of-the-art propagation models, and an improved coalescence model for fusion into antideuterons. Our predictions are then compared with AMS-02 data on the antiproton flux. We find that the AMS-02 antiproton data severely constrain the Galactic PBH density, setting bounds that depend significantly on the parameters of the lognormal mass distribution, and that are comparable to or slightly stronger than bounds set from diverse messengers. We also discuss prospects for future detection of antideuterons. Given the bounds from AMS-02 antiproton data, we predict that if antideuterons were to be measured by AMS-02 or GAPS, since the secondary contribution is subdominant, they would clearly be a signal of new physics, only part of which could, however, be explained by PBH evaporation.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 18:00:02 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 13:24:03 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"De Romeri",
"Valentina",
""
],
[
"Donato",
"Fiorenza",
""
],
[
"Maurin",
"David",
""
],
[
"Stefanuto",
"Lorenzo",
""
],
[
"Tolino",
"Agnese",
""
]
] |
2505.04819
|
Tristan Aur\'egan
|
Tristan Aur\'egan and Luc Deike
|
Surface bubble lifetime in the presence of a turbulent air flow, and the effect of surface layer renewal
| null | null | null | null |
physics.flu-dyn
|
http://creativecommons.org/licenses/by/4.0/
|
Surface bubbles in the ocean are critical in moderating several fluxes between the atmosphere and the ocean. In this paper, we experimentally investigate the drainage and lifetime of surface bubbles in solutions containing surfactants and salts, subjected to turbulence in the air surrounding them modelling the wind above the ocean. We carefully construct a setup allowing us to repeatably measure the mean lifetime of a series of surface bubbles, while varying the solution and the wind speed or humidity of the air. To that end, we show that renewing the surface layer is critical to avoid a change of the physical properties of the interface. We show that the drainage of the bubbles is well modelled by taking into account the outwards viscous flow and convective evaporation. The mean lifetime of surface bubbles in solutions containing no salt is controlled by evaporation and independent on surfactant concentration. When salt is added, the same scaling is valid only at high surfactant concentrations. At low concentrations, the lifetime is always smaller and independent of wind speed, owing to the presence of impurities triggering a thick bursting event. When the mean lifetime is controlled by evaporation, the probability density of lifetime is very narrow around its mean, while when impurities are present, a broad distribution is observed.
|
[
{
"version": "v1",
"created": "Wed, 07 May 2025 21:45:26 GMT"
},
{
"version": "v2",
"created": "Mon, 12 May 2025 13:41:47 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 15:37:26 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Aurégan",
"Tristan",
""
],
[
"Deike",
"Luc",
""
]
] |
2505.04909
|
Yubo Ma
|
Yubo Ma, Songtao Zheng, Yang Zhang
|
Thermodynamic relation on black branes with arbitrary cosmological constant
|
7pages,0 figures
| null | null | null |
hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We investigated the Goon-Penco(GP) relationship in (n+1)-dimensional black branes with an arbitrary cosmological constant. Our analysis revealed that the GP relation preserved its form in four-dimensional and (n+1)-dimensional spacetimes, demonstrating its universal behavior with respect to dimensionality. Furthermore, we established that the GP relation exhibits universality across all states of the black hole, including those associated with the event horizon and the cosmological horizon. These findings confirm that the GP relationship remains valid for (n+1)-dimensional black holes and black branes with an arbitrary cosmological constant, independent of the coordinate system employed.
|
[
{
"version": "v1",
"created": "Thu, 08 May 2025 02:55:46 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 02:48:40 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 08:57:37 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ma",
"Yubo",
""
],
[
"Zheng",
"Songtao",
""
],
[
"Zhang",
"Yang",
""
]
] |
2505.05124
|
Anton Baykalov
|
Anton A. Baykalov, Alice Devillers, Cheryl E. Praeger
|
Proper partial linear spaces affording imprimitive rank 3 automorphism groups
|
50 pages, address of authors updated, some typos fixed
| null | null | null |
math.GR math.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A partial linear space is a point--line incidence structure such that each line is incident with at least two points and each pair of points is incident with at most one line. It is said to be proper if there exists at least one non-collinear point pair, and at least one line incident with more than two points. The highest degree of symmetry for a proper partial linear space occurs when the automorphism group $G$ is transitive on ordered pairs of collinear points, and on ordered pairs of non-collinear points, that is to say, $G$ is a transitive rank $3$ group on the points. While the primitive rank 3 partial linear spaces are essentially classified, we present the first substantial classification of a family of imprimitive rank $3$ examples.
We classify all imprimitive rank $3$ proper partial linear spaces such that the rank $3$ group is semiprimitive. In particular, this includes all partial linear spaces with a rank 3 imprimitive automorphism group that is innately transitive or quasiprimitive. We construct several infinite families of examples and ten individual examples. The examples in the infinite families admit a rank $3$ action of a linear or unitary group, and to our knowledge most of these examples have not appeared before in the literature.
|
[
{
"version": "v1",
"created": "Thu, 08 May 2025 10:56:58 GMT"
},
{
"version": "v2",
"created": "Fri, 09 May 2025 16:48:06 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 09:13:02 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Baykalov",
"Anton A.",
""
],
[
"Devillers",
"Alice",
""
],
[
"Praeger",
"Cheryl E.",
""
]
] |
2505.05225
|
Mengze Hong
|
Mengze Hong, Wailing Ng, Chen Jason Zhang, Di Jiang
|
QualBench: Benchmarking Chinese LLMs with Localized Professional Qualifications for Vertical Domain Evaluation
|
Accepted by EMNLP 2025 Main Conference. Homepage: https://github.com/mengze-hong/QualBench
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The rapid advancement of Chinese LLMs underscores the need for vertical-domain evaluations to ensure reliable applications. However, existing benchmarks often lack domain coverage and provide limited insights into the Chinese working context. Leveraging qualification exams as a unified framework for expertise evaluation, we introduce QualBench, the first multi-domain Chinese QA benchmark dedicated to localized assessment of Chinese LLMs. The dataset includes over 17,000 questions across six vertical domains, drawn from 24 Chinese qualifications to align with national policies and professional standards. Results reveal an interesting pattern of Chinese LLMs consistently surpassing non-Chinese models, with the Qwen2.5 model outperforming the more advanced GPT-4o, emphasizing the value of localized domain knowledge in meeting qualification requirements. The average accuracy of 53.98% reveals the current gaps in domain coverage within model capabilities. Furthermore, we identify performance degradation caused by LLM crowdsourcing, assess data contamination, and illustrate the effectiveness of prompt engineering and model fine-tuning, suggesting opportunities for future improvements through multi-domain RAG and Federated Learning.
|
[
{
"version": "v1",
"created": "Thu, 08 May 2025 13:16:49 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 11:11:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Hong",
"Mengze",
""
],
[
"Ng",
"Wailing",
""
],
[
"Zhang",
"Chen Jason",
""
],
[
"Jiang",
"Di",
""
]
] |
2505.05272
|
Andr\'e Milagre
|
Andr\'e Milagre, Darius Jur\v{c}iukonis, Lu\'is Lavoura
|
Vacuum stability conditions for new $SU(2)$ multiplets
|
33 pages, 2 figures; v2: added references and footnotes, corrected the Acknowledgments
|
Prog. Theor. Exp. Phys. 2025, 093B01 (2025)
|
10.1093/ptep/ptaf105
| null |
hep-ph hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We consider the addition to the Standard Model of a scalar $SU(2)$ multiplet $\Delta_n$ with dimension $n$ going from $1$ to $6$. The multiplet $\Delta_n$ is assumed to have null vacuum expectation value and an arbitrary (free) hypercharge. We determine the shape of the phase space for the new terms that appear in the scalar potential (SP); we observe in particular that, in the case of a 6-plet, the phase space is slightly concave along one of its boundaries. We determine the bounded-from-below and vacuum stability conditions on the SP for each value of $n$.
|
[
{
"version": "v1",
"created": "Thu, 08 May 2025 14:18:09 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:09:47 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Milagre",
"André",
""
],
[
"Jurčiukonis",
"Darius",
""
],
[
"Lavoura",
"Luís",
""
]
] |
2505.05755
|
Dhruvesh Patel
|
Dhruvesh Patel, Aishwarya Sahoo, Avinash Amballa, Tahira Naseem, Tim G. J. Rudner, Andrew McCallum
|
Insertion Language Models: Sequence Generation with Arbitrary-Position Insertions
|
Additional related work. Code available at: https://dhruveshp.com/projects/ilm
| null | null | null |
cs.CL cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Autoregressive models (ARMs), which predict subsequent tokens one-by-one ``from left to right,'' have achieved significant success across a wide range of sequence generation tasks. However, they struggle to accurately represent sequences that require satisfying sophisticated constraints or whose sequential dependencies are better addressed by out-of-order generation. Masked Diffusion Models (MDMs) address some of these limitations, but the process of unmasking multiple tokens simultaneously in MDMs can introduce incoherences, and MDMs cannot handle arbitrary infilling constraints when the number of tokens to be filled in is not known in advance. In this work, we introduce Insertion Language Models (ILMs), which learn to insert tokens at arbitrary positions in a sequence -- that is, they select jointly both the position and the vocabulary element to be inserted. By inserting tokens one at a time, ILMs can represent strong dependencies between tokens, and their ability to generate sequences in arbitrary order allows them to accurately model sequences where token dependencies do not follow a left-to-right sequential structure. To train ILMs, we propose a tailored network parameterization and use a simple denoising objective. Our empirical evaluation demonstrates that ILMs outperform both ARMs and MDMs on common planning tasks. Furthermore, we show that ILMs outperform MDMs and perform on par with ARMs in an unconditional text generation task while offering greater flexibility than MDMs in arbitrary-length text infilling. The code is available at: https://dhruveshp.com/projects/ilm .
|
[
{
"version": "v1",
"created": "Fri, 09 May 2025 03:29:15 GMT"
},
{
"version": "v2",
"created": "Fri, 16 May 2025 02:40:07 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 16:59:44 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Patel",
"Dhruvesh",
""
],
[
"Sahoo",
"Aishwarya",
""
],
[
"Amballa",
"Avinash",
""
],
[
"Naseem",
"Tahira",
""
],
[
"Rudner",
"Tim G. J.",
""
],
[
"McCallum",
"Andrew",
""
]
] |
2505.07478
|
Rui Li
|
Rui Li
|
Low-energy effective Hamiltonian and end states of an inverted HgTe nanowire
|
6 pages, 4 figure, published version
|
Physica E 174 (2025) 116354
|
10.1016/j.physe.2025.116354
| null |
cond-mat.mes-hall
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The band inversion transition in a cylindrical HgTe nanowire is inducible via varying the nanowire radius. Here we derive the low-energy effective Hamiltonian describing the band structure of the HgTe nanowire close to the fundamental band gap. Because both the $E_{1}$ and $H_{1}$ subbands have quadratic dependence on $k_{z}$ when the gap closes, we need to consider at least three subbands, i.e., the $E_{1}$, $H_{1}$, and $H_{2}$ subbands, in building the effective Hamiltonian. The resulting effective Hamiltonian is block diagonal and each block is a $3\times3$ matrix. End states are found in the inverted regime when we solve the effective Hamiltonian with open boundary condition.
|
[
{
"version": "v1",
"created": "Mon, 12 May 2025 12:10:59 GMT"
},
{
"version": "v2",
"created": "Sun, 31 Aug 2025 01:56:47 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Rui",
""
]
] |
2505.08068
|
Ye-Ling Zhou
|
Gao-Xiang Fang, Zhi-Wei Wang and Ye-Ling Zhou
|
Asymptotic grand unification in SO(10) with one extra dimension
|
33 pages, 4 figures, 5 tables, refs added, published version
|
JHEP 08 (2025) 042
|
10.1007/JHEP08(2025)042
| null |
hep-ph hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Asymptotic grand unification provides an alternative approach to gradually unify gauge couplings in the UV limit, where they reach a non-trivial UV fixed point. Using an economical and realistic particle content setup, we demonstrate that asymptotic grand unification can be achieved in a 5D SO(10) model with one extra dimension. The top, bottom and tau masses are split, and the smallness of the neutrino mass is explained via inverse seesaw. One intermediate scale, the Pati-Salam symmetry breaking scale, is included below the compactification scale. Due to the absence of large-dimensional Higgs representations, gauge couplings exhibit asymptotic safety and are thus asymptotically unified, regardless of their initial values. In contrast, Yukawa couplings can achieve asymptotic freedom if the negative gauge contributions dominate over the positive Yukawa terms, requiring exact unification at the compactification scale. The widely-used 126-dimensional Higgs is not recommended in this 5D asymptotic SO(10) GUT, as it tends to drive the gauge beta function positive, compromising asymptotic safety.
|
[
{
"version": "v1",
"created": "Mon, 12 May 2025 21:08:37 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 00:52:48 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fang",
"Gao-Xiang",
""
],
[
"Wang",
"Zhi-Wei",
""
],
[
"Zhou",
"Ye-Ling",
""
]
] |
2505.08398
|
N. Asger Mortensen
|
A. J. Chaves, Line Jelver, D. R. da Costa, Joel D. Cox, N. Asger Mortensen, Nuno M. R. Peres
|
Nonlocal electrodynamics of two-dimensional anisotropic magneto-plasmons
| null | null |
10.1515/nanoph-2025-0233
| null |
cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
We present a hydrodynamic model, grounded in Madelung's formalism, to describe collective electronic motion in anisotropic materials. This model incorporates nonlocal contributions from the Thomas-Fermi quantum pressure and quantum effects arising from the Bohm potential. We derive analytical expressions for the magnetoplasmon dispersion and nonlocal optical conductivity. To demonstrate the applicability of the model, we examine electrons in the conduction band of monolayer phosphorene, an exemplary anisotropic two-dimensional electron gas. The dispersion of plasmons derived from our hydrodynamic approach is closely aligned with that predicted by ab~initio calculations. Then, we use our model to analyze few-layer black phosphorus, whose measured infrared optical response is hyperbolic. Our results reveal that the incorporation of nonlocal and quantum effects in the optical conductivity prevents black phosphorus from supporting hyperbolic surface plasmon polaritons. We further demonstrate that the predicted wavefront generated by an electric dipole exhibits a significant difference between the local and nonlocal descriptions for the optical conductivity. This study underscores the necessity of moving beyond local approximations when investigating anisotropic systems capable of hosting strongly confined plasmon-polaritons.
|
[
{
"version": "v1",
"created": "Tue, 13 May 2025 09:53:43 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chaves",
"A. J.",
""
],
[
"Jelver",
"Line",
""
],
[
"da Costa",
"D. R.",
""
],
[
"Cox",
"Joel D.",
""
],
[
"Mortensen",
"N. Asger",
""
],
[
"Peres",
"Nuno M. R.",
""
]
] |
2505.08957
|
Ying Feng
|
Ying Feng, Piotr Indyk
|
Even Faster Algorithm for the Chamfer Distance
|
Simplified Section 4
| null | null | null |
cs.CG cs.DS
|
http://creativecommons.org/licenses/by/4.0/
|
For two d-dimensional point sets A, B of size up to n, the Chamfer distance from A to B is defined as CH(A,B) = \sum_{a \in A} \min_{b \in B} \|a-b\|. The Chamfer distance is a widely used measure for quantifying dissimilarity between sets of points, used in many machine learning and computer vision applications. A recent work of Bakshi et al, NeuriPS'23, gave the first near-linear time (1+eps)-approximate algorithm, with a running time of O(ndlog(n)/eps^2). In this paper we improve the running time further, to O(nd(loglog(n)+log(1/eps))/eps^2). When eps is a constant, this reduces the gap between the upper bound and the trivial Omega(dn) lower bound significantly, from O(log n) to O(loglog n).
|
[
{
"version": "v1",
"created": "Tue, 13 May 2025 20:49:04 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 22:27:54 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Feng",
"Ying",
""
],
[
"Indyk",
"Piotr",
""
]
] |
2505.09016
|
Riwa Karam
|
Riwa Karam, Ruoyu Lin, Brooks A. Butler, Magnus Egerstedt
|
Resource Allocation with Multi-Team Collaboration Based on Hamilton's Rule
| null | null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by/4.0/
|
This paper presents a multi-team collaboration strategy based on Hamilton's rule from ecology that facilitates resource allocation among multiple teams, where agents are considered as shared resource among all teams that must be allocated appropriately. We construct an algorithmic framework that allows teams to make bids for agents that consider the costs and benefits of transferring agents while also considering relative mission importance for each team. This framework is applied to a multi-team coverage control mission to demonstrate its effectiveness. It is shown that the necessary criteria of a mission evaluation function are met by framing it as a function of the locational coverage cost of each team with respect to agent gain and loss, and these results are illustrated through simulations.
|
[
{
"version": "v1",
"created": "Tue, 13 May 2025 23:11:30 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 03:48:54 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Karam",
"Riwa",
""
],
[
"Lin",
"Ruoyu",
""
],
[
"Butler",
"Brooks A.",
""
],
[
"Egerstedt",
"Magnus",
""
]
] |
2505.09360
|
Jia Long Chen
|
Jia Long Chen and Xiao-Yu Yan
|
Spectral properties of a class of Sierpinski-type Moran measures on $ \mathbb{R}^n $
| null | null | null | null |
math.CA math.DS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Let the infinite convolutions \begin{equation*}
\mu_{\{R_{k}\},\{D_{k}\}}=\delta_{R_{1}^{-1}D_{1}}*\delta_{R_{1}^{-1}R_{2}^{-1}D_{2}}*\delta_{R_{1}^{-1}R_{2}^{-1}R_{3}^{-1}D_{3}}*\dotsi \end{equation*}
be generated by the sequence of pairs \(\{ (R_k,D_k) \}_{k=1}^{\infty} \), where $ R_k\in M_n(\mathbb{Z})$ is an expanding integer matric, $D_k$ is a finite integer digit sets that satisfies the following two conditions: (i). \( \# D_k = m \) and \( m>2 \) is a prime; (ii). \( \{x: \sum_{d\in D_{k}}e^{2\pi i\langle d,x \rangle}=0\} =\cup_{i=1}^{\phi(k)}\cup_{j=1}^{m-1}(\frac{j}{m}\nu_{k,i}+\mathbb{Z}^{n}) \) for some \( \nu_{k,i} \in \{ (l_1, \cdots, l_n)^t : l_i \in [1, m-1] \cap \mathbb{Z}, 1\leq i\leq n \} \). In this paper, we study the spectrality of $\mu_{\{R_{k}\},\{D_{k}\}}$, and some necessary and sufficient conditions for \( L^{2}(\mu_{\{R_{k}\},\{D_{k}\}}) \) to have an orthogonal exponential function basis are established. Finally, we discuss the explanations and applications of our results.
