id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 3
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-08-22 00:00:00
| authors_parsed
listlengths 1
3.31k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2508.15316
|
Abdul Rehman
|
Abdul Rehman, Jian-Jun Zhang, and Xiaosong Yang
|
CUPE: Contextless Universal Phoneme Encoder for Language-Agnostic Speech Processing
|
Accepted in: 8th International Conference on Natural Language and Speech Processing (ICNLSP 2025)
| null | null | null |
cs.CL cs.LG eess.AS
|
http://creativecommons.org/licenses/by/4.0/
|
Universal phoneme recognition typically requires analyzing long speech segments and language-specific patterns. Many speech processing tasks require pure phoneme representations free from contextual influence, which motivated our development of CUPE - a lightweight model that captures key phoneme features in just 120 milliseconds, about one phoneme's length. CUPE processes short, fixed-width windows independently and, despite fewer parameters than current approaches, achieves competitive cross-lingual performance by learning fundamental acoustic patterns common to all languages. Our extensive evaluation through supervised and self-supervised training on diverse languages, including zero-shot tests on the UCLA Phonetic Corpus, demonstrates strong cross-lingual generalization and reveals that effective universal speech processing is possible through modeling basic acoustic patterns within phoneme-length windows.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:27:10 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Rehman",
"Abdul",
""
],
[
"Zhang",
"Jian-Jun",
""
],
[
"Yang",
"Xiaosong",
""
]
] |
2508.15317
|
Zhaorui Tan
|
Zhaorui Tan, Yijie Hu, Xi Yang, Qiufeng Wang, Anh Nguyen, Kaizhu Huang
|
Saving for the future: Enhancing generalization via partial logic regularization
| null | null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Generalization remains a significant challenge in visual classification tasks, particularly in handling unknown classes in real-world applications. Existing research focuses on the class discovery paradigm, which tends to favor known classes, and the incremental learning paradigm, which suffers from catastrophic forgetting. Recent approaches such as the L-Reg technique employ logic-based regularization to enhance generalization but are bound by the necessity of fully defined logical formulas, limiting flexibility for unknown classes. This paper introduces PL-Reg, a novel partial-logic regularization term that allows models to reserve space for undefined logic formulas, improving adaptability to unknown classes. Specifically, we formally demonstrate that tasks involving unknown classes can be effectively explained using partial logic. We also prove that methods based on partial logic lead to improved generalization. We validate PL-Reg through extensive experiments on Generalized Category Discovery, Multi-Domain Generalized Category Discovery, and long-tailed Class Incremental Learning tasks, demonstrating consistent performance improvements. Our results highlight the effectiveness of partial logic in tackling challenges related to unknown classes.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:27:33 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Tan",
"Zhaorui",
""
],
[
"Hu",
"Yijie",
""
],
[
"Yang",
"Xi",
""
],
[
"Wang",
"Qiufeng",
""
],
[
"Nguyen",
"Anh",
""
],
[
"Huang",
"Kaizhu",
""
]
] |
2508.15318
|
Qian-Rui Lee
|
Qian-Rui Lee, Daw-Wei Wang
|
Flow Matching at Scale: A Machine Learning Framework for Efficient Large-Size Sampling of Many-Body Systems
| null | null | null | null |
cond-mat.stat-mech cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a machine learning framework based on Flow Matching to overcome the scaling limitations of Markov Chain Monte Carlo (MCMC) methods. We demonstrate its capability in the 2D XY model, where a single network, trained only on configurations from a small ($32\times 32$) lattice at sparse temperature points, generates reliable samples for a significantly larger system ($128\times 128$) across a continuous temperature range without retraining. The generated configurations show strong agreement with key thermodynamic observables and correctly capture the signatures of the Berezinskii-Kosterlitz-Thouless (BKT) transition. This dual generalization is enabled by the Flow Matching framework, which allows us to learn a continuous, temperature-conditioned mapping. At the same time, the inductive biases of the underlying CNN architecture ensure that the learned local physical rules are scale-invariant. This "train-small, generate-large" capability establishes a new paradigm for efficiently studying critical phenomena, offering a significant computational advantage for exploring the thermodynamic limit. The method can be directly applied to other classical or quantum many-body systems described by continuous fields on a lattice.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:29:25 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lee",
"Qian-Rui",
""
],
[
"Wang",
"Daw-Wei",
""
]
] |
2508.15319
|
Shuhan Zhou
|
Weijun Xu, Shuhan Zhou
|
Interface fluctuations for $1$D stochastic Allen-Cahn equation -- singular regime
|
50 pages. Comments are welcome
| null | null | null |
math.PR math.AP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study interface fluctuations for the $1$D stochastic Allen-Cahn equation perturbed by half a spatial derivative of the spacetime white noise. This half derivative makes the solution distribution-valued, so that proper renormalization is needed to make sense of the solution.
We show that if the noise is sufficiently small, then an analogue of the classical results by \cite{Fun95,BBDMP98} holds in this singular regime. More precisely, for initial data close to the traveling wave solution of the deterministic equation, under proper long time scaling, the solution still stays close to the family of traveling waves, and the interface location moves according to an approximate diffusion process. There is one interesting difference between our singular regime and the classical situation: even if the solution and its approximate phase separation point are both well defined, the intended diffusion describing the movement of the canonical candidate of the phase point is not (even for fixed $\eps$). Two infinite quantities arise from the derivation of such an SDE, one due to singularity of the noise, and the other from renormalization. Magically, it turns out that they cancel out each other, thus making the derivation of the interface SDE valid in the $\eps \rightarrow 0$ limit.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:29:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xu",
"Weijun",
""
],
[
"Zhou",
"Shuhan",
""
]
] |
2508.15320
|
Nicholas Mueller
|
Nicholas Mueller, Santiago Badia, Yiran Zhao
|
Reduced basis solvers for unfitted methods on parameterized domains
|
22 pages, 7 figures, 5 tables
| null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we present a unified framework for reduced basis approximations of parametrized partial differential equations defined on parameter-dependent domains. Our approach combines unfitted finite element methods with both classical and tensor-based reduced basis techniques -- particularly the tensor-train reduced basis method -- to enable efficient and accurate model reduction on general geometries. To address the challenge of reconciling geometric variability with fixed-dimensional snapshot representations, we adopt a deformation-based strategy that maps a reference configuration to each parameterized domain. Furthermore, we introduce a localization procedure to construct dictionaries of reduced subspaces and hyper-reduction approximations, which are obtained via matrix discrete empirical interpolation in our work. We extend the proposed framework to saddle-point problems by adapting the supremizer enrichment strategy to unfitted methods and deformed configurations, demonstrating that the supremizer operator can be defined on the reference configuration without loss of stability. Numerical experiments on two- and three-dimensional problems -- including Poisson, linear elasticity, incompressible Stokes and Navier-Stokes equations -- demonstrate the flexibility, accuracy and efficiency of the proposed methodology.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:34:06 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mueller",
"Nicholas",
""
],
[
"Badia",
"Santiago",
""
],
[
"Zhao",
"Yiran",
""
]
] |
2508.15321
|
Ivan Spirandelli
|
Ivan Spirandelli, Dmitriy Morozov, Arnur Nigmetov, Myfanwy E. Evans
|
Topological potentials guiding protein self-assembly
|
19 pages, 6 figures
| null | null | null |
cond-mat.soft cs.CG math.AT
|
http://creativecommons.org/licenses/by/4.0/
|
The simulated self-assembly of molecular building blocks into functional complexes is a key area of study in computational biology and materials science. Self-assembly simulations of proteins, driven by short-range non-polar interactions, can find the biologically correct assembly as the energy minimizing state. Short-ranged potentials produce rugged energy landscapes however, which lead to simulations becoming trapped in non-functional, local minimizers.
Successful self-assembly simulations depend both on the physical realism of the driving potentials as well as their ability to efficiently explore the configuration space.
We introduce a long-range topological potential, quantified via weighted total persistence, and combine it with the morphometric approach to solvation-free energy. This combination improves the assembly success rate in simulations of the tobacco mosaic virus dimer and other protein complexes by up to sixteen-fold compared with the morphometric model alone. It further enables successful simulation in systems that don't otherwise assemble during the examined timescales.