|
[
{
"version": "v1",
"created": "Wed, 14 May 2025 13:09:25 GMT"
},
{
"version": "v2",
"created": "Mon, 01 Sep 2025 12:17:02 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 01:40:02 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chen",
"Jia Long",
""
],
[
"Yan",
"Xiao-Yu",
""
]
] |
2505.09728
|
Zhaoxuan Liu
|
Zhaoxuan Liu, John D. Silverman, Emanuele Daddi, Boris S. Kalita, Annagrazia Puglisi, Qinyue Fei, Alvio Renzini, Daichi Kashino, Francesco Valentino, Jeyhan S. Kartaltepe, Daizhong Liu, Pablo G. P\'erez-Gonz\'alez, Jed McKinney, Caitlin M. Casey, Xuheng Ding, Andreas Faisst, Maximilien Franco, Darshan Kakkad, Anton M. Koekemoer, Erini Lambrides, Steven Gillman, Ghassem Gozaliasl, Henry Joy McCracken, Jason Rhodes, Brant E. Robertson, Giulia Rodighiero, Wiphu Rujopakarn, Tomoko L. Suzuki, Takumi S. Tanaka, Brittany N. Vanderhoof, Aswin P. Vijayan, Olivia R. Cooper, Aidan Kaminsky, Georgios E. Magdis and Namrata Roy
|
A PAH deficit in the starburst core of a distant spiral galaxy
|
13 pages (including appendix), 8 figures. Accepted for publication in MNRAS
| null |
10.1093/mnras/staf1248
| null |
astro-ph.GA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present high-resolution and spatially-matched observations with JWST and ALMA of a starburst galaxy (PACS-830) at $z=1.46$. The NIRCam observations mainly trace the stellar light while the CO ($J$=5--4) observations map the dense molecular gas at kpc scales. Both datasets reveal the morphology to be that of a gas/dust rich bulge with two extending arms, together resembling a grand-design spiral galaxy. The more pronounced arm contributes 21 $\pm$ 6\% of the total CO emission. These results demonstrate that starburst activity at high redshift can be triggered, without undergoing a highly disruptive major merger. We assess the strength and distribution of star formation using two tracers: (1) Polycyclic Aromatic Hydrocarbons (PAHs) emission detected at $8~\mu$m ($L_8$) with a MIRI/F1800W image, and (2) $L_\mathrm{IR}$, inferred from the CO ($J$=5--4) map. The spatial profiles of the $L_\mathrm{IR}$ and $L_8$ are dissimilar, thus leading to a significant deficit of mid-IR ($L_8$) emission in the nucleus. We hypothesize that this is due to the destruction of PAH molecules by the intense ionizing radiation field or decreased emission in the photodissociation region, as seen in nearby star-forming regions and consistent with the galaxy-wide properties of distant starbursts. This study reveals spatial variations in the $L_8$ to $L_\mathrm{IR}$ ratio for the first time at $z>1$, in agreement with expectations from theory. Our analysis underscores the pivotal role of joint high-resolution observations with JWST and ALMA in discerning the different phases of the interstellar medium (ISM) and revealing internal physics in galaxy substructures.
|
[
{
"version": "v1",
"created": "Wed, 14 May 2025 18:36:39 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Jul 2025 18:08:57 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Liu",
"Zhaoxuan",
""
],
[
"Silverman",
"John D.",
""
],
[
"Daddi",
"Emanuele",
""
],
[
"Kalita",
"Boris S.",
""
],
[
"Puglisi",
"Annagrazia",
""
],
[
"Fei",
"Qinyue",
""
],
[
"Renzini",
"Alvio",
""
],
[
"Kashino",
"Daichi",
""
],
[
"Valentino",
"Francesco",
""
],
[
"Kartaltepe",
"Jeyhan S.",
""
],
[
"Liu",
"Daizhong",
""
],
[
"Pérez-González",
"Pablo G.",
""
],
[
"McKinney",
"Jed",
""
],
[
"Casey",
"Caitlin M.",
""
],
[
"Ding",
"Xuheng",
""
],
[
"Faisst",
"Andreas",
""
],
[
"Franco",
"Maximilien",
""
],
[
"Kakkad",
"Darshan",
""
],
[
"Koekemoer",
"Anton M.",
""
],
[
"Lambrides",
"Erini",
""
],
[
"Gillman",
"Steven",
""
],
[
"Gozaliasl",
"Ghassem",
""
],
[
"McCracken",
"Henry Joy",
""
],
[
"Rhodes",
"Jason",
""
],
[
"Robertson",
"Brant E.",
""
],
[
"Rodighiero",
"Giulia",
""
],
[
"Rujopakarn",
"Wiphu",
""
],
[
"Suzuki",
"Tomoko L.",
""
],
[
"Tanaka",
"Takumi S.",
""
],
[
"Vanderhoof",
"Brittany N.",
""
],
[
"Vijayan",
"Aswin P.",
""
],
[
"Cooper",
"Olivia R.",
""
],
[
"Kaminsky",
"Aidan",
""
],
[
"Magdis",
"Georgios E.",
""
],
[
"Roy",
"Namrata",
""
]
] |
2505.10280
|
Artem Sboychakov
|
A.O. Sboychakov
|
High frequency permeability of the composite with ferromagnetic spherical shells
|
15 pages, 9 figures
| null | null | null |
cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
The paper studies high-frequency permeability of the composite materials consisting of hollow ferromagnetic particles embedded into the non-magnetic media. We model the ferromagnetic particles in composite by spherical shells: the thickness of the ferromagnetic region $d$ compared to the particles' diameter $D$ can vary in a wide range, from $d\ll D$ to $d\sim D$. We assume that the magnetization distribution in such a particle is non-uniform, but forms a vortex-like structure: the magnetization is twisted in some plane outside two vortex cores placed at the poles of the particle. We consider two types of magnetic anisotropy, which help to stabilize such a magnetic configuration: the easy-plane magnetic anisotropy and the ``circular'' uniaxial magnetic anisotropy with easy axis rotated together with the magnetization direction outside the vortex core. The high-frequency permeability of such a composite material has been studied in the limit of non-interacting particles. We study the dependence of the permeability on the ratio $d/D$. We showed that for $d\sim D$ the composite's permeability behaves in similar manner for both types of magnetic anisotropy. It was shown also that for the second type of magnetic anisotropy and in the limit $d/D\ll1$ the frequency dependence of the particle's susceptibility is quite similar to that for the thin film. At the same time, the magnetization oscillations in the ac field are non-homogeneous for both types of anisotropy.
|
[
{
"version": "v1",
"created": "Thu, 15 May 2025 13:28:32 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:40:58 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sboychakov",
"A. O.",
""
]
] |
2505.10674
|
Floyd Stecker
|
Floyd W. Stecker
|
Astrophysical Gamma-ray and Neutrino Production Following from the Physics of Photon-Nucleon Interactions
|
15 pages, 5 figures, new expaanded results
| null | null | null |
astro-ph.HE hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The aim of this paper is to present a more complete consideration of the theoretical concepts and experimental aspects of the physics of photoproduction interactions involving nuclei. This treatment is based in large part on the most recent and extensive empirical data on particle photoproduction interactions with protons. The implications of photoproduction involving helium nuclei are also treated and compared with photoproduction involving protons. It is found that photoproduction interactions with helium nuclei produce approximately 10% of the number of pions as compared with photoproduction interactions involving protons in astrophysical sources, assuming the cosmological abundance of He. In addition to the production of pions, we also demonstrate the relative effect of excited nucleon p-pi and p-2pi resonances and rho, eta, omega, and K meson production and decay channels that result in neutrino and gamma-ray production. The production of mesons other than pions is shown not to be significant for calculations of astrophysical gamma-ray and neutrino production. It is further shown that, for astrophysical purposes, the Delta(1232) resonance channel clearly dominates. However, a blend of nucleon resonances at 1400 GeV can contribute as much as 20% to the neutrino flux.
|
[
{
"version": "v1",
"created": "Thu, 15 May 2025 19:36:49 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 14:17:23 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Stecker",
"Floyd W.",
""
]
] |
2505.10978
|
Lang Feng
|
Lang Feng, Zhenghai Xue, Tingcong Liu, Bo An
|
Group-in-Group Policy Optimization for LLM Agent Training
|
Preprint
| null | null | null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent advances in group-based reinforcement learning (RL) have driven frontier large language models (LLMs) in single-turn tasks like mathematical reasoning. However, their scalability to long-horizon LLM agent training remains limited. Unlike static tasks, agent-environment interactions unfold over many steps and often yield sparse or delayed rewards, making credit assignment across individual steps significantly more challenging. In this work, we propose Group-in-Group Policy Optimization (GiGPO), a novel RL algorithm that achieves fine-grained credit assignment for LLM agents while preserving the appealing properties of group-based RL: critic-free, low memory, and stable convergence. GiGPO introduces a two-level structure for estimating relative advantage: (i) At the episode-level, GiGPO computes macro relative advantages based on groups of complete trajectories; (ii) At the step-level, GiGPO introduces an anchor state grouping mechanism that retroactively constructs step-level groups by identifying repeated environment states across trajectories. Actions stemming from the same state are grouped together, enabling micro relative advantage estimation. This hierarchical structure effectively captures both global trajectory quality and local step effectiveness without relying on auxiliary models or additional rollouts. We evaluate GiGPO on two challenging agent benchmarks, ALFWorld and WebShop, using Qwen2.5-1.5B-Instruct and Qwen2.5-7B-Instruct. Crucially, GiGPO delivers fine-grained per-step credit signals and achieves performance gains of > 12\% on ALFWorld and > 9\% on WebShop over the GRPO baseline: all while maintaining the same GPU memory overhead, identical LLM rollout, and incurring little to no additional time cost.
|
[
{
"version": "v1",
"created": "Fri, 16 May 2025 08:26:59 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 16:13:41 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Feng",
"Lang",
""
],
[
"Xue",
"Zhenghai",
""
],
[
"Liu",
"Tingcong",
""
],
[
"An",
"Bo",
""
]
] |
2505.13188
|
Juntian Zhu
|
Juntian Zhu, Miguel de Carvalho, Zhouwang Yang, Fengxiang He
|
When a Reinforcement Learning Agent Encounters Unknown Unknowns
| null | null | null | null |
cs.LG cs.AI stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
An AI agent might surprisingly find she has reached an unknown state which she has never been aware of -- an unknown unknown. We mathematically ground this scenario in reinforcement learning: an agent, after taking an action calculated from value functions $Q$ and $V$ defined on the {\it {aware domain}}, reaches a state out of the domain. To enable the agent to handle this scenario, we propose an {\it episodic Markov decision {process} with growing awareness} (EMDP-GA) model, taking a new {\it noninformative value expansion} (NIVE) approach to expand value functions to newly aware areas: when an agent arrives at an unknown unknown, value functions $Q$ and $V$ whereon are initialised by noninformative beliefs -- the averaged values on the aware domain. This design is out of respect for the complete absence of knowledge in the newly discovered state. The upper confidence bound momentum Q-learning is then adapted to the growing awareness for training the EMDP-GA model. We prove that (1) the regret of our approach is asymptotically consistent with the state of the art (SOTA) without exposure to unknown unknowns in an extremely uncertain environment, and (2) our computational complexity and space complexity are comparable with the SOTA -- these collectively suggest that though an unknown unknown is surprising, it will be asymptotically properly discovered with decent speed and an affordable cost.
|
[
{
"version": "v1",
"created": "Mon, 19 May 2025 14:45:58 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 11:20:30 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhu",
"Juntian",
""
],
[
"de Carvalho",
"Miguel",
""
],
[
"Yang",
"Zhouwang",
""
],
[
"He",
"Fengxiang",
""
]
] |
2505.13424
|
Jonathan Moussa
|
Jonathan E. Moussa
|
The Enduring Relevance of Semiempirical Quantum Mechanics
|
16 pages, 4 figures
| null | null | null |
physics.chem-ph cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
The development of semiempirical models to simplify quantum mechanical descriptions of atomistic systems is a practice that started soon after the discovery of quantum mechanics and continues to the present day. There are now many methods for atomistic simulation with many software implementations and many users, on a scale large enough to be considered as a software market. Semiempirical models occupied a large share of this market in its early days, but the research activity in atomistic simulation has steadily polarized over the last three decades towards general-purpose but expensive ab initio quantum mechanics methods and fast but special-purpose molecular mechanics methods. I offer perspective on recent trends in atomistic simulation from the middle ground of semiempirical modeling, to learn from its past success and consider its possible paths to future growth. In particular, there is a lot of ongoing research activity in combining semiempirical quantum mechanics with machine learning models and some unrealized possibilities of tighter integration between ab initio and semiempirical quantum mechanics with more flexible theoretical frameworks and more modular software components.
|
[
{
"version": "v1",
"created": "Mon, 19 May 2025 17:53:45 GMT"
},
{
"version": "v2",
"created": "Sun, 03 Aug 2025 00:36:25 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 16:18:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Moussa",
"Jonathan E.",
""
]
] |
2505.13754
|
Joshua Daymude
|
Devendra Parkar and Anya Chaturvedi and Joshua J. Daymude
|
Unsupervised Learning of Local Updates for Maximum Independent Set in Dynamic Graphs
|
13 pages, 2 figures, 2 tables, 3 algorithms
| null | null | null |
cs.LG cs.SI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present the first unsupervised learning model for finding Maximum Independent Sets (MaxIS) in dynamic graphs where edges change over time. Our method combines structural learning from graph neural networks (GNNs) with a learned distributed update mechanism that, given an edge addition or deletion event, modifies nodes' internal memories and infers their MaxIS membership in a single, parallel step. We parameterize our model by the update mechanism's radius and investigate the resulting performance-runtime tradeoffs for various dynamic graph topologies. We evaluate our model against a mixed integer programming solver and the state-of-the-art learning-based methods for MaxIS on static graphs (ICML 2020; NeurIPS 2020, 2023). Across synthetic and empirical dynamic graphs of 50-1,000 nodes, our model achieves competitive approximation ratios with excellent scalability; on large graphs, it significantly outperforms the state-of-the-art learning methods in solution quality, runtime, and memory usage. When generalizing to graphs of 10,000 nodes (100x larger than the ones used for training), our model produces MaxIS solutions 1.05-1.18x larger than any other learning method, even while maintaining competitive runtimes.
|
[
{
"version": "v1",
"created": "Mon, 19 May 2025 21:58:22 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 06:55:09 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Parkar",
"Devendra",
""
],
[
"Chaturvedi",
"Anya",
""
],
[
"Daymude",
"Joshua J.",
""
]
] |
2505.14210
|
Berend van Wachem
|
Gabriele Gennari and Christian Gorges and Fabian Denner and Berend van Wachem
|
A marching cubes based method for topology changes in three-dimensional two-phase flows with front tracking
| null |
Journal of Computational Physics, Vol 540, 114284 (2025)
|
10.1016/j.jcp.2025.114284
| null |
physics.flu-dyn physics.comp-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The handling of topology changes in two-phase flows, such as breakup or coalescence of interfaces, with front tracking is a well-known problem that requires an additional effort to perform explicit manipulations of the Lagrangian front. In this work, we present an approach that allows to perform topology changes with interfaces made of connected triangular elements. The methodology consists of replacing the fluid entities that undergo breakup/coalescence with the iso-surface corresponding to the indicator function value I = 0.5, which automatically returns the shape of the bodies after topology changes. The generation and triangulation of such surface is obtained by exploiting the marching cubes algorithm. Since we perform the reconstruction of the interface only for the bodies that experience breakup/coalescence, the increase in computational cost with respect to a classic front tracking scheme without topology changes is small. Using validation cases, we show that the proposed reconstruction procedure is second-order accurate for volume conservation and able to capture the physics of several two-phase flow configurations undergoing topology changes. The validation cases include the breakup of a droplet in simple shear flow and two rising bubbles in different regimes (peripheral and central breakups). Coalescence is tested by modelling the binary collision between two droplets. For the selected validation cases, an excellent agreement between the numerical results and experiments is observed. The proposed methodology is able to capture the details of such interfacial flows, by predicting accurately the coalescence/breakup dynamics, as well as the number, size and shapes of satellite droplets/bubbles after topology changes.
|
[
{
"version": "v1",
"created": "Tue, 20 May 2025 11:13:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Gennari",
"Gabriele",
""
],
[
"Gorges",
"Christian",
""
],
[
"Denner",
"Fabian",
""
],
[
"van Wachem",
"Berend",
""
]
] |
2505.14257
|
Jianfei Zhao
|
Jianfei Zhao, Feng Zhang, Xin Sun, Chong Feng
|
Mitigating Hallucination in Large Vision-Language Models through Aligning Attention Distribution to Information Flow
|
Accepted to Findings of EMNLP 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Due to the unidirectional masking mechanism, Decoder-Only models propagate information from left to right. LVLMs (Large Vision-Language Models) follow the same architecture, with visual information gradually integrated into semantic representations during forward propagation. Through systematic analysis, we observe that the majority of the visual information is absorbed into the semantic representations. However, the model's attention distribution does not exhibit sufficient emphasis on semantic representations. This misalignment between the attention distribution and the actual information flow undermines the model's visual understanding ability and contributes to hallucinations. To address this issue, we enhance the model's visual understanding by leveraging the core information embedded in semantic representations. Specifically, we identify attention heads that focus on core semantic representations based on their attention distributions. Then, through a two-stage optimization paradigm, we propagate the advantages of these attention heads across the entire model, aligning the attention distribution with the actual information flow. We evaluate our method on three image captioning benchmarks using five different LVLMs, demonstrating its effectiveness in significantly reducing hallucinations. Further experiments reveal a trade-off between reduced hallucinations and richer details. Notably, our method allows for manual adjustment of the model's conservativeness, enabling flexible control to meet diverse real-world requirements.
|
[
{
"version": "v1",
"created": "Tue, 20 May 2025 12:10:13 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 11:34:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhao",
"Jianfei",
""
],
[
"Zhang",
"Feng",
""
],
[
"Sun",
"Xin",
""
],
[
"Feng",
"Chong",
""
]
] |
2505.16022
|
Wei Liu
|
Wei Liu, Siya Qi, Xinyu Wang, Chen Qian, Yali Du, Yulan He
|
NOVER: Incentive Training for Language Models via Verifier-Free Reinforcement Learning
|
20 pages, 5 tables, 12 figures. accepted to EMNLP 2025
| null | null | null |
cs.CL cs.AI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent advances such as DeepSeek R1-Zero highlight the effectiveness of incentive training, a reinforcement learning paradigm that computes rewards solely based on the final answer part of a language model's output, thereby encouraging the generation of intermediate reasoning steps. However, these methods fundamentally rely on external verifiers, which limits their applicability to domains like mathematics and coding where such verifiers are readily available. Although reward models can serve as verifiers, they require high-quality annotated data and are costly to train. In this work, we propose NOVER, NO-VERifier Reinforcement Learning, a general reinforcement learning framework that requires only standard supervised fine-tuning data with no need for an external verifier. NOVER enables incentive training across a wide range of text-to-text tasks and outperforms the model of the same size distilled from large reasoning models such as DeepSeek R1 671B by 7.7 percent. Moreover, the flexibility of NOVER enables new possibilities for optimizing large language models, such as inverse incentive training.
|
[
{
"version": "v1",
"created": "Wed, 21 May 2025 21:12:35 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 20:28:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Liu",
"Wei",
""
],
[
"Qi",
"Siya",
""
],
[
"Wang",
"Xinyu",
""
],
[
"Qian",
"Chen",
""
],
[
"Du",
"Yali",
""
],
[
"He",
"Yulan",
""
]
] |
2505.16893
|
Shuichi Nishino
|
Shuichi Nishino, Tomohiro Shiraishi, Teruyuki Katsuoka, Ichiro Takeuchi
|
Statistical Test for Saliency Maps of Graph Neural Networks via Selective Inference
| null | null | null | null |
stat.ML cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Graph Neural Networks (GNNs) have gained prominence for their ability to process graph-structured data across various domains. However, interpreting GNN decisions remains a significant challenge, leading to the adoption of saliency maps for identifying salient subgraphs composed of influential nodes and edges. Despite their utility, the reliability of GNN saliency maps has been questioned, particularly in terms of their robustness to input noise. In this study, we propose a statistical testing framework to rigorously evaluate the significance of saliency maps. Our main contribution lies in addressing the inflation of the Type I error rate caused by double-dipping of data, leveraging the framework of Selective Inference. Our method provides statistically valid $p$-values while controlling the Type I error rate, ensuring that identified salient subgraphs contain meaningful information rather than random artifacts. The method is applicable to a variety of saliency methods with piecewise linearity (e.g., Class Activation Mapping). We validate our method on synthetic and real-world datasets, demonstrating its capability in assessing the reliability of GNN interpretations.
|
[
{
"version": "v1",
"created": "Thu, 22 May 2025 16:50:55 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 07:31:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Nishino",
"Shuichi",
""
],
[
"Shiraishi",
"Tomohiro",
""
],
[
"Katsuoka",
"Teruyuki",
""
],
[
"Takeuchi",
"Ichiro",
""
]
] |
2505.17067
|
Kristin Qi
|
Kristin Qi, Jiali Cheng, Youxiang Zhu, Hadi Amiri, Xiaohui Liang
|
Unveil Multi-Picture Descriptions for Multilingual Mild Cognitive Impairment Detection via Contrastive Learning
|
IEEE Global Communications Conference (GlobeCom) 2025
| null | null | null |
cs.CL cs.AI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Detecting Mild Cognitive Impairment from picture descriptions is critical yet challenging, especially in multilingual and multiple picture settings. Prior work has primarily focused on English speakers describing a single picture (e.g., the 'Cookie Theft'). The TAUKDIAL-2024 challenge expands this scope by introducing multilingual speakers and multiple pictures, which presents new challenges in analyzing picture-dependent content. To address these challenges, we propose a framework with three components: (1) enhancing discriminative representation learning via supervised contrastive learning, (2) involving image modality rather than relying solely on speech and text modalities, and (3) applying a Product of Experts (PoE) strategy to mitigate spurious correlations and overfitting. Our framework improves MCI detection performance, achieving a +7.1% increase in Unweighted Average Recall (UAR) (from 68.1% to 75.2%) and a +2.9% increase in F1 score (from 80.6% to 83.5%) compared to the text unimodal baseline. Notably, the contrastive learning component yields greater gains for the text modality compared to speech. These results highlight our framework's effectiveness in multilingual and multi-picture MCI detection.
|
[
{
"version": "v1",
"created": "Mon, 19 May 2025 03:03:08 GMT"
},
{
"version": "v2",
"created": "Mon, 26 May 2025 08:18:33 GMT"
},
{
"version": "v3",
"created": "Mon, 28 Jul 2025 16:28:49 GMT"
},
{
"version": "v4",
"created": "Mon, 01 Sep 2025 05:15:02 GMT"
},
{
"version": "v5",
"created": "Wed, 03 Sep 2025 03:38:04 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Qi",
"Kristin",
""
],
[
"Cheng",
"Jiali",
""
],
[
"Zhu",
"Youxiang",
""
],
[
"Amiri",
"Hadi",
""
],
[
"Liang",
"Xiaohui",
""
]
] |
2505.17137
|
Kristin Qi
|
Kristin Qi, Youxiang Zhu, Caroline Summerour, John A. Batsis, Xiaohui Liang
|
Cog-TiPRO: Iterative Prompt Refinement with LLMs to Detect Cognitive Decline via Longitudinal Voice Assistant Commands
|
IEEE Global Communications Conference (GlobeCom) 2025
| null | null | null |
cs.CL cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Early detection of cognitive decline is crucial for enabling interventions that can slow neurodegenerative disease progression. Traditional diagnostic approaches rely on labor-intensive clinical assessments, which are impractical for frequent monitoring. Our pilot study investigates voice assistant systems (VAS) as non-invasive tools for detecting cognitive decline through longitudinal analysis of speech patterns in voice commands. Over an 18-month period, we collected voice commands from 35 older adults, with 15 participants providing daily at-home VAS interactions. To address the challenges of analyzing these short, unstructured and noisy commands, we propose Cog-TiPRO, a framework that combines (1) LLM-driven iterative prompt refinement for linguistic feature extraction, (2) HuBERT-based acoustic feature extraction, and (3) transformer-based temporal modeling. Using iTransformer, our approach achieves 73.80% accuracy and 72.67% F1-score in detecting MCI, outperforming its baseline by 27.13%. Through our LLM approach, we identify linguistic features that uniquely characterize everyday command usage patterns in individuals experiencing cognitive decline.
|
[
{
"version": "v1",
"created": "Thu, 22 May 2025 05:40:12 GMT"
},
{
"version": "v2",
"created": "Mon, 28 Jul 2025 16:30:46 GMT"
},
{
"version": "v3",
"created": "Sun, 31 Aug 2025 03:04:11 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 03:28:34 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Qi",
"Kristin",
""
],
[
"Zhu",
"Youxiang",
""
],
[
"Summerour",
"Caroline",
""
],
[
"Batsis",
"John A.",
""
],
[
"Liang",
"Xiaohui",
""
]
] |
2505.17879
|
Zengrui Han
|
Zengrui Han, Lu Bai, Ziwei Huang, and Xiang Cheng
|
LLM4SG: Adapting Large Language Model for Scatterer Generation via Synesthesia of Machines
| null | null | null | null |
eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, a novel large language model (LLM)-based method for scatterer generation (LLM4SG) is proposed for sixth-generation (6G) artificial intelligence (AI)-native communications. To provide a solid data foundation, we construct a new synthetic intelligent sensing-communication dataset for Synesthesia of Machines (SoM) in vehicle-to-vehicle (V2V) communications, named SynthSoM-V2V, covering multiple V2V scenarios with multiple frequency bands and multiple vehicular traffic densities (VTDs). Leveraging the powerful cross-modal representation capabilities of LLMs, LLM4SG is designed to capture the general mapping relationship from light detection and ranging (LiDAR) point clouds to electromagnetic scatterers via SoM. To address the inherent and significant differences across multi-modal data, synergistically optimized four-module architecture, i.e., preprocessor, embedding, backbone, and output modules, are designed by considering sensing characteristics and electromagnetic propagation. The embedding module achieves effective cross-domain alignment of the sensing-communication domain and the natural language domain.The backbone network is adapted in a task-guided manner with low rank adaptation (LoRA), where a carefully selected subset of layers is fine tuned to preserve general knowledge and reduce training cost. The proposed LLM4SG is evaluated for scatterer generation by benchmarking against ray-tracing (RT) and conventional deep learning models. Simulation results demonstrate that the proposed LLM4SG achieves superior performance in both full-sample and cross-condition generalization testing. It significantly outperforms conventional deep learning models across different frequency bands, scenarios, and VTDs, and demonstrates the capability to provide the massive and high-quality channel small-scale fading data required by AI-native 6G systems.
|
[
{
"version": "v1",
"created": "Fri, 23 May 2025 13:28:26 GMT"
},
{
"version": "v2",
"created": "Mon, 02 Jun 2025 14:17:22 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 03:21:11 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Han",
"Zengrui",
""
],
[
"Bai",
"Lu",
""
],
[
"Huang",
"Ziwei",
""
],
[
"Cheng",
"Xiang",
""
]
] |
2505.19099
|
Kun Xiang
|
Kun Xiang and Heng Li and Terry Jingchen Zhang and Yinya Huang and Zirong Liu and Peixin Qu and Jixi He and Jiaqi Chen and Yu-Jie Yuan and Jianhua Han and Hang Xu and Hanhui Li and Mrinmaya Sachan and Xiaodan Liang
|
SeePhys: Does Seeing Help Thinking? -- Benchmarking Vision-Based Physics Reasoning
|
46 pages
| null | null | null |
cs.AI physics.ed-ph physics.pop-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We present SeePhys, a large-scale multimodal benchmark for LLM reasoning grounded in physics questions ranging from middle school to PhD qualifying exams. The benchmark covers 7 fundamental domains spanning the physics discipline, incorporating 21 categories of highly heterogeneous diagrams. In contrast to prior works where visual elements mainly serve auxiliary purposes, our benchmark features a substantial proportion of vision-essential problems (75%) that mandate visual information extraction for correct solutions. Through extensive evaluation, we observe that even the most advanced visual reasoning models (e.g., Gemini-2.5-pro and o4-mini) achieve sub-60% accuracy on our benchmark. These results reveal fundamental challenges in current large language models' visual understanding capabilities, particularly in: (i) establishing rigorous coupling between diagram interpretation and physics reasoning, and (ii) overcoming their persistent reliance on textual cues as cognitive shortcuts.
|
[
{
"version": "v1",
"created": "Sun, 25 May 2025 11:28:34 GMT"
},
{
"version": "v2",
"created": "Wed, 28 May 2025 12:02:43 GMT"
},
{
"version": "v3",
"created": "Fri, 30 May 2025 18:22:56 GMT"
},
{
"version": "v4",
"created": "Tue, 17 Jun 2025 14:31:36 GMT"
},
{
"version": "v5",
"created": "Mon, 21 Jul 2025 12:44:10 GMT"
},
{
"version": "v6",
"created": "Wed, 03 Sep 2025 05:05:25 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Xiang",
"Kun",
""
],
[
"Li",
"Heng",
""
],
[
"Zhang",
"Terry Jingchen",
""
],
[
"Huang",
"Yinya",
""
],
[
"Liu",
"Zirong",
""
],
[
"Qu",
"Peixin",
""
],
[
"He",
"Jixi",
""
],
[
"Chen",
"Jiaqi",
""
],
[
"Yuan",
"Yu-Jie",
""
],
[
"Han",
"Jianhua",
""
],
[
"Xu",
"Hang",
""
],
[
"Li",
"Hanhui",
""
],
[
"Sachan",
"Mrinmaya",
""
],
[
"Liang",
"Xiaodan",
""
]
] |
2505.19269
|
Therese Landry
|
Anshu, David Jekel, Therese Basa Landry
|
Quantum Wasserstein distances for quantum permutation groups
|
27 pages
| null | null | null |
math.OA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We seek an analog for the quantum permutation group $S_n^+$ of the normalized Hamming distance for permutations. We define three distances on the tracial state space of $C(S_n^+)$ that generalize the $L^1$-Wasserstein distance of probability measures on $S_n$ equipped with the normalized Hamming metric, for which we demonstrate basic metric properties, subadditivity under convolution, and density of the Lipschitz elements in the $\mathrm{C}^{\ast}$-algebra.
|
[
{
"version": "v1",
"created": "Sun, 25 May 2025 18:59:33 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 11:54:33 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Anshu",
"",
""
],
[
"Jekel",
"David",
""
],
[
"Landry",
"Therese Basa",
""
]
] |
2505.19758
|
Yu-Shuai Li
|
Yu-Shuai Li
|
The nonleptonic decays $\Xi_{cc}^{++}\to\Xi_{c}^{(\prime)+}\pi^{+}$ within the nonrelativistic quark model
|
13 pages, 5 figures, 8 tables, to be published in EPJC
| null |
10.1140/epjc/s10052-025-14670-4
|
Eur. Phys. J. C 85, 938 (2025)
|
hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this work, we study the nonleptonic decays $\Xi_{cc}^{++}\to\Xi_{c}^{(\prime)+}\pi^{+}$ with considering $\Xi_{c}-\Xi_{c}^{\prime}$ mixing. The relevant decay amplitudes are evaluated within the framework of nonrelativistic quark model, combining the baryon spatial wave functions adopted from solving the Schr\"{o}dinger equation with a nonrelativistic potential. With the mixing angle ranging $\theta\in(-18.2^{\circ},-14.3^{\circ})$, we successfully reproduce the measured ratio $R=\mathcal{B}[\Xi_{cc}^{++}\to\Xi_{c}^{\prime+}\pi^{+}]/\mathcal{B}[\Xi_{cc}^{++}\to\Xi_{c}^{+}\pi^{+}]$ reported by the LHCb Collaboration. Furthermore, we estimate the branching fractions as $\mathcal{B}[\Xi_{cc}^{++}\to\Xi_{c}^{+}\pi^{+}]=(3.2\sim4.3)\%$ and $\mathcal{B}[\Xi_{cc}^{++}\to\Xi_{c}^{\prime+}\pi^{+}]=(4.2\sim6.0)\%$, and the asymmetry parameters as $\alpha[\Xi_{cc}^{++}\to\Xi_{c}^{+}\pi^{+}]=(-0.80\sim-0.81)$ and $\alpha[\Xi_{cc}^{++}\to\Xi_{c}^{\prime+}\pi^{+}]=(-0.61\sim-0.62)$. The measurements of absolute branching fractions and asymmetry parameters by the ongoing LHCb and Belle II experiments will be helpful for further testing our numerical results and confirming the mixing angle.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 09:37:23 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 11:57:54 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Li",
"Yu-Shuai",
""
]
] |
2505.19988
|
Theodore Johnson
|
Vladislav Shkapenyuk, Divesh Srivastava, Theodore Johnson, Parisa Ghane
|
Automatic Metadata Extraction for Text-to-SQL
|
37 pages
| null | null | null |
cs.DB
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Large Language Models (LLMs) have recently become sophisticated enough to automate many tasks ranging from pattern finding to writing assistance to code generation. In this paper, we examine text-to-SQL generation. We have observed from decades of experience that the most difficult part of query development lies in understanding the database contents. These experiences inform the direction of our research.
Text-to-SQL benchmarks such as SPIDER and Bird contain extensive metadata that is generally not available in practice. Human-generated metadata requires the use of expensive Subject Matter Experts (SMEs), who are often not fully aware of many aspects of their databases. In this paper, we explore techniques for automatic metadata extraction to enable text-to-SQL generation.
We explore the use of two standard and one newer metadata extraction techniques: profiling, query log analysis, and SQL-to text generation using an LLM. We use BIRD benchmark [JHQY+23] to evaluate the effectiveness of these techniques. BIRD does not provide query logs on their test database, so we prepared a submission that uses profiling alone, and does not use any specially tuned model (we used GPT-4o). From Sept 1 to Sept 23, 2024, and Nov 11 through Nov 23, 2024 we achieved the highest score both with and without using the "oracle" information provided with the question set. We regained the number 1 spot on Mar 11, 2025, and are still at #1 at the time of the writing (May, 2025).