Compared to previous topology-based work, which has been primarily descriptive, our approach uses topological measures as an active energetic bias that is independent of electrostatics or chemical specificity and depends only on atomic coordinates. Therefore the method can, in principle, be applied to arbitrary systems where such coordinates are optimized. Integrating topological descriptions into an energy function offers a general strategy for overcoming kinetic barriers in molecular simulations, with potential applications in drug design, materials development, and the study of complex self-assembly processes.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:35:24 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Spirandelli",
"Ivan",
""
],
[
"Morozov",
"Dmitriy",
""
],
[
"Nigmetov",
"Arnur",
""
],
[
"Evans",
"Myfanwy E.",
""
]
] |
2508.15322
|
Alexey Golovnev
|
Alexey Golovnev, Mustafa M. Hemida
|
More on Bianchi I spacetimes and f(T) gravity
|
18 pages
| null | null | null |
gr-qc hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Bianchi I cosmological solutions in f(T) gravity are discussed. We start from diagonal metrics and tetrads and show that their dynamical equations are pretty much tractable analytically, with a possible arena for physical applications. Then we derive a very bad unpredictability of the teleparallel connection in these configurations. Namely, even for the simple isotropic Friedmann universes, one might apply an arbitrary time-dependent spatial rotation to the standard tetrad without changing anything in the cosmological equations.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:36:03 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Golovnev",
"Alexey",
""
],
[
"Hemida",
"Mustafa M.",
""
]
] |
2508.15323
|
Zechuan Liu
|
Chong-Yuan Xu, Ze-Chuan Liu, Yong Xu
|
Fermion-to-Fermion Low-Density Parity-Check Codes
|
6 pages, 5 figures
| null | null | null |
quant-ph cs.DS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Simulating fermionic systems on qubit-based quantum computers often demands significant computational resources due to the requirement to map fermions to qubits. Thus, designing a fault-tolerant quantum computer that operates directly with fermions offers an effective solution to this challenge. Here, we introduce a protocol for fault-tolerant fermionic quantum computation utilizing fermion-to-fermion low-density parity-check codes. Our method employs a fermionic low-density parity-check code memory, which transfers its state to fermionic color code processors assisted by lattice surgery, where logical operations are subsequently performed. To construct the fermionic low-density parity-check memory, we develop a systematic approach for creating fermionic stabilizer codes based on self-dual Calderbank-Shor-Steane codes. We present examples demonstrating these codes' significantly improved coding rate compared to the fermionic color code. Finally, we simulate the dynamics of a fermionic system using our protocol, illustrating effective error suppression.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:37:03 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xu",
"Chong-Yuan",
""
],
[
"Liu",
"Ze-Chuan",
""
],
[
"Xu",
"Yong",
""
]
] |
2508.15324
|
Martin Plesch
|
Sergej Faleti\v{c}, Assen Kyuldjiev, Thomas Lindner, Dorottya Schnider, \'Eva Izsa, Mih\'aly H\"om\"ostrei, P\'eter Jenei, Boyka Aneva, Daniela D. Ivanova, Hynek N\v{e}mec, Franti\v{s}ek Kundracik, and Martin Plesch
|
Development of soft and hard skills of high-school students via Young Physicists' Tournament
|
11 pages, 3 figures
| null |
10.1088/1361-6404/adedd6
| null |
physics.ed-ph
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
The Young Physicists' Tournament (YPT) inspires high-school students to immerse themselves into an inquiry process closely resembling the real physics research. Using questionnaires' replies from the students and their teachers engaged in YPT, we investigated the perception of teachers and students on how the preparation for YPT contributes to the development of hard skills (physics content knowledge, mathematics, modelling...) and soft skills (communication, team work, organization...) of the students, and how it compares to regular classes and other activities. This comparison shows a positive role of YPT among students, and even a more positive perception among teachers. The significant development of a wide range of advanced skills justifies the substantial effort, resources and dedication of both teachers and students required for the participation in YPT.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:37:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Faletič",
"Sergej",
""
],
[
"Kyuldjiev",
"Assen",
""
],
[
"Lindner",
"Thomas",
""
],
[
"Schnider",
"Dorottya",
""
],
[
"Izsa",
"Éva",
""
],
[
"Hömöstrei",
"Mihály",
""
],
[
"Jenei",
"Péter",
""
],
[
"Aneva",
"Boyka",
""
],
[
"Ivanova",
"Daniela D.",
""
],
[
"Němec",
"Hynek",
""
],
[
"Kundracik",
"František",
""
],
[
"Plesch",
"Martin",
""
]
] |
2508.15325
|
Bingsheng Shen
|
Bingsheng Shen, Zhengchun Zhou, Yang Yang, Pingzhi Fan
|