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 13:43:43 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Jun 2025 15:23:03 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Shkapenyuk",
"Vladislav",
""
],
[
"Srivastava",
"Divesh",
""
],
[
"Johnson",
"Theodore",
""
],
[
"Ghane",
"Parisa",
""
]
] |
2505.20015
|
Ramon Ferrer-i-Cancho
|
Ramon Ferrer-i-Cancho
|
On the class of coding optimality of human languages and the origins of Zipf's law
|
a few typos corrected, in press in Europhysics Letters
| null | null | null |
cs.CL physics.soc-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Here we present a new class of optimality for coding systems. Members of that class are displaced linearly from optimal coding and thus exhibit Zipf's law, namely a power-law distribution of frequency ranks. Within that class, Zipf's law, the size-rank law and the size-probability law form a group-like structure. We identify human languages that are members of the class. All languages showing sufficient agreement with Zipf's law are potential members of the class. In contrast, there are communication systems in other species that cannot be members of that class for exhibiting an exponential distribution instead but dolphins and humpback whales might. We provide a new insight into plots of frequency versus rank in double logarithmic scale. For any system, a straight line in that scale indicates that the lengths of optimal codes under non-singular coding and under uniquely decodable encoding are displaced by a linear function whose slope is the exponent of Zipf's law. For systems under compression and constrained to be uniquely decodable, such a straight line may indicate that the system is coding close to optimality. We provide support for the hypothesis that Zipf's law originates from compression and define testable conditions for the emergence of Zipf's law in compressing systems.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 14:05:45 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Jun 2025 17:00:20 GMT"
},
{
"version": "v3",
"created": "Wed, 04 Jun 2025 11:35:43 GMT"
},
{
"version": "v4",
"created": "Fri, 18 Jul 2025 14:57:19 GMT"
},
{
"version": "v5",
"created": "Tue, 02 Sep 2025 18:22:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ferrer-i-Cancho",
"Ramon",
""
]
] |
2505.20203
|
Elliott Thornley
|
Elliott Thornley
|
Shutdownable Agents through POST-Agency
| null | null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Many fear that future artificial agents will resist shutdown. I present an idea - the POST-Agents Proposal - for ensuring that doesn't happen. I propose that we train agents to satisfy Preferences Only Between Same-Length Trajectories (POST). I then prove that POST - together with other conditions - implies Neutrality+: the agent maximizes expected utility, ignoring the probability distribution over trajectory-lengths. I argue that Neutrality+ keeps agents shutdownable and allows them to be useful.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 16:44:17 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 16:55:20 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Thornley",
"Elliott",
""
]
] |
2505.20244
|
Miguel Angel Moreno-Mateos
|
Miguel Angel Moreno-Mateos, Simon Wiesheier, Ali Esmaeili, Mokarram Hossain, Paul Steinmann
|
Biaxial characterization of soft elastomers: experiments and data-adaptive configurational forces for fracture
| null | null |
10.1016/j.jmps.2025.106339
| null |
cond-mat.soft cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Understanding the fracture mechanics of soft solids remains a fundamental challenge due to their complex, nonlinear responses under large deformations. While multiaxial loading is key to probing their mechanical behavior, the role of such loading in fracture processes is still poorly understood. Here, we present a combined experimental-computational framework to investigate fracture in soft elastomers under equi-biaxial loading. We report original equi-biaxial quasi-static experiments on five elastomeric materials, revealing a spectrum of material and fracture behavior, from brittle-like to highly deformable response with crack tip strains exceeding 150 %. Motivated by these observations, we develop a hybrid computational testbed that mirrors the experimental setup and enables virtual biaxial tests. Central to this framework are two components: a data-adaptive formulation of hyperelastic energy functions that flexibly captures material behavior, and a post-processing implementation of the Configurational Force Method, providing a computationally efficient estimate of the J-integral at the crack tip. Our data-adaptive framework for hyperelastic energy functions proves versatility to capture with high accuracy the hyperelastic behavior observed in the biaxial experiments. This is important because accurately capturing the constitutive behaviour of soft solids is key for a reliable application of the Configurational Force Method to soft solids. In the limit of crack onset, a critical value of the crack tip configurational force allows for a criterion of fracture toughness. Together, our experimental, theoretical, and computational contributions offer a new paradigm for characterizing and designing soft materials with tailored fracture properties.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 17:22:11 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Moreno-Mateos",
"Miguel Angel",
""
],
[
"Wiesheier",
"Simon",
""
],
[
"Esmaeili",
"Ali",
""
],
[
"Hossain",
"Mokarram",
""
],
[
"Steinmann",
"Paul",
""
]
] |
2505.20353
|
Dong Liu
|
Dong Liu, Yanxuan Yu, Jiayi Zhang, Yifan Li, Ben Lengerich, Ying Nian Wu
|
FastCache: Fast Caching for Diffusion Transformer Through Learnable Linear Approximation
| null | null | null | null |
cs.LG cs.AI cs.CV cs.MM cs.PF
|
http://creativecommons.org/licenses/by/4.0/
|
Diffusion Transformers (DiT) are powerful generative models but remain computationally intensive due to their iterative structure and deep transformer stacks. To alleviate this inefficiency, we propose FastCache, a hidden-state-level caching and compression framework that accelerates DiT inference by exploiting redundancy within the model's internal representations. FastCache introduces a dual strategy: (1) a spatial-aware token selection mechanism that adaptively filters redundant tokens based on hidden state saliency, and (2) a transformer-level cache that reuses latent activations across timesteps when changes are statistically insignificant. These modules work jointly to reduce unnecessary computation while preserving generation fidelity through learnable linear approximation. Theoretical analysis shows that FastCache maintains bounded approximation error under a hypothesis-testing-based decision rule. Empirical evaluations across multiple DiT variants demonstrate substantial reductions in latency and memory usage, with best generation output quality compared to other cache methods, as measured by FID and t-FID. Code implementation of FastCache is available on GitHub at https://github.com/NoakLiu/FastCache-xDiT.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 05:58:49 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 06:56:21 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Liu",
"Dong",
""
],
[
"Yu",
"Yanxuan",
""
],
[
"Zhang",
"Jiayi",
""
],
[
"Li",
"Yifan",
""
],
[
"Lengerich",
"Ben",
""
],
[
"Wu",
"Ying Nian",
""
]
] |
2505.20357
|
Jun Tian
|
Jun Tian, He Wang, Jibo He, Yu Pan, Shuo Cao, and Qingquan Jiang
|
Learning and Interpreting Gravitational-Wave Features from CNNs with a Random Forest Approach
| null |
2025 Mach. Learn.: Sci. Technol. 6 035045
|
10.1088/2632-2153/adfc27
| null |
cs.LG gr-qc physics.data-an
|
http://creativecommons.org/licenses/by/4.0/
|
Convolutional neural networks (CNNs) have become widely adopted in gravitational wave (GW) detection pipelines due to their ability to automatically learn hierarchical features from raw strain data. However, the physical meaning of these learned features remains underexplored, limiting the interpretability of such models. In this work, we propose a hybrid architecture that combines a CNN-based feature extractor with a random forest (RF) classifier to improve both detection performance and interpretability. Unlike prior approaches that directly connect classifiers to CNN outputs, our method introduces four physically interpretable metrics - variance, signal-to-noise ratio (SNR), waveform overlap, and peak amplitude - computed from the final convolutional layer. These are jointly used with the CNN output in the RF classifier to enable more informed decision boundaries. Tested on long-duration strain datasets, our hybrid model outperforms a baseline CNN model, achieving a relative improvement of 21\% in sensitivity at a fixed false alarm rate of 10 events per month. Notably, it also shows improved detection of low-SNR signals (SNR $\le$ 10), which are especially vulnerable to misclassification in noisy environments. Feature attribution via the RF model reveals that both CNN-extracted and handcrafted features contribute significantly to classification decisions, with learned variance and CNN outputs ranked among the most informative. These findings suggest that physically motivated post-processing of CNN feature maps can serve as a valuable tool for interpretable and efficient GW detection, bridging the gap between deep learning and domain knowledge.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 07:33:27 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 09:08:31 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tian",
"Jun",
""
],
[
"Wang",
"He",
""
],
[
"He",
"Jibo",
""
],
[
"Pan",
"Yu",
""
],
[
"Cao",
"Shuo",
""
],
[
"Jiang",
"Qingquan",
""
]
] |
2505.20543
|
Mattia Miotto
|
Mattia Miotto, Giorgio Gosti, Maria Rosito, Michela Dell'Omo, Viola Folli, Valeria de Turris, Giancarlo Ruocco, Alessandro Rosa, Matteo Paoluzzi
|
Rosette formations as symmetry-breaking events: theory and experiment
|
14 pages, 7 figures
| null | null | null |
q-bio.CB cond-mat.soft cond-mat.stat-mech physics.bio-ph
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Multicellular rosettes are observed in different situations such as morphogenesis, wound healing, and cancer progression. While some molecular insights have been gained to explain the presence of these assemblies of five or more cells around a common center, what are the tunable, global features that favors/hinders their formation is still largely unknown. Here, we made use of a Voronoi dynamical model to investigate the ingredients driving the emergence of rosettes characterized by different degree of stability and organization. We found that (i) breaking the local spatial symmetry of the system, i.e., introducing curvature-inducing defects, allows for the formation of rosette-like structures (ii) whose probability of formation depends on the characteristic of the cellular layer. In particular, a trade-off between tissue fluidity and single cell deformability dictates the assembly of transient rosettes, that are strongly stabilized in the presence of cell alignment interactions. To test our model predictions, we performed fluorescence microscopy experiments on rosette-forming neural populations derived from induced pluripotent stem cells, finding significant agreement. Overall, our work may set the stage to gain an unifying understanding of the plethora of biophysical mechanisms involving the occurrence of rosette-like structure both in physiology and their altered formation in pathology.
|
[
{
"version": "v1",
"created": "Mon, 26 May 2025 22:12:36 GMT"
},
{
"version": "v2",
"created": "Sun, 17 Aug 2025 18:28:29 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Miotto",
"Mattia",
""
],
[
"Gosti",
"Giorgio",
""
],
[
"Rosito",
"Maria",
""
],
[
"Dell'Omo",
"Michela",
""
],
[
"Folli",
"Viola",
""
],
[
"de Turris",
"Valeria",
""
],
[
"Ruocco",
"Giancarlo",
""
],
[
"Rosa",
"Alessandro",
""
],
[
"Paoluzzi",
"Matteo",
""
]
] |
2505.20797
|
Antonio Tudisco
|
Antonio Tudisco, Deborah Volpe, and Giovanna Turvani
|
Multi-VQC: A Novel QML Approach for Enhancing Healthcare Classification
| null | null |
10.1109/QSW67625.2025.00013
| null |
cs.LG cs.ET
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Accurate and reliable diagnosis of diseases is crucial in enabling timely medical treatment and enhancing patient survival rates. In recent years, Machine Learning has revolutionized diagnostic practices by creating classification models capable of identifying diseases. However, these classification problems often suffer from significant class imbalances, which can inhibit the effectiveness of traditional models. Therefore, the interest in Quantum models has arisen, driven by the captivating promise of overcoming the limitations of the classical counterpart thanks to their ability to express complex patterns by mapping data in a higher-dimensional computational space.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 07:00:33 GMT"
},
{
"version": "v2",
"created": "Mon, 04 Aug 2025 07:18:49 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tudisco",
"Antonio",
""
],
[
"Volpe",
"Deborah",
""
],
[
"Turvani",
"Giovanna",
""
]
] |
2505.21360
|
Dhanesh Ramachandram
|
Dhanesh Ramachandram and Ananya Raval
|
CRISP-NAM: Competing Risks Interpretable Survival Prediction with Neural Additive Models
|
Added Feature Importance Diagrams and co-author
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Competing risks are crucial considerations in survival modelling, particularly in healthcare domains where patients may experience multiple distinct event types. We propose CRISP-NAM (Competing Risks Interpretable Survival Prediction with Neural Additive Models), an interpretable neural additive model for competing risks survival analysis which extends the neural additive architecture to model cause-specific hazards while preserving feature-level interpretability. Each feature contributes independently to risk estimation through dedicated neural networks, allowing for visualization of complex non-linear relationships between covariates and each competing risk. We demonstrate competitive performance on multiple datasets compared to existing approaches.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 15:52:15 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 12:41:47 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 18:49:10 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 16:09:10 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ramachandram",
"Dhanesh",
""
],
[
"Raval",
"Ananya",
""
]
] |
2505.21761
|
Tomppa Pakarinen
|
Tomppa Pakarinen
|
On the Role of Demagnetizing Tensors in Arbitrary Orientations of General Ellipsoid: Implications for MRI Safety Assessment
|
This work has been submitted to the IOP for possible publication
| null | null | null |
physics.med-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This work explores the behaviour of demagnetizing tensors for general ellipsoids under arbitrary rotations in homogeneous magnetic fields. The work is motivated by the concerns in magnetic resonance imaging safety and their practical evaluation in clinical environments. Whereas demagnetizing tensor is a well-defined concept in the principal axes, its transformation under three-dimensional reorientation is often overlooked - a justifiable omission for solutions derived from Poisson equation, where the tensor can be directly rotated. However, this does not hold for common approximations, where such tensor is not explicitly defined. This work demonstrates the validity of directly rotating the orthogonal basis solutions, derived from Poissons equation, and uses the procedure to evaluate a practical approximation, based on orthogonal area-projections. The tensor rotation approach is also applied to generalize force and torque calculations for ellipsoids under three-dimensional re-orientation. The results show an exact match for translation force and torque when compared to the standard single axis rotation. Additionally, a unique connection to the well-known MRI magic angle is found as the point of convergence for prolate spheroid aspect ratios while solving the corresponding saturation field magnitudes. Finally, the evaluated approximate method was demonstrated to perform fairly across prolate and oblate spheroids. Similar approximations might extend to irregular shapes, but numerical validation would likely remain preferable due to the complexity of internal field distributions.
|
[
{
"version": "v1",
"created": "Tue, 27 May 2025 20:56:11 GMT"
},
{
"version": "v2",
"created": "Tue, 26 Aug 2025 10:27:45 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 09:45:41 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Pakarinen",
"Tomppa",
""
]
] |
2505.22482
|
Dani\"el Cox
|
Dani\"el W.S. Cox, Harish Sasikumar and Ivo M. Vellekoop
|
Inline calibration of spatial light modulators in nonlinear microscopy
|
10 pages, 5 figures
| null |
10.1364/AO.567871
| null |
physics.optics
|
http://creativecommons.org/licenses/by/4.0/
|
We present a method for calibrating the response of a phase-only spatial light modulator in nonlinear microscopy. Our method uses the microscope image itself as calibration measurement and requires no additional hardware components. Our method is adapted to the nonlinear signals encountered in multi-photon excitation fluorescence microscopes, and works well even under low light conditions and with strong photobleaching.
|
[
{
"version": "v1",
"created": "Wed, 28 May 2025 15:31:30 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Cox",
"Daniël W. S.",
""
],
[
"Sasikumar",
"Harish",
""
],
[
"Vellekoop",
"Ivo M.",
""
]
] |
2505.22874
|
Navindra Singh
|
Navindra D. Singh, James Leung, Ji Feng, Alma K. Gonz\'alez-Alcalde, Arial Tolentino, David Tuft, Juchen Guo, and Luat T. Vuong
|
Spectrum Selective Interfaces and Materials towards Non-photothermal Saltwater Evaporation: Demonstration with a White Ceramic Wick
|
18 pages, 5 figures, submitted to ACS Advanced Materials & Interfaces, 73 references
| null | null | null |
cond-mat.mtrl-sci physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Most solar desalination efforts are photothermal: they evaporate water with ``black'' materials that absorb as much sunlight as possible. Such ``brine-boiling'' methods are limited by the high thermal mass of water, i.e., its capacity to store and release heat. Here, we study the light-enhanced evaporation by a hard, white, aluminum nitride wick, and propose a route to selectively target salt-water bonds instead of bulk heating via deep-UV interactions. Through experiments and analyses that isolate the effects of light absorption and heating in aluminum nitride, we provide experimental evidence of a light-driven, spectrum-selective path to non-photothermal saltwater evaporation. Leverage of these light-matter interactions in white ceramic wicks may achieve low-cost, low-energy desalination, reduce the heat island effects of traditional solar technologies, and contribute to new cooling technologies where drought is also a concern.
|
[
{
"version": "v1",
"created": "Wed, 28 May 2025 21:16:44 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:02:55 GMT"
},
{
"version": "v3",
"created": "Mon, 30 Jun 2025 17:54:35 GMT"
},
{
"version": "v4",
"created": "Tue, 02 Sep 2025 19:35:10 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Singh",
"Navindra D.",
""
],
[
"Leung",
"James",
""
],
[
"Feng",
"Ji",
""
],
[
"González-Alcalde",
"Alma K.",
""
],
[
"Tolentino",
"Arial",
""
],
[
"Tuft",
"David",
""
],
[
"Guo",
"Juchen",
""
],
[
"Vuong",
"Luat T.",
""
]
] |
2505.23633
|
Long Zhang
|
Xiao-Dong Lin, Long Zhang
|
Measuring topological invariants of even-dimensional line-gapped non-Hermitian systems through quench dynamics
|
12 pages, 5 figures; complementary to arXiv: 2410.13241; to appear in PRB
| null | null | null |
cond-mat.quant-gas cond-mat.mes-hall quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The accurate determination of non-Hermitian (NH) topological invariants plays a central role in the study of NH topological phases. In this work, we propose a general framework for directly measuring NH topological invariants in even-dimensional systems with real line gaps through quench dynamics. Our approach hinges on constructing an auxiliary Hermitian matrix topologically equivalent to the original NH Hamiltonian, enabling topological characterization via reduced-dimensional momentum subspaces called band-inversion surfaces (BISs). A key insight lies in the emergence of chiral symmetry in the NH Hamiltonian specifically on BISs -- a critical property that allows extension of the dynamical characterization scheme previously developed for odd-dimensional NH systems with chiral or sublattice symmetry [Lin et al., Phys. Rev. Res. 7, L012060 (2025)]. We show that NH topological invariants can be extracted from the winding patterns of a dynamical field constructed from post-quench spin textures on BISs. We demonstrate our approach through a detailed analysis of NH Chern insulators and then extend the framework to higher even-dimensional systems by introducing second-order BISs for characterization. The framework is also generalized to imaginary line-gapped topological phases. This work establishes an experimentally accessible protocol for detecting NH topological invariants in quantum platforms.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 16:41:09 GMT"
},
{
"version": "v2",
"created": "Mon, 02 Jun 2025 09:48:03 GMT"
},
{
"version": "v3",
"created": "Wed, 04 Jun 2025 14:12:11 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 16:11:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Lin",
"Xiao-Dong",
""
],
[
"Zhang",
"Long",
""
]
] |
2505.23643
|
Boris K\"opf
|
Manuel Costa, Boris K\"opf, Aashish Kolluri, Andrew Paverd, Mark Russinovich, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Santiago Zanella-B\'eguelin
|
Securing AI Agents with Information-Flow Control
| null | null | null | null |
cs.CR cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
As AI agents become increasingly autonomous and capable, ensuring their security against vulnerabilities such as prompt injection becomes critical. This paper explores the use of information-flow control (IFC) to provide security guarantees for AI agents. We present a formal model to reason about the security and expressiveness of agent planners. Using this model, we characterize the class of properties enforceable by dynamic taint-tracking and construct a taxonomy of tasks to evaluate security and utility trade-offs of planner designs. Informed by this exploration, we present Fides, a planner that tracks confidentiality and integrity labels, deterministically enforces security policies, and introduces novel primitives for selectively hiding information. Its evaluation in AgentDojo demonstrates that this approach enables us to complete a broad range of tasks with security guarantees. A tutorial to walk readers through the the concepts introduced in the paper can be found at https://github.com/microsoft/fides
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 16:50:41 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 12:30:47 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Costa",
"Manuel",
""
],
[
"Köpf",
"Boris",
""
],
[
"Kolluri",
"Aashish",
""
],
[
"Paverd",
"Andrew",
""
],
[
"Russinovich",
"Mark",
""
],
[
"Salem",
"Ahmed",
""
],
[
"Tople",
"Shruti",
""
],
[
"Wutschitz",
"Lukas",
""
],
[
"Zanella-Béguelin",
"Santiago",
""
]
] |
2505.23980
|
Bayu Adhi Tama
|
Bayu Adhi Tama, Mansa Krishna, Homayra Alam, Mostafa Cham, Omar Faruque, Gong Cheng, Jianwu Wang, Mathieu Morlighem, Vandana Janeja
|
DeepTopoNet: A Framework for Subglacial Topography Estimation on the Greenland Ice Sheets
|
Accepted as Full Application Track Paper in SIGSPATIAL 2025
| null | null | null |
cs.CV cs.LG eess.IV
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding Greenland's subglacial topography is critical for projecting the future mass loss of the ice sheet and its contribution to global sea-level rise. However, the complex and sparse nature of observational data, particularly information about the bed topography under the ice sheet, significantly increases the uncertainty in model projections. Bed topography is traditionally measured by airborne ice-penetrating radar that measures the ice thickness directly underneath the aircraft, leaving data gap of tens of kilometers in between flight lines. This study introduces a deep learning framework, which we call as DeepTopoNet, that integrates radar-derived ice thickness observations and BedMachine Greenland data through a novel dynamic loss-balancing mechanism. Among all efforts to reconstruct bed topography, BedMachine has emerged as one of the most widely used datasets, combining mass conservation principles and ice thickness measurements to generate high-resolution bed elevation estimates. The proposed loss function adaptively adjusts the weighting between radar and BedMachine data, ensuring robustness in areas with limited radar coverage while leveraging the high spatial resolution of BedMachine predictions i.e. bed estimates. Our approach incorporates gradient-based and trend surface features to enhance model performance and utilizes a CNN architecture designed for subgrid-scale predictions. By systematically testing on the Upernavik Isstr{\o}m) region, the model achieves high accuracy, outperforming baseline methods in reconstructing subglacial terrain. This work demonstrates the potential of deep learning in bridging observational gaps, providing a scalable and efficient solution to inferring subglacial topography.
|
[
{
"version": "v1",
"created": "Thu, 29 May 2025 20:12:27 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 23:41:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tama",
"Bayu Adhi",
""
],
[
"Krishna",
"Mansa",
""
],
[
"Alam",
"Homayra",
""
],
[
"Cham",
"Mostafa",
""
],
[
"Faruque",
"Omar",
""
],
[
"Cheng",
"Gong",
""
],
[
"Wang",
"Jianwu",
""
],
[
"Morlighem",
"Mathieu",
""
],
[
"Janeja",
"Vandana",
""
]
] |
2505.24698
|
Andres Munoz-Arcentales Ph. D.
|
Rodrigo Men\'endez, Andres Munoz-Arcentales, Joaqu\'in Salvach\'ua, Carlos Aparicio, Irene Plaza, Gabriel Huecas
|
Next Generation Authentication for Data Spaces: An Authentication Flow Based On Grant Negotiation And Authorization Protocol For Verifiable Presentations (GNAP4VP)
|
Accepted in The 22nd International Conference on Mobile Systems and Pervasive Computing (MobiSPC) \\ August 4-6, 2025, Leuven, Belgium
| null |
10.1016/j.procs.2025.07.176
|
Procedia Computer Science Volume 265, 2025, Pages 226-235
|
cs.CR cs.ET
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Identity verification in Data Spaces is a fundamental aspect of ensuring security and privacy in digital environments. This paper presents an identity verification protocol tailored for shared data environments within Data Spaces. This protocol extends the Grant Negotiation and Authorization Protocol (GNAP) and integrates OpenID Connect for Verifiable Presentations (OIDC4VP) along with support for Linked Verifiable Presentations (LVP), providing a robust foundation for secure and privacy-preserving interactions. The proposed solution adheres to the principles of Self-Sovereign Identity (SSI) to facilitate decentralized, user-centric identity management while maintaining flexibility through protocol negotiation. Two alternative interaction flows are introduced: a "Wallet-Driven Interaction" utilizing OIDC4VP, and a "LVP Authorization" model for fully automated machine-to-machine communication. These flows address critical challenges encountered in Data Spaces, including privacy, interoperability, and regulatory compliance while simultaneously ensuring scalability and minimizing trust assumptions. The paper provides a detailed technical design, outlining the implementation considerations, and demonstrating how the proposed flows guarantee verifiable, secure, and efficient interactions between participants. This work contributes towards the establishment of a more trustworthy and sovereign digital infrastructure, in alignment with emerging European data governance initiatives.
|
[
{
"version": "v1",
"created": "Fri, 30 May 2025 15:20:39 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Menéndez",
"Rodrigo",
""
],
[
"Munoz-Arcentales",
"Andres",
""
],
[
"Salvachúa",
"Joaquín",
""
],
[
"Aparicio",
"Carlos",
""
],
[
"Plaza",
"Irene",
""
],
[
"Huecas",
"Gabriel",
""
]
] |
2506.01211
|
William B. James
|
William B. James
|
Iola Walker: A Mobile Footfall Detection System for Music Composition
| null | null | null | null |
cs.MM eess.AS
|
http://creativecommons.org/licenses/by/4.0/
|
This outing is part of a larger music technology research project. The objective is to find a method for materially enhancing music using hardware and software. There is a strong likelihood that there exists a new medium for experiencing music via a wearable device that ordinary listeners prefer over the current state of the art. If such a medium is discovered, it is a step towards altruistic, prosocial reform in the music industry. A new playback system infrastructure has a chance to soothe some of the societal problems tied to the larger entertainment industry ecosystem. Iola walker is a music playback system that allows musicians to compose music that changes in accordance with the listener's gait. Artifacts are available here: https://github.com/willbjames/iolawalker
|
[
{
"version": "v1",
"created": "Sun, 01 Jun 2025 23:13:46 GMT"
},
{
"version": "v2",
"created": "Tue, 05 Aug 2025 23:33:47 GMT"
},
{
"version": "v3",
"created": "Sun, 10 Aug 2025 19:16:44 GMT"
},
{
"version": "v4",
"created": "Tue, 02 Sep 2025 19:19:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"James",
"William B.",
""
]
] |
2506.01269
|
Hansung Choi
|
Hansung Choi and Daewon Seo
|
Region-of-Interest-Guided Deep Joint Source-Channel Coding for Image Transmission
| null | null | null | null |
cs.IT math.IT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Deep joint source-channel coding (deepJSCC) methods have shown promising improvements in communication performance over wireless networks. However, existing approaches primarily focus on enhancing overall image reconstruction quality, which may not fully align with user experiences, often driven by the quality of regions of interest (ROI). Motivated by this, we propose ROI-guided joint source-channel coding (ROI-JSCC), a novel deepJSCC framework that prioritizes high-quality transmission of ROI. The ROI-JSCC consists of four key components: (1) Image ROI embedding, (2) ROI-guided split processing, (3) ROI-based loss function design, and (4) ROI-adaptive bandwidth allocation. Together, these components allow ROI-JSCC to selectively enhance the ROI reconstruction quality at varying ROI positions while maintaining overall image quality with minimal computational overhead. Experimental results under diverse communication environments demonstrate that ROI-JSCC significantly improves ROI reconstruction quality while maintaining competitive average image quality compared to recent state-of-the-art methods.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 02:42:50 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 05:20:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Choi",
"Hansung",
""
],
[
"Seo",
"Daewon",
""
]
] |
2506.01326
|
Zhiyuan Wang
|
Zhiyuan Wang, Bokui Chen, Yinya Huang, Qingxing Cao, Ming He, Jianping Fan, Xiaodan Liang
|
ORMind: A Cognitive-Inspired End-to-End Reasoning Framework for Operations Research
|
Accepted by Annual Meetings of the Association for Computational Linguistics 2025
| null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Operations research (OR) is widely deployed to solve critical decision-making problems with complex objectives and constraints, impacting manufacturing, logistics, finance, and healthcare outcomes. While Large Language Models (LLMs) have shown promising results in various domains, their practical application in industry-relevant operations research (OR) problems presents significant challenges and opportunities. Preliminary industrial applications of LLMs for operations research face two critical deployment challenges: 1) Self-correction focuses on code syntax rather than mathematical accuracy, causing costly errors; 2) Complex expert selection creates unpredictable workflows that reduce transparency and increase maintenance costs, making them impractical for time-sensitive business applications. To address these business limitations, we introduce ORMind, a cognitive-inspired framework that enhances optimization through counterfactual reasoning. Our approach emulates human cognition, implementing an end-to-end workflow that systematically transforms requirements into mathematical models and executable solver code. It is currently being tested internally in Lenovo's AI Assistant, with plans to enhance optimization capabilities for both business and consumer customers. Experiments demonstrate that ORMind outperforms existing methods, achieving a 9.5\% improvement on the NL4Opt dataset and a 14.6\% improvement on the ComplexOR dataset.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 05:11:21 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 12:09:34 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Wang",
"Zhiyuan",
""
],
[
"Chen",
"Bokui",
""
],
[
"Huang",
"Yinya",
""
],
[
"Cao",
"Qingxing",
""
],
[
"He",
"Ming",
""
],
[
"Fan",
"Jianping",
""
],
[
"Liang",
"Xiaodan",
""
]
] |
2506.01729
|
Jad Wehbeh
|
J. Wehbeh, E. C. Kerrigan
|
Update-Aware Robust Optimal Model Predictive Control for Nonlinear Systems
|
6 pages, 2 figures, published in the IEEE Control System Letters (2025)
| null |
10.1109/LCSYS.2025.3576534
| null |
eess.SY cs.SY math.OC
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Robust optimal or min-max model predictive control (MPC) approaches aim to guarantee constraint satisfaction over a known, bounded uncertainty set while minimizing a worst-case performance bound. Traditionally, these methods compute a trajectory that meets the desired properties over a fixed prediction horizon, apply a portion of the resulting input, and then re-solve the MPC problem using newly obtained measurements at the next time step. However, this approach fails to account for the fact that the control trajectory will be updated in the future, potentially leading to conservative designs. In this paper, we present a novel update-aware robust optimal MPC algorithm for decreasing horizon problems on nonlinear systems that explicitly accounts for future control trajectory updates. This additional insight allows our method to provably expand the feasible solution set and guarantee improved worst-case performance bounds compared to existing techniques. Our approach formulates the trajectory generation problem as a sequence of nested existence-constrained semi-infinite programs (SIPs), which can be efficiently solved using local reduction techniques. To demonstrate its effectiveness, we evaluate our approach on a planar quadrotor problem, where it clearly outperforms an equivalent method that does not account for future updates at the cost of increased computation time.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 14:40:53 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:17:14 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Wehbeh",
"J.",
""
],
[
"Kerrigan",
"E. C.",
""
]
] |
2506.01742
|
Jad Wehbeh
|
J. Wehbeh, E. C. Kerrigan
|
Smooth Logic Constraints in Nonlinear Optimization and Optimal Control Problems
|
6 pages, 7 figures, accepted for publication at the 2025 IEEE Conference on Decision and Control
| null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
In some optimal control problems, complex relationships between states and inputs cannot be easily represented using continuous constraints, necessitating the use of discrete logic instead. This paper presents a method for incorporating such logic constraints directly within continuous optimization frameworks, eliminating the need for binary variables or specialized solvers. Our approach reformulates arbitrary logic constraints under minimal assumptions as max-min constraints, which are then smoothed by introducing auxiliary variables into the optimization problem. When these reformulated constraints are satisfied, they guarantee that the original logical conditions hold, ensuring correctness in the optimization process. We demonstrate the effectiveness of this method on two planar quadrotor control tasks with complex logic constraints. Compared to existing techniques for encoding logic in continuous optimization, our approach achieves faster computational performance and improved convergence to feasible solutions.
|
[
{
"version": "v1",
"created": "Mon, 02 Jun 2025 14:50:17 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:17:57 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Wehbeh",
"J.",
""
],
[
"Kerrigan",
"E. C.",
""
]
] |
2506.02716
|
Sudin Ganguly
|
Sourav Karmakar, Sudin Ganguly, and Santanu K. Maiti
|
Reentrant localization in a quasiperiodic chain with correlated hopping sequences
|
7 pages, 5 figures. Comments are welcome
| null |
10.1088/1361-648X/ae025a
| null |
cond-mat.dis-nn cond-mat.mes-hall cond-mat.str-el physics.comp-ph physics.data-an
|
http://creativecommons.org/licenses/by/4.0/
|
Quasiperiodic systems are known to exhibit localization transitions in low dimensions, wherein all electronic states become localized beyond a critical disorder strength. Interestingly, recent studies have uncovered a reentrant localization (RL) phenomenon: upon further increasing the quasiperiodic modulation strength beyond the localization threshold, a subset of previously localized states can become delocalized again within a specific parameter window. While RL transitions have been primarily explored in systems with simple periodic modulations, such as dimerized or long-range hopping integrals, the impact of more intricate or correlated hopping structures on RL behavior remains largely elusive. In this work, we investigate the localization behavior in a one-dimensional lattice featuring staggered, correlated on-site potentials following the Aubry-Andr\'{e}-Harper model, along with off-diagonal hopping modulations structured according to quasiperiodic Fibonacci and Bronze Mean sequences. By systematically analyzing the fractal dimension, inverse participation ratio, and normalized participation ratio, we demonstrate the occurrence of RL transitions induced purely by the interplay between quasiperiodic on-site disorder and correlated hopping. We further examine the parameter space to determine the specific regimes that give rise to RL. Our findings highlight the crucial role of underlying structural correlations in governing localization-delocalization transitions in low-dimensional quasiperiodic systems, where the correlated disorder manifests in both diagonal and off-diagonal terms.
|
[
{
"version": "v1",
"created": "Tue, 03 Jun 2025 10:15:19 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 12:40:57 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Karmakar",
"Sourav",
""
],
[
"Ganguly",
"Sudin",
""
],
[
"Maiti",
"Santanu K.",
""
]
] |
2506.03590
|
Minh Luu
|
Minh Luu, Surya Jasper, Khoi Le, Evan Pan, Michael Quinn, Aakash Tyagi, Jiang Hu
|
VCDiag: Classifying Erroneous Waveforms for Failure Triage Acceleration
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Failure triage in design functional verification is critical but time-intensive, relying on manual specification reviews, log inspections, and waveform analyses. While machine learning (ML) has improved areas like stimulus generation and coverage closure, its application to RTL-level simulation failure triage, particularly for large designs, remains limited. VCDiag offers an efficient, adaptable approach using VCD data to classify failing waveforms and pinpoint likely failure locations. In the largest experiment, VCDiag achieves over 94% accuracy in identifying the top three most likely modules. The framework introduces a novel signal selection and statistical compression approach, achieving over 120x reduction in raw data size while preserving features essential for classification. It can also be integrated into diverse Verilog/SystemVerilog designs and testbenches.
|
[
{
"version": "v1",
"created": "Wed, 04 Jun 2025 05:44:03 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Jul 2025 02:56:18 GMT"
},
{
"version": "v3",
"created": "Thu, 24 Jul 2025 06:09:00 GMT"
},
{
"version": "v4",
"created": "Fri, 15 Aug 2025 18:20:05 GMT"
},
{
"version": "v5",
"created": "Wed, 03 Sep 2025 17:48:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Luu",
"Minh",
""
],
[
"Jasper",
"Surya",
""
],
[
"Le",
"Khoi",
""
],
[
"Pan",
"Evan",
""
],
[
"Quinn",
"Michael",
""
],
[
"Tyagi",
"Aakash",
""
],
[
"Hu",
"Jiang",
""
]
] |
2506.05052
|
Xiao-Dong Ma
|
Yi Liao, Xiao-Dong Ma, and Hao-Lin Wang
|
Chiral perturbation theory for baryon-number-violating nucleon decay into a vector meson
|
v2: minor revisions, 6 pages + supplemental material, matched the published version as a Letter in PRD
|
Phys. Rev. D 112, L031704 (2025)
|
10.1103/fzq9-tfcp
| null |
hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
In a recent work [New chiral structures for baryon number violating nucleon decays, arXiv:2504.14855], we identified generic baryon-number-violating (BNV) structures containing triple light quarks and achieved their leading-order chiral realizations involving octet pseudoscalars and baryons. Although many two-body nucleon decays into a vector meson have been experimentally searched for and stringently constrained, a consistent theoretical framework for their calculation is still lacking. In this Letter, we fill the gap by implementing chiral matching of all these triple-quark interactions onto hadronic interactions involving octet vector mesons, baryons, and pseudoscalars. This paves the way for a consistent and comprehensive study of all relevant BNV processes. As an illustration of application, we show how degeneracy in the parameter space of Wilson coefficients can be broken by synthesizing experimental constraints on nucleon decays into a vector or pseudoscalar meson when relevant hadronic low-energy constants can be reasonably determined.
|
[
{
"version": "v1",
"created": "Thu, 05 Jun 2025 13:58:05 GMT"
},
{
"version": "v2",
"created": "Mon, 01 Sep 2025 02:56:13 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Liao",
"Yi",
""
],
[
"Ma",
"Xiao-Dong",
""
],
[
"Wang",
"Hao-Lin",
""
]
] |
2506.05657
|
Anarya Ray
|
Anarya Ray
|
Emulating compact binary population synthesis simulations with uncertainty quantification and model comparison using Bayesian normalizing flows
|
16 pages, 4 figures
| null | null | null |
astro-ph.HE cs.LG gr-qc
|
http://creativecommons.org/licenses/by/4.0/
|
Population synthesis simulations of compact binary coalescences~(CBCs) play a crucial role in extracting astrophysical insights from an ensemble of gravitational wave~(GW) observations. However, realistic simulations can be costly to implement for a dense grid of initial conditions. Normalizing flows can emulate population synthesis runs to enable simulation-based inference from observed catalogs and data augmentation for feature prediction in rarely synthesizable sub-populations. However, flow predictions can be wrought with uncertainties, especially for sparse training sets. In this work, we develop a method for quantifying and marginalizing uncertainties in the emulators by implementing the Bayesian Normalizing flow, a conditional density estimator constructed from Bayesian neural networks. Using the exact likelihood function naturally associated with density estimators, we sample the posterior distribution of flow parameters with suitably chosen priors to quantify and marginalize over flow uncertainties. We demonstrate the accuracy, calibration, inference, and data-augmentation impacts of the estimated uncertainties for simulations of binary black hole populations formed through common envelope evolution. We outline the applications of the proposed methodology in the context of simulation-based inference from growing GW catalogs and feature prediction, with state-of-the-art binary evolution simulators, now marginalized over model and data uncertainties.
|
[
{
"version": "v1",
"created": "Fri, 06 Jun 2025 01:17:52 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 17:47:46 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Ray",
"Anarya",
""
]
] |
2506.05668
|
Jiajun He
|
Jiajun He, Jos\'e Miguel Hern\'andez-Lobato, Yuanqi Du, Francisco Vargas
|
RNE: plug-and-play diffusion inference-time control and energy-based training
|
48 pages; 15 figures; Add more experiments on energy-based training, fix several typos and an error in RNC-TDS paragraph
| null | null | null |
cs.LG stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Diffusion models generate data by removing noise gradually, which corresponds to the time-reversal of a noising process. However, access to only the denoising kernels is often insufficient. In many applications, we need the knowledge of the marginal densities along the generation trajectory, which enables tasks such as inference-time control. To address this gap, in this paper, we introduce the Radon-Nikodym Estimator (RNE). Based on the concept of the density ratio between path distributions, it reveals a fundamental connection between marginal densities and transition kernels, providing a flexible plug-and-play framework that unifies diffusion density estimation, inference-time control, and energy-based diffusion training under a single perspective. Experiments demonstrated that RNE delivers strong results in inference-time control applications, such as annealing and model composition, with promising inference-time scaling performance. Moreover, RNE provides a simple yet efficient regularisation for training energy-based diffusion.
|
[
{
"version": "v1",
"created": "Fri, 06 Jun 2025 01:32:14 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Jun 2025 16:51:44 GMT"
},
{
"version": "v3",
"created": "Sun, 15 Jun 2025 21:26:55 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 13:30:36 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"He",
"Jiajun",
""
],
[
"Hernández-Lobato",
"José Miguel",
""
],
[
"Du",
"Yuanqi",
""
],
[
"Vargas",
"Francisco",
""
]
] |
2506.07625
|
Steven Finch
|
Steven Finch
|
Half-Iterates and Delta Conjectures
|
Two distinct earlier preprints are relevant. The first covers half-iterates of $x(1+x)$, sin$(x)$ & exp$(x/e)$ and appears at arXiv:2506.07625v1. The second covers half-iterates of $x$exp$(x)$, $x+1/x$ & arcsinh$(x)$ and appears at arXiv:2506.07625v2. Reading these will help to motivate the study of $\delta$ in the current paper. 9 pages; 2 figures
| null | null | null |
math.NT cs.DM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The vivid contrast between two competing algorithms for solving Abel's equation $g(\theta(x)) = g(x) + 1$, given $\theta(x)$, is easily sketched. EJ is faster and more efficient, but ML evaluates a limit characterizing the principal solution $g(x)$ directly. EJ finds $g(x)+\delta$, where $\delta$ is possibly nonzero but independent of $x$. If we were to know an exact expression for $\delta$, then the "intrinsicality" of ML would be subsumed by EJ. Filling this gap in our knowledge is the aim of this paper.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 10:42:45 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 02:12:58 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Jul 2025 17:14:48 GMT"
},
{
"version": "v4",
"created": "Wed, 13 Aug 2025 12:58:44 GMT"
},
{
"version": "v5",
"created": "Tue, 26 Aug 2025 02:24:00 GMT"
},
{
"version": "v6",
"created": "Wed, 03 Sep 2025 13:32:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Finch",
"Steven",
""
]
] |
2506.07940
|
Chris Subia-Waud
|
Christopher Subia-Waud (Rayonlabs Team)
|
Gradients: When Markets Meet Fine-tuning -- A Distributed Approach to Model Optimisation
| null | null | null | null |
cs.AI cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Current AutoML platforms leave substantial performance untapped. Testing 180 fine-tuning tasks across models from 70M to 70B parameters, we found that HuggingFace AutoTrain, TogetherAI, Databricks, and Google Cloud consistently produce suboptimal configurations. Gradients, built on the Bittensor network, attacks this problem through competition. Independent miners race to find optimal hyperparameters, earning rewards proportional to their models' performance. This tournament drives exploration of configuration spaces that single-strategy methods never examine. In our experiments, Gradients achieved a 100\% win rate against TogetherAI, Databricks, and Google Cloud, and beat HuggingFace AutoTrain in 82.8\% of experiments. Mean improvements reached 42.1\% against commercial platforms. Retrieval-augmented generation tasks saw 30-40\% gains; diffusion models improved 23.4\% on person-specific generation. When miners compete for rewards, they develop optimization strategies that centralized approaches overlook. These findings demonstrate that decentralized systems with economic incentives can systematically outperform traditional AutoML, suggesting market dynamics may be key to achieving superior fine-tuning results. Code is available at https://github.com/rayonlabs/G.O.D.