Some Optimal and Near Optimal Doppler Resilient Complementary Sequence Sets
| null | null | null | null |
cs.IT math.IT
|
http://creativecommons.org/licenses/by/4.0/
|
Sequences with excellent ambiguity functions are very useful in radar detection and modern mobile communications. Doppler resilient complementary sequence (DRCS) is a new type of sequence proposed recently, which can achieve lower ambiguity function sidelobes by summing the ambiguity functions of subsequences. In this paper, we introduce some new constructions of DRCS sets (DRCSSs) based on one-coincidence frequency-hopping sequence sets (OC-FHSSs), almost difference sets (ADSs), some specific sequences, etc. Critically, the proposed DRCSSs are optimal or near optimal.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:38:45 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Shen",
"Bingsheng",
""
],
[
"Zhou",
"Zhengchun",
""
],
[
"Yang",
"Yang",
""
],
[
"Fan",
"Pingzhi",
""
]
] |
2508.15326
|
Weijiang Lai
|
Weijiang Lai, Beihong Jin, Jiongyan Zhang, Yiyuan Zheng, Jian Dong, Jia Cheng, Jun Lei, Xingxing Wang
|
Exploring Scaling Laws of CTR Model for Online Performance Improvement
| null |
Recsys 2025
|
10.1145/3705328.3748046
| null |
cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
CTR models play a vital role in improving user experience and boosting business revenue in many online personalized services. However, current CTR models generally encounter bottlenecks in performance improvement. Inspired by the scaling law phenomenon of LLMs, we propose a new paradigm for improving CTR predictions: first, constructing a CTR model with accuracy scalable to the model grade and data size, and then distilling the knowledge implied in this model into its lightweight model that can serve online users. To put it into practice, we construct a CTR model named SUAN (Stacked Unified Attention Network). In SUAN, we propose the UAB as a behavior sequence encoder. A single UAB unifies the modeling of the sequential and non-sequential features and also measures the importance of each user behavior feature from multiple perspectives. Stacked UABs elevate the configuration to a high grade, paving the way for performance improvement. In order to benefit from the high performance of the high-grade SUAN and avoid the disadvantage of its long inference time, we modify the SUAN with sparse self-attention and parallel inference strategies to form LightSUAN, and then adopt online distillation to train the low-grade LightSUAN, taking a high-grade SUAN as a teacher. The distilled LightSUAN has superior performance but the same inference time as the LightSUAN, making it well-suited for online deployment. Experimental results show that SUAN performs exceptionally well and holds the scaling laws spanning three orders of magnitude in model grade and data size, and the distilled LightSUAN outperforms the SUAN configured with one grade higher. More importantly, the distilled LightSUAN has been integrated into an online service, increasing the CTR by 2.81% and CPM by 1.69% while keeping the average inference time acceptable. Our source code is available at https://github.com/laiweijiang/SUAN.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:41:15 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lai",
"Weijiang",
""
],
[
"Jin",
"Beihong",
""
],
[
"Zhang",
"Jiongyan",
""
],
[
"Zheng",
"Yiyuan",
""
],
[
"Dong",
"Jian",
""
],
[
"Cheng",
"Jia",
""
],
[
"Lei",
"Jun",
""
],
[
"Wang",
"Xingxing",
""
]
] |
2508.15327
|
Xiancheng Gao
|
Xiancheng Gao, Yufeng Shi, Wengang Zhou and Houqiang Li
|
Search-Based Credit Assignment for Offline Preference-Based Reinforcement Learning
|
7 pages, 6 figures, under review
| null | null | null |
cs.AI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Offline reinforcement learning refers to the process of learning policies from fixed datasets, without requiring additional environment interaction. However, it often relies on well-defined reward functions, which are difficult and expensive to design. Human feedback is an appealing alternative, but its two common forms, expert demonstrations and preferences, have complementary limitations. Demonstrations provide stepwise supervision, but they are costly to collect and often reflect limited expert behavior modes. In contrast, preferences are easier to collect, but it is unclear which parts of a behavior contribute most to a trajectory segment, leaving credit assignment unresolved. In this paper, we introduce a Search-Based Preference Weighting (SPW) scheme to unify these two feedback sources. For each transition in a preference labeled trajectory, SPW searches for the most similar state-action pairs from expert demonstrations and directly derives stepwise importance weights based on their similarity scores. These weights are then used to guide standard preference learning, enabling more accurate credit assignment that traditional approaches struggle to achieve. We demonstrate that SPW enables effective joint learning from preferences and demonstrations, outperforming prior methods that leverage both feedback types on challenging robot manipulation tasks.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:41:45 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gao",