|
[
{
"version": "v1",
"created": "Mon, 09 Jun 2025 17:00:38 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 09:39:22 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Subia-Waud",
"Christopher",
"",
"Rayonlabs Team"
]
] |
2506.08489
|
Dibya Chakraborty
|
Dibya Chakraborty, Suratna Das
|
Behaviour of $\alpha$-attractors in Warm Inflation
|
20 pages, 13 figures, matches published version in PRD
| null | null | null |
astro-ph.CO gr-qc hep-ph hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
The $\alpha$-attractor models of inflation have remained one of the preferred inflationary models for nearly a decade now. The unique attractor nature of these models in the $n_s-r$ plane have put these models in the sweet-spot of the $n_s-r$ measurement of Planck observations. In this article, we analyse the behaviour of such attractor models in a Warm Inflation setup to investigate whether the attractor nature of these models can be retained even when the dynamics deviates from the standard Cold inflationary dynamics. We have chosen to analyse these models in a strongly dissipative Warm inflation setup, namely the Minimal Warm Inflation, as in such a setup the inflationary dynamics significantly departs from the standard Cold inflation dynamics. We observe that the departure from the standard Cold inflation dynamics destroys the unique attractor nature of such models. The analysis clearly indicates that the attractor nature of these $\alpha$-attractor models is quite unique to the standard Cold inflationary dynamics. However, on a positive note, the analysis indicates that the $\alpha$-attractor models may be made in tune with the recent ACT results in Warm inflation for certain parameter ranges.
|
[
{
"version": "v1",
"created": "Tue, 10 Jun 2025 06:29:50 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 06:23:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chakraborty",
"Dibya",
""
],
[
"Das",
"Suratna",
""
]
] |
2506.08731
|
Eleni-Rosalina Andrinopoulou
|
Eleni-Rosalina Andrinopoulou, Emrah Gecili and Rhonda D Szczesniak
|
Unveiling the Impact of Social and Environmental Determinants of Health on Lung Function Decline in Cystic Fibrosis through Data Integration using the US Registry
| null | null | null | null |
stat.AP
|
http://creativecommons.org/licenses/by/4.0/
|
Integrating diverse data sources offers a comprehensive view of patient health and holds potential for improving clinical decision-making. In Cystic Fibrosis (CF), which is a genetic disorder primarily affecting the lungs, biomarkers that track lung function decline such as FEV1 serve as important predictors for assessing disease progression. Prior research has shown that incorporating social and environmental determinants of health improves prognostic accuracy. To investigate the lung function decline among individuals with CF, we integrate data from the U.S. Cystic Fibrosis Foundation Patient Registry with social and environmental health information. Our analysis focuses on the relationship between lung function and the deprivation index, a composite measure of socioeconomic status. We used advanced multivariate mixed-effects models, which allow for the joint modelling of multiple longitudinal outcomes with flexible functional forms. This methodology provides an understanding of interrelationships among outcomes, addressing the complexities of dynamic health data. We examine whether this relationship varies with patients' exposure duration to high-deprivation areas, analyzing data across time and within individual US states. Results show a strong relation between lung function and the area under the deprivation index curve across all states. These results underscore the importance of integrating social and environmental determinants of health into clinical models of disease progression. By accounting for broader contextual factors, healthcare providers can gain deeper insights into disease trajectories and design more targeted intervention strategies.
|
[
{
"version": "v1",
"created": "Tue, 10 Jun 2025 12:29:07 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Jul 2025 11:36:25 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 08:46:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Andrinopoulou",
"Eleni-Rosalina",
""
],
[
"Gecili",
"Emrah",
""
],
[
"Szczesniak",
"Rhonda D",
""
]
] |
2506.09154
|
Fabio Antonini Dr
|
Fabio Antonini, Thomas Callister, Fani Dosopoulou, Isobel Romero-Shaw, Debatri Chattopadhyay
|
Inferring the pair-instability mass gap from gravitational wave data
|
Accepted by PRD
| null | null | null |
astro-ph.HE astro-ph.SR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We use hierarchical Bayesian inference with non-parametric Gaussian process models to investigate the effective inspiral spin parameter, $\chi_{\rm eff}$, as a function of primary black hole mass in the third gravitational-wave transient catalog (GWTC-3). Our analysis reveals a transition in the population at a primary mass of $46^{+7}_{-5}\,M_\odot$. Beyond this mass, the $\chi_{\rm eff}$ distribution broadens, becomes consistent with being symmetric around zero, and has a median of $-0.03^{+0.36}_{-0.59}$ (90\% credibility). These results are consistent with the presence of a pair-instability mass gap that is repopulated by black holes that are the remnant of a previous merger, formed in dense star clusters. However, asymmetric distributions skewed toward positive $\chi_{\rm eff}$ are not excluded by current data. Below the inferred transition mass, we constrain the fraction of second-generation black holes to be $\lesssim 10\%$. These results provide model-independent support for a high-mass and high-spin population of black holes in the data, consistent with earlier work using parametric models. Imminent gravitational-wave data releases will be essential to sharpen constraints on spin symmetry and clarify the origin of the black holes.
|
[
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:06:29 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 07:39:51 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Antonini",
"Fabio",
""
],
[
"Callister",
"Thomas",
""
],
[
"Dosopoulou",
"Fani",
""
],
[
"Romero-Shaw",
"Isobel",
""
],
[
"Chattopadhyay",
"Debatri",
""
]
] |
2506.09537
|
Hiromasa Takaura
|
Hiromasa Takaura, Robert V. Harlander, Fabian Lange
|
A new approach to quark mass determination using the gradient flow
|
33 pages, 5 figures, 2 ancillary files, published version, minor modifications
| null | null |
YITP-25-89, P3H-25-037, TTK-25-15, ZU-TH 42/25
|
hep-lat hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a new method to determine quark masses using ratios of the vacuum-expectation values (VEVs) of flowed quark bilinear operators. They can be expressed as functions of the flow time $t$ and the ${\overline {\rm MS}}$ quark mass $\overline{m}$, which can then be determined by matching with the corresponding lattice results. Motivated by this, we evaluate these VEVs perturbatively through next-to-leading order in the strong coupling. We provide the results as expansions in the limits of small and large $\overline{m}^2 t$. To this end, we develop a new expansion technique based on the Laplace transform. Additionally, we present numerical results with the exact mass dependence over a wide range of $\overline{m}^2t$. We discuss the expected perturbative precision for the mass determination based on our next-to-leading order perturbative calculations, and possible non-perturbative corrections.
|
[
{
"version": "v1",
"created": "Wed, 11 Jun 2025 09:13:31 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 12:49:03 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Takaura",
"Hiromasa",
""
],
[
"Harlander",
"Robert V.",
""
],
[
"Lange",
"Fabian",
""
]
] |
2506.09710
|
Hiroyasu Satoh
|
Hiroyasu Satoh, Hemangi Madhusudan Shah
|
Conformal Vector Fields on Complex Hyperbolic Space
|
23 pages
| null | null | null |
math.DG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this article, we show that there does not exist any non-trivial conformal vector field on complex hyperbolic space. That is, we show that any conformal vector field on a complex hyperbolic space is indeed a Killing vector field.
|
[
{
"version": "v1",
"created": "Wed, 11 Jun 2025 13:23:14 GMT"
},
{
"version": "v2",
"created": "Sat, 16 Aug 2025 15:39:49 GMT"
},
{
"version": "v3",
"created": "Tue, 02 Sep 2025 23:27:06 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Satoh",
"Hiroyasu",
""
],
[
"Shah",
"Hemangi Madhusudan",
""
]
] |
2506.09785
|
Alexander Marusov
|
Alexander Marusov, Aleksandr Yugay, Alexey Zaytsev
|
A theoretical framework for self-supervised contrastive learning for continuous dependent data
| null | null | null | null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Self-supervised learning (SSL) has emerged as a powerful approach to learning representations, particularly in the field of computer vision. However, its application to dependent data, such as temporal and spatio-temporal domains, remains underexplored. Besides, traditional contrastive SSL methods often assume \emph{semantic independence between samples}, which does not hold for dependent data exhibiting complex correlations. We propose a novel theoretical framework for contrastive SSL tailored to \emph{continuous dependent data}, which allows the nearest samples to be semantically close to each other. In particular, we propose two possible \textit{ground truth similarity measures} between objects -- \emph{hard} and \emph{soft} closeness. Under it, we derive an analytical form for the \textit{estimated similarity matrix} that accommodates both types of closeness between samples, thereby introducing dependency-aware loss functions. We validate our approach, \emph{Dependent TS2Vec}, on temporal and spatio-temporal downstream problems. Given the dependency patterns presented in the data, our approach surpasses modern ones for dependent data, highlighting the effectiveness of our theoretically grounded loss functions for SSL in capturing spatio-temporal dependencies. Specifically, we outperform TS2Vec on the standard UEA and UCR benchmarks, with accuracy improvements of $4.17$\% and $2.08$\%, respectively. Furthermore, on the drought classification task, which involves complex spatio-temporal patterns, our method achieves a $7$\% higher ROC-AUC score.
|
[
{
"version": "v1",
"created": "Wed, 11 Jun 2025 14:23:47 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 09:59:25 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 15:45:50 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Marusov",
"Alexander",
""
],
[
"Yugay",
"Aleksandr",
""
],
[
"Zaytsev",
"Alexey",
""
]
] |
2506.10285
|
Archishna Bhattacharyya
|
Archishna Bhattacharyya
|
Sequential transmission at short times
|
18 pages, no figures, v3: minor updates
| null | null | null |
quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We show that it is possible to transmit and preserve information at short time scales over an n-fold composition of quantum channels $(\Xi^n)_{n \in \mathbb{N}}$ modelled as a discrete quantum Markov semigroup, long enough to generate entanglement at some finite $n$. This is achieved by interspersing the action of noise with quantum error correction in succession. We show this by means of a non-trivial lower bound on the one-shot quantum capacity in the sequential setting as a function of $n$, in an attempt to model a linear quantum network and assess its capabilities to distribute entanglement. Intriguingly, the rate of transmission of such a network turns out to be a property of the spectrum of the channels composed in sequence, and the maximum possible error in transmission can be bounded as a function of the noise model only. As an application, we derive an exact error bound for the infinite dimensional pure-loss channel believed to be the dominant source of noise in networks precluding the distribution of entanglement. We exemplify our results by analysing the amplitude damping channel and its bosonic counterpart.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 01:48:59 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 20:20:38 GMT"
},
{
"version": "v3",
"created": "Tue, 02 Sep 2025 19:23:54 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Bhattacharyya",
"Archishna",
""
]
] |
2506.10733
|
Marco Zaro
|
Stefano Frixione, Fabio Maltoni, Davide Pagani, Marco Zaro
|
Precision phenomenology at multi-TeV muon colliders
|
48 pages, 18 figures; version accepted for publication
|
JHEP 09 (2025) 036
|
10.1007/JHEP09(2025)036
|
TIF-UNIMI-2025-12, IRMP-CP3-25-21, COMETA-2025-21
|
hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Future lepton colliders, such as those based on linear $e^+e^-$ or circular $\mu^+\mu^-$ accelerators, are expected to attain centre-of-mass energies in the multi-TeV range. In this regime the impact of QED and of weak radiation, in both the initial and the final state, can become a leading effect. By employing a general framework presented in a companion paper - suitable for any flavour of colliding leptons - we improve next-to-leading order electroweak predictions by including higher-order contributions, which encompass, but are not limited to, vector-boson-fusion processes. We apply this approach to the study of $t\bar{t}$ and $W^+W^-$ production at a muon collider operating at centre-of-mass energies up to 10 TeV. We show that such an approach, where both QED and weak contributions are included at fixed order, in addition to the all-order resummation of initial-state QED effects, can provide predictions for arbitrary observables in all of the phase space which are precise at the percent level.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 14:24:41 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Jun 2025 12:54:11 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 14:12:14 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Frixione",
"Stefano",
""
],
[
"Maltoni",
"Fabio",
""
],
[
"Pagani",
"Davide",
""
],
[
"Zaro",
"Marco",
""
]
] |
2506.10786
|
Ding-Fu Shao
|
Shui-Sen Zhang, Zi-An Wang, Bo Li, Wen-Jian Lu, Mingliang Tian, Yu-Ping Sun, Haifeng Du, Ding-Fu Shao
|
Deterministic Switching of the N\'{e}el Vector by Asymmetric Spin Torque
|
Added the case for field-assisted switching
| null | null | null |
cond-mat.mtrl-sci cond-mat.mes-hall
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
N\'{e}el vector, the order parameter of collinear antiferromagnets, serves as a state variable in associated antiferromagnetic (AFM) spintronic devices to encode information. A deterministic switching of N\'{e}el vector is crucial for the write-in operation, which, however, remains a challenging problem in AFM spintronics. Here we demonstrate, based on analytical derivation and macro-spin simulations, that N\'{e}el vector switching can be generally achieved via a current-induced spin torque, provided the spin accumulations responsible for this torque are non-identical between opposite sublattices. This condition occurs widely in AFM films, as symmetry equivalence between sublattice-dependent spin accumulations is usually absent, allowing unequal spin accumulations induced by Edelstein effect or a spin current. Unlike previously studied spin torques induced by uniform or staggered spin accumulations -- where either the field-like or damping-like component dominates exclusively -- the asymmetric spin torque features cooperative contributions from both components, leading to N\'{e}el vector dynamics that are fundamentally distinct from previous expectations. The switching conditions derived analytically agree well with simulation results and suggest various directions for further optimization. Our work establishes a general mechanism for current-induced N\'{e}el vector switching, which is in principle feasible for all collinear antiferromagnets, and thus paves the route to realize efficient writing in antiferromagnetic spintronics.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 14:58:18 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 06:37:24 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Zhang",
"Shui-Sen",
""
],
[
"Wang",
"Zi-An",
""
],
[
"Li",
"Bo",
""
],
[
"Lu",
"Wen-Jian",
""
],
[
"Tian",
"Mingliang",
""
],
[
"Sun",
"Yu-Ping",
""
],
[
"Du",
"Haifeng",
""
],
[
"Shao",
"Ding-Fu",
""
]
] |
2506.10926
|
Ivan Christov
|
SungGyu Chun, Ivan C. Christov, Jie Feng
|
Experimental investigation of the flow rate--pressure drop relation of a viscoelastic Boger fluid in a deformable channel
|
7 pages, 4 figures; v2: minor improvements, accepted for publication in Physical Review Applied
|
Phys. Rev. Applied 24 (2025) 034001
|
10.1103/hcfn-z67n
| null |
physics.flu-dyn cond-mat.soft
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study the steady flow-induced deformation between an incompressible non-Newtonian fluid and a three-dimensional (3D) deformable channel. Specifically, we provide a comprehensive experimental--theoretical framework for such flows of constant-viscosity viscoelastic (i.e., Boger) fluids, which allows us to quantify the influence of the fluid's viscoelasticity on the flow rate--pressure drop ($q-\Delta p$) relation. For a flow-rate-controlled regime and weakly viscoelastic flow, we find excellent agreement between the theoretical prediction and experimental results, specifically providing an experimental demonstration of how both wall compliance and fluid viscoelasticity decrease the pressure drop. To definitively demonstrate the coupled effect of wall compliance and fluid viscoelasticity (change of cross-sectional area and curving of streamlines), we perform experiments in an equivalent rigid channel, showing that both the Boger fluid and a viscosity-matched Newtonian fluid have the same pressure drop. Thus, the experimentally verified theory for the $q-\Delta p$ relation of weakly viscoelastic flows of Boger fluids in compliant channels can now be used for the design of microfluidics and soft hydraulic systems operating with complex fluids.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 17:35:56 GMT"
},
{
"version": "v2",
"created": "Mon, 18 Aug 2025 14:11:56 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Chun",
"SungGyu",
""
],
[
"Christov",
"Ivan C.",
""
],
[
"Feng",
"Jie",
""
]
] |
2506.11270
|
Raam Uzdin Dr.
|
Jader P. Santos, Raam Uzdin
|
Drift-resilient mid-circuit measurement and state preparation error mitigation for dynamic circuits
|
V2: New experiments and improved presentation
| null | null | null |
quant-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Quantum error mitigation (QEM) for dynamic circuits, i.e., those incorporating mid-circuit measurements and feedforward, is important for two key reasons. First, quantum error correction (QEC) circuits are instances of dynamic circuits, and therefore a dynamic circuit-compatible QEM can extend circuit depth and address errors that QEC struggles with. Second, recent studies show that dynamic circuits can significantly outperform purely unitary ones. However, mid-circuit measurement errors remain a major bottleneck. Current solutions rely on readout noise characterization that is vulnerable to temporal noise drifts. To the best of our knowledge, no readout mitigation schemes are resilient to temporal noise drifts. By introducing parity-based noise amplification in repeated measurements, we derive and experimentally demonstrate a drift-resilient protocol for addressing preparation, mid-circuit, and terminating measurement errors without requiring calibration or characterization. Drift resilience increases the longest possible execution time (in terms of shots) and enables flexibility by combining data from non-consecutive times. For platforms such as trapped ions, where the measurements are highly disruptive, we provide an alternative reset-based mitigation scheme. We demonstrate our methods experimentally on IBMQ and Quantinuum hardware. Combined with the Layered-KIK gate error mitigation protocol, the presented readout mitigation approach enables "End-to-end" mitigation for dynamic circuits, that can improve the outcomes of QEC experiments, and that covers the widest range of errors to the best of our knowledge. Other applications of the presented methods include a faster alternative to gate-set tomography and diagnostics of defective qubits during the execution of the target algorithm.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 20:22:09 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 17:23:26 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Santos",
"Jader P.",
""
],
[
"Uzdin",
"Raam",
""
]
] |
2506.11457
|
Bernard Parent
|
Felipe Martin Rodriguez Fuentes and Bernard Parent
|
Vibrational-Electron Heating in Plasma Flows: A Thermodynamically Consistent Model
|
18 pages, 16 figures
| null | null | null |
physics.plasm-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Accurate prediction of electron temperature ($T_{\rm e}$) in non-equilibrium {plasma} flows is critical, yet hampered by inadequate models for electron heating from vibrationally excited states. Prior models often relied on ad-hoc scaling or flawed applications of detailed balance that failed to ensure the convergence of electron temperature and species-specific vibrational temperature ($T_{\rm v}$) at thermal equilibrium. This paper introduces a novel, thermodynamically consistent electron heating model derived rigorously from the principle of detailed balance. By assuming a Boltzmann vibrational distribution and employing an effective activation energy, our approach yields a simple heating-to-cooling ratio of $\exp(\theta_{\rm v}/T_{\rm e}-\theta_{\rm v}/T_{\rm v})$, where $\theta_{\rm v}$ is the characteristic vibrational temperature of the species under consideration. This formulation guarantees that $T_{\rm e}$ correctly converges to $T_{\rm v}$ at equilibrium. A key advantage is that our model can utilize total cooling rates determined from swarm experiments, leading to higher accuracy at low electron temperatures. For re-entry flows, the proposed approach predicts an electron temperature several times lower than previous models which results in improved agreement with some flight test data. These more reliable predictions can significantly enhance the modeling fidelity of plasma-assisted combustion, laser-induced plasmas, and various hypersonic plasma technologies such as electron transpiration cooling or magnetohydrodynamic force generators.
|
[
{
"version": "v1",
"created": "Fri, 13 Jun 2025 04:28:10 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 02:31:17 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fuentes",
"Felipe Martin Rodriguez",
""
],
[
"Parent",
"Bernard",
""
]
] |
2506.12100
|
Reza Fayyazi
|
Reza Fayyazi, Michael Zuzak, Shanchieh Jay Yang
|
LLM Embedding-based Attribution (LEA): Quantifying Source Contributions to Generative Model's Response for Vulnerability Analysis
| null | null | null | null |
cs.CR cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large Language Models (LLMs) are increasingly used for cybersecurity threat analysis, but their deployment in security-sensitive environments raises trust and safety concerns. With over 21,000 vulnerabilities disclosed in 2025, manual analysis is infeasible, making scalable and verifiable AI support critical. When querying LLMs, dealing with emerging vulnerabilities is challenging as they have a training cut-off date. While Retrieval-Augmented Generation (RAG) can inject up-to-date context to alleviate the cut-off date limitation, it remains unclear how much LLMs rely on retrieved evidence versus the model's internal knowledge, and whether the retrieved information is meaningful or even correct. This uncertainty could mislead security analysts, mis-prioritize patches, and increase security risks. Therefore, this work proposes LLM Embedding-based Attribution (LEA) to analyze the generated responses for vulnerability exploitation analysis. More specifically, LEA quantifies the relative contribution of internal knowledge vs. retrieved content in the generated responses. We evaluate LEA on 500 critical vulnerabilities disclosed between 2016 and 2025, across three RAG settings -- valid, generic, and incorrect -- using three state-of-the-art LLMs. Our results demonstrate LEA's ability to detect clear distinctions between non-retrieval, generic-retrieval, and valid-retrieval scenarios with over 95% accuracy on larger models. Finally, we demonstrate the limitations posed by incorrect retrieval of vulnerability information and raise a cautionary note to the cybersecurity community regarding the blind reliance on LLMs and RAG for vulnerability analysis. LEA offers security analysts with a metric to audit RAG-enhanced workflows, improving the transparent and trustworthy deployment of AI in cybersecurity threat analysis.
|
[
{
"version": "v1",
"created": "Thu, 12 Jun 2025 21:20:10 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 17:18:40 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Fayyazi",
"Reza",
""
],
[
"Zuzak",
"Michael",
""
],
[
"Yang",
"Shanchieh Jay",
""
]
] |
2506.12121
|
Paul Draghis
|
Paul A. Draghis, Jon M. Miller, Laura Brenneman, Elisa Costantini, Luigi C. Gallo, Mark Reynolds, John A. Tomsick, Abderahmen Zoghbi
|
The population of NuSTAR Black Hole X-ray Binaries
|
21 pages + appendix & references, 13 figures. Published in ApJ. Updated to match journal version
|
ApJ 989 227 (2025)
|
10.3847/1538-4357/adec85
| null |
astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
The spin of a black hole (BH) encodes information about its formation and evolution history. Yet the understanding of the distribution of BH spins in X-ray binaries (XBs), of the models used to measure spin, and of their impact on systematic uncertainties remains incomplete. In this work, we expand on previous analyses of the entire NuSTAR archive of accreting BH XBs. Prior work compiled a sample of 245 spectral fits using the relativistic reflection method for NuSTAR observations of 36 BH systems. Here, we aim to probe two aspects: the connection between BH spin and binary system properties, and the relationships between parameters in the spectral fits. We identify moderate negative correlations between spin uncertainty and both BH mass and system inclination, and a moderate positive correlation with distance. We also point out tentative multidimensional degeneracies between inclination, disk density, Fe abundance, ionization, and the presence or absence of absorption features from ionized outflows linked to disk winds. Lastly, we provide a comprehensive view of the observed distribution of BH spins in XBs, in comparison to spins inferred from gravitational waves. We find that the distribution of BH spins in XBs can be described by a beta distribution with $\alpha=5.66$ and $\beta=1.09$. This data set is highly complex, and the analysis presented here does not fully explore all potential parameter correlations. We make the full data set available in Zenodo to the community to encourage further exploration.
|
[
{
"version": "v1",
"created": "Fri, 13 Jun 2025 18:00:00 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 14:56:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Draghis",
"Paul A.",
""
],
[
"Miller",
"Jon M.",
""
],
[
"Brenneman",
"Laura",
""
],
[
"Costantini",
"Elisa",
""
],
[
"Gallo",
"Luigi C.",
""
],
[
"Reynolds",
"Mark",
""
],
[
"Tomsick",
"John A.",
""
],
[
"Zoghbi",
"Abderahmen",
""
]
] |
2506.12389
|
Zhiyuan Su
|
Zhiyuan Su, Sunhao Dai, Xiao Zhang
|
Revisiting Clustering of Neural Bandits: Selective Reinitialization for Mitigating Loss of Plasticity
|
Some proof details are being revised
| null |
10.1145/3711896.3737105
| null |
cs.LG cs.AI stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Clustering of Bandits (CB) methods enhance sequential decision-making by grouping bandits into clusters based on similarity and incorporating cluster-level contextual information, demonstrating effectiveness and adaptability in applications like personalized streaming recommendations. However, when extending CB algorithms to their neural version (commonly referred to as Clustering of Neural Bandits, or CNB), they suffer from loss of plasticity, where neural network parameters become rigid and less adaptable over time, limiting their ability to adapt to non-stationary environments (e.g., dynamic user preferences in recommendation). To address this challenge, we propose Selective Reinitialization (SeRe), a novel bandit learning framework that dynamically preserves the adaptability of CNB algorithms in evolving environments. SeRe leverages a contribution utility metric to identify and selectively reset underutilized units, mitigating loss of plasticity while maintaining stable knowledge retention. Furthermore, when combining SeRe with CNB algorithms, the adaptive change detection mechanism adjusts the reinitialization frequency according to the degree of non-stationarity, ensuring effective adaptation without unnecessary resets. Theoretically, we prove that SeRe enables sublinear cumulative regret in piecewise-stationary environments, outperforming traditional CNB approaches in long-term performances. Extensive experiments on six real-world recommendation datasets demonstrate that SeRe-enhanced CNB algorithms can effectively mitigate the loss of plasticity with lower regrets, improving adaptability and robustness in dynamic settings.
|
[
{
"version": "v1",
"created": "Sat, 14 Jun 2025 07:58:27 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 14:39:36 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Su",
"Zhiyuan",
""
],
[
"Dai",
"Sunhao",
""
],
[
"Zhang",
"Xiao",
""
]
] |
2506.12467
|
Fang Qin
|
Fang Qin, Rui Chen
|
Tunable corner states in topological insulators with long-range hoppings and diverse shapes
|
10 pages, 4 figures, updated references, published version
|
Phys. Rev. B 112, 125405 (2025)
|
10.1103/2pw8-wy6t
| null |
cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
In this work we develop a theoretical framework for the control of corner modes in higher-order topological insulators (HOTIs) featuring long-range hoppings and diverse geometries, enabling precise tunability of their spatial positions. First, we demonstrate that the locations of corner states can be finely tuned by varying long-range hoppings in a circular HOTI, as revealed by a detailed edge theory analysis and the condition of vanishing Dirac mass. Moreover, we show that long-range hoppings in different directions (e.g., $x$ and $y$) have distinct effects on the positioning of corner states. Second, we investigate HOTIs with various polygonal geometries and find that the presence and location of corner modes depend sensitively on the shape. In particular, a corner hosts a localized mode if the Dirac masses of its two adjacent edges have opposite signs, while no corner mode emerges if the masses share the same sign. Our findings offer a versatile approach for the controlled manipulation of corner modes in HOTIs, opening avenues for the design and implementation of higher-order topological materials.
|
[
{
"version": "v1",
"created": "Sat, 14 Jun 2025 12:08:18 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 13:47:54 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 02:45:08 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Qin",
"Fang",
""
],
[
"Chen",
"Rui",
""
]
] |
2506.12943
|
Ofir Alon
|
O. E. Alon and L. S. Cederbaum
|
Fragmentation of a trapped multiple-species bosonic mixture
|
25 pages, 2 figures
|
Physics 7, 38 (2025)
|
10.3390/physics7030038
| null |
cond-mat.quant-gas quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We consider a multiple-species mixture of interacting bosons, $N_1$ bosons of mass $m_1$, $N_2$ bosons of mass $m_2$, and $N_3$ bosons of mass $m_3$ in a harmonic trap of frequency $\omega$. The corresponding intraspecies interaction strengths are $\lambda_{11}$, $\lambda_{22}$, and $\lambda_{33}$, and the interspecies interaction strengths are $\lambda_{12}$, $\lambda_{13}$, and $\lambda_{23}$. When the shape of all interactions are harmonic, this is the generic multiple-species harmonic-interaction model which is exactly solvable. We start by solving the many-particle Hamiltonian and concisely discussing the ground-state wavefunction and energy in explicit forms as functions of all parameters, the masses, numbers of particles, and the intraspecies and interspecies interaction strengths. We then move to compute explicitly the reduced one-particle density matrices for all the species and diagonalize them, thus generalizing the treatment in [J. Chem. Phys. {\bf 161}, 184307 (2024)]. The respective eigenvalues determine the degree of fragmentation of each species. As applications, we focus on aspects that do not appear for the respective single-species and two-species systems. For instance, placing a mixture of two kinds of bosons in a bath made by a third kind, and controlling the fragmentation of the former by coupling to the latter. Another example exploits the possibility of different connectivities (i.e., which species interacts with which species) in the mixture, and demonstrates how the fragmentation of species $3$ can be manipulated by the interaction between species $1$ and species $2$, when species $3$ and $1$ do not interact with each other. We thereby highlight properties of fragmentation that only appear in the multiple-species mixture. Further applications are briefly discussed.
|
[
{
"version": "v1",
"created": "Sun, 15 Jun 2025 19:06:19 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Alon",
"O. E.",
""
],
[
"Cederbaum",
"L. S.",
""
]
] |
2506.13265
|
Rohit Mohan
|
Rohit Mohan, Julia Hindel, Florian Drews, Claudius Gl\"aser, Daniele Cattaneo, Abhinav Valada
|
Open-Set LiDAR Panoptic Segmentation Guided by Uncertainty-Aware Learning
| null | null | null | null |
cs.CV cs.AI cs.LG cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Autonomous vehicles that navigate in open-world environments may encounter previously unseen object classes. However, most existing LiDAR panoptic segmentation models rely on closed-set assumptions, failing to detect unknown object instances. In this work, we propose ULOPS, an uncertainty-guided open-set panoptic segmentation framework that leverages Dirichlet-based evidential learning to model predictive uncertainty. Our architecture incorporates separate decoders for semantic segmentation with uncertainty estimation, embedding with prototype association, and instance center prediction. During inference, we leverage uncertainty estimates to identify and segment unknown instances. To strengthen the model's ability to differentiate between known and unknown objects, we introduce three uncertainty-driven loss functions. Uniform Evidence Loss to encourage high uncertainty in unknown regions. Adaptive Uncertainty Separation Loss ensures a consistent difference in uncertainty estimates between known and unknown objects at a global scale. Contrastive Uncertainty Loss refines this separation at the fine-grained level. To evaluate open-set performance, we extend benchmark settings on KITTI-360 and introduce a new open-set evaluation for nuScenes. Extensive experiments demonstrate that ULOPS consistently outperforms existing open-set LiDAR panoptic segmentation methods.
|
[
{
"version": "v1",
"created": "Mon, 16 Jun 2025 09:03:51 GMT"
},
{
"version": "v2",
"created": "Tue, 12 Aug 2025 22:57:17 GMT"
},
{
"version": "v3",
"created": "Tue, 02 Sep 2025 20:38:10 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Mohan",
"Rohit",
""
],
[
"Hindel",
"Julia",
""
],
[
"Drews",
"Florian",
""
],
[
"Gläser",
"Claudius",
""
],
[
"Cattaneo",
"Daniele",
""
],
[
"Valada",
"Abhinav",
""
]
] |
2506.13431
|
Aman Sharma
|
Aman Sharma, Mithilesh Nayak, Natalia Chepiga, Fr\'ed\'eric Mila
|
Excitations and dynamical structure factor of $J_1-J_2$ spin-$3/2$ and spin-$5/2$ Heisenberg spin chains
| null | null |
10.1103/pddz-42x9
| null |
cond-mat.str-el
|
http://creativecommons.org/licenses/by/4.0/
|
We study the dynamical structure factor of the frustrated spin-$3/2$ $J_1$-$J_2$ Heisenberg chains, with particular focus on the partially dimerized phase that emerges between two Kosterlitz-Thouless transitions. Using a valence bond solid ansatz corroborated by density matrix renormalization group simulations, we investigate the nature of magnon and spinon excitations through the single-mode approximation. We show that the magnon develops an incommensurate dispersion at $J_2 \approx 0.32J_1$, while the spinons, viewed as domain walls between degenerate valence bond solid states, become incommensurate at $J_2 \approx 0.4J_1$ beyond the Lifshitz point ($J_2 \approx 0.388J_1$). The dynamical structure factor exhibits rich spectral features shaped by the interplay between these excitations, with magnons appearing as resonances embedded in the spinon continuum. The spinon gap shows a nonmonotonic behavior, reaching a peak near the center of the partially dimerized phase and closing at the boundaries, suggesting the appearance of a floating phase as a result of the condensation of incommensurate spinons. Comparative analysis with the spin-$5/2$ case confirms the universality of these phenomena across half-integer higher-spin systems. Our results provide detailed insight into how fractionalization and incommensurate condensation govern the spectral properties of frustrated spin chains, offering a unified picture across different spin magnitudes.
|
[
{
"version": "v1",
"created": "Mon, 16 Jun 2025 12:47:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Sharma",
"Aman",
""
],
[
"Nayak",
"Mithilesh",
""
],
[
"Chepiga",
"Natalia",
""
],
[
"Mila",
"Frédéric",
""
]
] |
2506.13554
|
Ronald Katende
|
Ronald Katende
|
Non-Asymptotic Stability and Consistency Guarantees for Physics-Informed Neural Networks via Coercive Operator Analysis
| null | null | null | null |
cs.LG cs.NA math.FA math.NA
|
http://creativecommons.org/licenses/by/4.0/
|
We present a unified theoretical framework for analyzing the stability and consistency of Physics-Informed Neural Networks (PINNs), grounded in operator coercivity, variational formulations, and non-asymptotic perturbation theory. PINNs approximate solutions to partial differential equations (PDEs) by minimizing residual losses over sampled collocation and boundary points. We formalize both operator-level and variational notions of consistency, proving that residual minimization in Sobolev norms leads to convergence in energy and uniform norms under mild regularity. Deterministic stability bounds quantify how bounded perturbations to the network outputs propagate through the full composite loss, while probabilistic concentration results via McDiarmid's inequality yield sample complexity guarantees for residual-based generalization. A unified generalization bound links residual consistency, projection error, and perturbation sensitivity. Empirical results on elliptic, parabolic, and nonlinear PDEs confirm the predictive accuracy of our theoretical bounds across regimes. The framework identifies key structural principles, such as operator coercivity, activation smoothness, and sampling admissibility, that underlie robust and generalizable PINN training, offering principled guidance for the design and analysis of PDE-informed learning systems.
|
[
{
"version": "v1",
"created": "Mon, 16 Jun 2025 14:41:15 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 04:37:00 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Katende",
"Ronald",
""
]
] |
2506.14612
|
Howard Su
|
Howard Su and Huan-Hsin Tseng
|
On Quantum BSDE Solver for High-Dimensional Parabolic PDEs
|
6 pages, 5 figures
| null | null | null |
q-fin.MF q-fin.CP quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We propose a quantum machine learning framework for approximating solutions to high-dimensional parabolic partial differential equations (PDEs) that can be reformulated as backward stochastic differential equations (BSDEs). In contrast to popular quantum-classical network hybrid approaches, this study employs the pure Variational Quantum Circuit (VQC) as the core solver without trainable classical neural networks. The quantum BSDE solver performs pathwise approximation via temporal discretization and Monte Carlo simulation, framed as model-based reinforcement learning. We benchmark VQCbased and classical deep neural network (DNN) solvers on two canonical PDEs as representatives: the Black-Scholes and nonlinear Hamilton-Jacobi-Bellman (HJB) equations. The VQC achieves lower variance and improved accuracy in most cases, particularly in highly nonlinear regimes and for out-of-themoney options, demonstrating greater robustness than DNNs. These results, obtained via quantum circuit simulation, highlight the potential of VQCs as scalable and stable solvers for highdimensional stochastic control problems.
|
[
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:10:42 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 16:51:16 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Su",
"Howard",
""
],
[
"Tseng",
"Huan-Hsin",
""
]
] |
2506.15079
|
Yikai Hou
|
Wenyu Luo, Yikai Hou and Peng Tang
|
Neural Canonical Polyadic Factorization for Traffic Analysis
| null | null | null | null |
cs.LG stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Modern intelligent transportation systems rely on accurate spatiotemporal traffic analysis to optimize urban mobility and infrastructure resilience. However, pervasive missing data caused by sensor failures and heterogeneous sensing gaps fundamentally hinders reliable traffic modeling. This paper proposes a Neural Canonical Polyadic Factorization (NCPF) model that synergizes low-rank tensor algebra with deep representation learning for robust traffic data imputation. The model innovatively embeds CP decomposition into neural architecture through learnable embedding projections, where sparse traffic tensors are encoded into dense latent factors across road segments, time intervals, and mobility metrics. A hierarchical feature fusion mechanism employs Hadamard products to explicitly model multilinear interactions, while stacked multilayer perceptron layers nonlinearly refine these representations to capture complex spatiotemporal couplings. Extensive evaluations on six urban traffic datasets demonstrate NCPF's superiority over six state-of-the-art baselines. By unifying CP decomposition's interpretable factor analysis with neural network's nonlinear expressive power, NCPF provides a principled yet flexible approaches for high-dimensional traffic data imputation, offering critical support for next-generation transportation digital twins and adaptive traffic control systems.
|
[
{
"version": "v1",
"created": "Wed, 18 Jun 2025 02:45:25 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jun 2025 04:06:21 GMT"
},
{
"version": "v3",
"created": "Wed, 09 Jul 2025 14:45:44 GMT"
},
{
"version": "v4",
"created": "Wed, 03 Sep 2025 13:29:28 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Luo",
"Wenyu",
""
],
[
"Hou",
"Yikai",
""
],
[
"Tang",
"Peng",
""
]
] |
2506.16480
|
Hajo Leschke
|
Hajo Leschke
|
The Quantum Formalism Revisited
|
differences to version v1: 44 pages, various minor corrections, additions, and stylistic improvements
| null | null | null |
quant-ph math-ph math.MP physics.hist-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
For the simple system of a point-like particle in the real line I compile and contrast, starting with a concise table, the structural elements of quantum mechanics with those of classical (statistical) mechanics. Despite many similarities, there are the well-known fundamental differences, following from the algebraic non-commutativity in the quantal structure. It was discovered by Werner Heisenberg (1901-1976) in June 1925 on the small island of Helgoland in the North Sea, as a consequence of understanding atomic spectral data within a matrix scheme consistent with energy conservation. I discuss the differences and exemplify their quantifications by the variance and entropic indeterminacy inequalities, by (pseudo-)classical bounds on quantum canonical partition functions, and by the correlation inequalities of John Bell (1928-1990) and others.
|
[
{
"version": "v1",
"created": "Thu, 19 Jun 2025 17:25:40 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 16:47:12 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Leschke",
"Hajo",
""
]
] |
2506.16513
|
Cheyanne Shariat
|
Cheyanne Shariat, Kareem El-Badry, and Smadar Naoz
|
10,000 Resolved Triples from Gaia: Empirical Constraints on Triple Star Populations
|
Accepted to PASP. Relevant code and data can be found at https://github.com/cheyanneshariat/gaia_triples
| null |
10.1088/1538-3873/adfb30
| null |
astro-ph.SR astro-ph.GA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a catalog of $\sim 10,000$ resolved triple star systems within 500 pc of the Sun, constructed using Gaia data. The triples include main-sequence, red giant, and white dwarf components spanning separations of 10 to 50,000 au. A well-characterized selection function allows us to constrain intrinsic demographics of the triple star population. We find that (a) all systems are compatible with being hierarchical and dynamically stable; (b) mutual orbital inclinations are isotropic for wide triples but show modest alignment as the systems become more compact; (c) primary masses follow a Kroupa initial mass function weighted by the triple fraction; (d) inner binary orbital periods, eccentricities, and mass ratios mirror those of isolated binaries, including a pronounced twin excess (mass ratios greater than 0.95) out to separations of 1000+ au, suggesting a common formation pathway; (e) tertiary mass ratios follow a power-law distribution with slope $-1.4$; (f) tertiary orbits are consistent with a log-normal period distribution and thermal eccentricities, subject to dynamical stability. Informed by these observations, we develop a publicly available prescription for generating mock triple star populations. Finally, we estimate the catalog's completeness and infer the intrinsic triple fraction, which rises steadily with primary mass: from $5\%$ at $\lesssim 0.5\,{\rm M_\odot}$ to $35\%$ at $2\,{\rm M_\odot}$. The public catalog provides a robust testbed for models of triple star formation and evolution.
|
[
{
"version": "v1",
"created": "Thu, 19 Jun 2025 18:00:02 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:54:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Shariat",
"Cheyanne",
""
],
[
"El-Badry",
"Kareem",
""
],
[
"Naoz",
"Smadar",
""
]
] |
2506.17072
|
Fellype Nascimento
|
Fellype do Nascimento, Bruno Henrique da Silva Leal, Konstantin Georgiev Kostov
|
A low-cost plasma source aimed for medical applications using Ar as the working gas
|
20 pages, 10 figures
|
Phys. Scr., vol. 100, no. 9, p. 095601, Sep. 2025
|
10.1088/1402-4896/adfe2b
| null |
physics.plasm-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Due to advances in equipment and intense research on plasma biomedical applications over the past few years, plasma devices are now suitable for clinical use. It has been demonstrated that plasma sources can be designed in a way that ensures their safe operation and application whilst preserving the relevant clinical results. However, the manufacturing and operating costs of such plasma devices remain high, which can be decisive for their adoption in medical procedures. In this work, a simple modification of a low-cost plasma jet configuration is proposed in order to evaluate the device's viability for medical applications using argon as working gas. Typically, plasma jets operating with argon (Ar) generate high electrical currents, which makes them unsuitable for medical applications. With the proposed device modification, the treatment is no longer carried out using the plasma jet directly, but rather with the post-discharge effluent enriched with reactive oxygen and nitrogen species. As a result, the electrical current reaching the target remains below the safety threshold established for plasma applications in humans. Tests on water activation by plasma jets with and without the proposed modification indicate that both devices can yield comparable outcomes.
|
[
{
"version": "v1",
"created": "Fri, 20 Jun 2025 15:21:02 GMT"
},
{
"version": "v2",
"created": "Tue, 12 Aug 2025 12:24:41 GMT"
},
{
"version": "v3",
"created": "Tue, 02 Sep 2025 13:02:10 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Nascimento",
"Fellype do",
""
],
[
"Leal",
"Bruno Henrique da Silva",
""
],
[
"Kostov",
"Konstantin Georgiev",
""
]
] |
2506.17769
|
Alexander Popkov
|
A. V. Popkov (MIPT, LPI), Y. Y. Kovalev (MPIfR), A. V. Plavin (BHI Harvard), L. Y. Petrov (NASA GSFC), I. N. Pashchenko (LPI)
|
Dim cores of radio-bright AGN jets: VLBI and Gaia astrometry pinpoint different parsec-scale features
|
17 pages, 9 figures, accepted by MNRAS
| null | null | null |
astro-ph.HE astro-ph.GA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Astrometry with the very long baseline radio interferometry (VLBI) allows to determine the position of a point close to the source's brightest compact detail at milliarcsecond scales. For most active galactic nuclei (AGNs), this compact detail is the opaque core of the radio jet. Rare cases of sources whose brightest detail is not the core but a prominent jet feature parsecs away from the core have been reported, but such sources remained elusive. In this work, we use a novel method for a systematic search of these sources. We scrutinize the AGNs for which the offset between their coordinates determined with VLBI and Gaia is statistically significant and coincides with the vector between two dominant features in their VLBI images, using publicly available archival multi-frequency data. We find 35 sources whose VLBI coordinates are associated with a bright component of their jet separated by several to tens of mas from the radio core. Their Gaia coordinates, in turn, correspond to the jet origin close to the radio jet core. The previously published jet directions of most of them must be reversed. These sources exhibit atypically low brightness temperatures of the radio cores, down to 10^9 K in the host galaxy frame, and, at the same time, extreme brightness of the dominating jet components. We argue that these bright components are standing shock fronts and discuss possible physical explanations for the low core brightness, such as ineffective particle heating, atypical absorption, or differential Doppler boosting.
|
[
{
"version": "v1",
"created": "Sat, 21 Jun 2025 17:44:24 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 11:24:59 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Popkov",
"A. V.",
"",
"MIPT, LPI"
],
[
"Kovalev",
"Y. Y.",
"",
"MPIfR"
],
[
"Plavin",
"A. V.",
"",
"BHI Harvard"
],
[
"Petrov",
"L. Y.",
"",
"NASA GSFC"
],
[
"Pashchenko",
"I. N.",
"",
"LPI"
]
] |
2506.17948
|
Mahzabin Tamanna
|
Mahzabin Tamanna, Yash Chandrani, Matthew Burrows, Brandon Wroblewski, Laurie Williams, Dominik Wermke
|
Your Build Scripts Stink: The State of Code Smells in Build Scripts
|
13 pages, 5 tables, 2 figures
| null | null | null |
cs.SE
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Build scripts are files that automate the process of compiling source code, managing dependencies, running tests, and packaging software into deployable artifacts. These scripts are ubiquitous in modern software development pipelines for streamlining testing and delivery. While developing build scripts, practitioners may inadvertently introduce code smells. Code smells are recurring patterns of poor coding practices that may lead to build failures or increase risk and technical debt. The goal of this study is to aid practitioners in avoiding code smells in build scripts through an empirical study of build scripts and issues on GitHub. We employed a mixed-methods approach, combining qualitative and quantitative analysis. We conducted a qualitative analysis of 2000 build-script-related GitHub issues. Next, we developed a static analysis tool, Sniffer, to identify code smells in 5882 build scripts of Maven, Gradle, CMake, and Make files, collected from 4877 open-source GitHub repositories. We identified 13 code smell categories, with a total of 10,895 smell occurrences, where 3184 were in Maven, 1214 in Gradle, 337 in CMake, and 6160 in Makefiles.
Our analysis revealed that Insecure URLs were the most prevalent code smell in Maven build scripts, while Hardcoded Paths/URLs were commonly observed in both Gradle and CMake scripts. Wildcard Usage emerged as the most frequent smell in Makefiles. The co-occurrence analysis revealed strong associations between specific smell pairs of Hardcoded Paths/URLs with Duplicates, and Inconsistent Dependency Management with Empty or Incomplete Tags, indicating potential underlying issues in the build script structure and maintenance practices. Based on our findings, we also recommended strategies to mitigate the existence of code smells in build scripts to improve the efficiency, reliability, and maintainability of software projects.
|
[
{
"version": "v1",
"created": "Sun, 22 Jun 2025 09:01:42 GMT"
},
{
"version": "v2",
"created": "Tue, 02 Sep 2025 18:32:25 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Tamanna",
"Mahzabin",
""
],
[
"Chandrani",
"Yash",
""
],
[
"Burrows",
"Matthew",
""
],
[
"Wroblewski",
"Brandon",
""
],
[
"Williams",
"Laurie",
""
],
[
"Wermke",
"Dominik",
""
]
] |
2506.18072
|
Hong Ding
|
Yunhao Liu and Suyang Xi and Shiqi Liu and Hong Ding and Chicheng Jin and Chong Zhong and Junjun He and Catherine C. Liu and Yiqing Shen
|
Multimodal Medical Image Binding via Shared Text Embeddings
|
10 pages, 3 figures
| null | null | null |
eess.IV cs.AI cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Medical image analysis increasingly relies on the integration of multiple imaging modalities to capture complementary anatomical and functional information, enabling more accurate diagnosis and treatment planning. Achieving aligned feature representations across these diverse modalities is therefore important for effective multimodal analysis. While contrastive language-image pre-training (CLIP) and its variant have enabled image-text alignments, they require explicitly paired data between arbitrary two modalities, which is difficult to acquire in medical contexts. To address the gap, we present Multimodal Medical Image Binding with Text (M\textsuperscript{3}Bind), a novel pre-training framework that enables seamless alignment of multiple medical imaging modalities through a shared text representation space without requiring explicit paired data between any two medical image modalities. Specifically, based on the insight that different images can naturally bind with text, M\textsuperscript{3}Bind first fine-tunes pre-trained CLIP-like image-text models to align their modality-specific text embedding space while preserving their original image-text alignments. Subsequently, we distill these modality-specific text encoders into a unified model, creating a shared text embedding space. Experiments on X-ray, CT, retina, ECG, and pathological images on multiple downstream tasks demonstrate that M\textsuperscript{3}Bind achieves state-of-the-art performance in zero-shot, few-shot classification and cross-modal retrieval tasks compared to its CLIP-like counterparts. These results validate M\textsuperscript{3}Bind's effectiveness in achieving cross-image-modal alignment for medical analysis.
|
[
{
"version": "v1",
"created": "Sun, 22 Jun 2025 15:39:25 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 16:04:55 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Liu",
"Yunhao",
""
],
[
"Xi",
"Suyang",
""
],
[
"Liu",
"Shiqi",
""
],
[
"Ding",
"Hong",
""
],
[
"Jin",
"Chicheng",
""
],
[
"Zhong",
"Chong",
""
],
[
"He",
"Junjun",
""
],
[
"Liu",
"Catherine C.",
""
],
[
"Shen",
"Yiqing",
""
]
] |
2506.18096
|
Yuxuan Huang
|
Yuxuan Huang, Yihang Chen, Haozheng Zhang, Kang Li, Huichi Zhou, Meng Fang, Linyi Yang, Xiaoguang Li, Lifeng Shang, Songcen Xu, Jianye Hao, Kun Shao, and Jun Wang
|
Deep Research Agents: A Systematic Examination And Roadmap
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The rapid progress of Large Language Models (LLMs) has given rise to a new category of autonomous AI systems, referred to as Deep Research (DR) agents. These agents are designed to tackle complex, multi-turn informational research tasks by leveraging a combination of dynamic reasoning, adaptive long-horizon planning, multi-hop information retrieval, iterative tool use, and the generation of structured analytical reports. In this paper, we conduct a detailed analysis of the foundational technologies and architectural components that constitute Deep Research agents. We begin by reviewing information acquisition strategies, contrasting API-based retrieval methods with browser-based exploration. We then examine modular tool-use frameworks, including code execution, multimodal input processing, and the integration of Model Context Protocols (MCPs) to support extensibility and ecosystem development. To systematize existing approaches, we propose a taxonomy that differentiates between static and dynamic workflows, and we classify agent architectures based on planning strategies and agent composition, including single-agent and multi-agent configurations. We also provide a critical evaluation of current benchmarks, highlighting key limitations such as restricted access to external knowledge, sequential execution inefficiencies, and misalignment between evaluation metrics and the practical objectives of DR agents. Finally, we outline open challenges and promising directions for future research. A curated and continuously updated repository of DR agent research is available at: {https://github.com/ai-agents-2030/awesome-deep-research-agent}.
|
[
{
"version": "v1",
"created": "Sun, 22 Jun 2025 16:52:48 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 15:32:23 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Huang",
"Yuxuan",
""
],
[
"Chen",
"Yihang",
""
],
[
"Zhang",
"Haozheng",
""
],
[
"Li",
"Kang",
""
],
[
"Zhou",
"Huichi",
""
],
[
"Fang",
"Meng",
""
],
[
"Yang",
"Linyi",
""
],
[
"Li",
"Xiaoguang",
""
],
[
"Shang",
"Lifeng",
""
],
[
"Xu",
"Songcen",
""
],
[
"Hao",
"Jianye",
""
],
[
"Shao",
"Kun",
""
],
[
"Wang",
"Jun",
""
]
] |
2506.18368
|
Anja Deli\'c
|
Anja Deli\'c and Matej Grci\'c and Sini\v{s}a \v{S}egvi\'c
|
Sequential keypoint density estimator: an overlooked baseline of skeleton-based video anomaly detection
|
ICCV 2025 Highlight
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Detecting anomalous human behaviour is an important visual task in safety-critical applications such as healthcare monitoring, workplace safety, or public surveillance. In these contexts, abnormalities are often reflected with unusual human poses. Thus, we propose SeeKer, a method for detecting anomalies in sequences of human skeletons. Our method formulates the skeleton sequence density through autoregressive factorization at the keypoint level. The corresponding conditional distributions represent probable keypoint locations given prior skeletal motion. We formulate the joint distribution of the considered skeleton as causal prediction of conditional Gaussians across its constituent keypoints. A skeleton is flagged as anomalous if its keypoint locations surprise our model (i.e. receive a low density). In practice, our anomaly score is a weighted sum of per-keypoint log-conditionals, where the weights account for the confidence of the underlying keypoint detector. Despite its conceptual simplicity, SeeKer surpasses all previous methods on the UBnormal and MSAD-HR datasets while delivering competitive performance on the ShanghaiTech dataset.
|
[
{
"version": "v1",
"created": "Mon, 23 Jun 2025 07:55:28 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 09:05:21 GMT"
},
{
"version": "v3",
"created": "Wed, 03 Sep 2025 08:11:22 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Delić",
"Anja",
""
],
[
"Grcić",
"Matej",
""
],
[
"Šegvić",
"Siniša",
""
]
] |
2506.18734
|
Shu-Min Wu
|
Xiao-Li Huang, Xiao-Ying Jiang, Yu-Xuan Wang, Si-Yu Liu, Zejun Wang, Shu-Min Wu
|
Can boundary configuration be tuned to optimize directional quantum steering harvesting?
|
22 pages, 8 figures,
|
JHEP 09 (2025) 023
|
10.1007/JHEP09(2025)023
| null |
quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We investigate the harvesting of quantum steering and its asymmetry between two static detectors locally interacting with a vacuum massless scalar field near an infinite, perfectly reflecting boundary. The detectors are arranged either parallel or orthogonal to the boundary, with detector $B$ assumed to have an energy gap greater than or equal to that of detector $A$. It is interesting to observe that, with increasing distance between the detectors and the boundary, the boundary tends to suppress quantum steering in one direction while enhancing it in the opposite direction. In the case of identical detectors, steering is symmetric when they are aligned parallel to the boundary. However, orthogonal alignment breaks this symmetry due to their unequal spatial proximity to the boundary. For non-identical detectors in the parallel configuration, the steering from $A$ to $B$ ($A \rightarrow B$) generally surpasses that from $B$ to $A$ ($B \rightarrow A$). In contrast, when the detectors are oriented orthogonally to the boundary, the relative strength of $A \rightarrow B$ and $B \rightarrow A$ steerability depends on the interplay between the boundary effects and the detectors' energy gap difference. Across most of the parameter space, the orthogonal alignment tends to enhance $B \rightarrow A$ steering while suppressing $A \rightarrow B$ steering compared to the parallel setup. These findings suggest that boundary configurations should be flexibly adjusted according to the directional dependence of steering harvesting in order to optimize quantum information extraction.
|
[
{
"version": "v1",
"created": "Mon, 23 Jun 2025 15:11:41 GMT"
},
{
"version": "v2",
"created": "Wed, 03 Sep 2025 04:13:53 GMT"
}
] | 2025-09-04T00:00:00 |
[
[
"Huang",
"Xiao-Li",
""
],
[
"Jiang",
"Xiao-Ying",
""
],
[
"Wang",
"Yu-Xuan",
""
],
[
"Liu",
"Si-Yu",
""
],
[
"Wang",
"Zejun",
""
],
[
"Wu",
"Shu-Min",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.