"Xiancheng",
""
],
[
"Shi",
"Yufeng",
""
],
[
"Zhou",
"Wengang",
""
],
[
"Li",
"Houqiang",
""
]
] |
2508.15328
|
Henriette Bast
|
Henriette Bast, Rafael C. Schick, Thomas Koehler, Franz Pfeiffer
|
Technical Note: Low-Dose Simulation for Grating-Based X-Ray Dark-Field Radiography Using a Virtually Decreased Irradiation Area
| null | null | null | null |
physics.med-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Background: X-ray dark-field radiography uses small-angle scattering to visualize the structural integrity of lung alveoli. To study the influence of dose reduction on clinical dark-field radiographs, one can simulate low-dose images by virtually reducing the irradiated area. However, these simulations can exhibit stripe artifacts. Purpose: Validation of the low-dose simulation algorithm reported by Schick & Bast et al., PLoS ONE, 2024. Furthermore, we want to demonstrate that stripe artifacts observed in simulated images at very low-dose levels are introduced by limitations of the algorithm and would not appear in actual low-dose dark-field images. Methods: Dark-field radiographs of an anthropomorphic chest phantom were acquired at different tube currents equaling different radiation doses. Based on the measurement with a high radiation dose, dark-field radiographs corresponding to lower radiation doses were simulated by virtually reducing the irradiated area. The simulated low-dose radiographs were evaluated by a quantitative comparison of the dark-field signal using different regions of interests. Results: Dark-field radiographs acquired at one quarter of the standard dose were artifact-free. The dark-field signal differed from the simulated radiographs by up to 10%. Algorithm-induced stripe artifacts decrease the image quality of the simulated radiographs. Conclusions: Virtually reducing the irradiation area is a simple approach to generate low-dose radiographs based on images acquired with scanning-based dark-field radiography. However, as the algorithm creates stripe artifacts in the dark-field images, particularly at higher dose reductions, that are not present in measured low-dose images, simulated images have reduced image quality compared to their measured counterparts.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:42:38 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bast",
"Henriette",
""
],
[
"Schick",
"Rafael C.",
""
],
[
"Koehler",
"Thomas",
""
],
[
"Pfeiffer",
"Franz",
""
]
] |
2508.15329
|
Shigeru Taniguchi
|
Shigeru Taniguchi
|
One-dimensional blast waves in a rarefied polyatomic gas with large bulk viscosity based on rational extended thermodynamics
|
9 pages, 3 figures
| null | null | null |
physics.flu-dyn
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The one-dimensional blast waves in a rarefied polyatomic gas with large bulk viscosity are studied based on rational extended thermodynamics (RET) with six independent fields: mass density, velocity, (equilibrium) pressure, and dynamic pressure. First, by using the method of Lie group theory, we derive a similarity solution of blast waves induced by an intense point explosion. We discuss the deviation from the well-known Sedov-von Neumann-Taylor solution due to the dynamic pressure. Second, we analyze the time evolution by numerically solving the field equations of the RET theory directly for both cases of the intense explosion corresponding to the similarity solution and moderately strong explosion with generic temperature dependence of the bulk viscosity. It is shown that the prediction by the RET theory shows quite different behaviors from those by the system of the Euler equations, and also the Navier-Stokes-Fourier theory when the relaxation time for the dynamic pressure is considerable.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:43:40 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Taniguchi",
"Shigeru",
""
]
] |
2508.15330
|
Katsuhiko Matsuzaki
|
Katsuhiko Matsuzaki and Taiki Resnick
|
Arithmetic Y-frieze patterns of width 3 and 4
| null | null | null | null |
math.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We determine all arithmetic Y-Frieze patterns of width $3$ and $4$. As a consequence, for $n=3,4$, we verify the surjectivity of a map $p_n$ which corresponds arithmetic Y-Frieze patterns of width $n$ to Coxeter's Frieze patterns.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:46:04 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Matsuzaki",
"Katsuhiko",
""
],
[
"Resnick",
"Taiki",
""
]
] |
2508.15331
|
Paul M\"ucksch
|
Paul M\"ucksch, Masahiko Yoshinaga
|
Milnor fibrations and oriented matroids
|
29 pages, 9 figures
| null | null | null |
math.AG math.AT math.CO math.GT
|
http://creativecommons.org/licenses/by/4.0/
|
We introduce a combinatorial model for the Milnor fibration of a complexified real arrangement using oriented matroids. It is a poset quasi-fibration, a notion recently introduced by the first author, whose domain is a subdivision of the Salvetti complex stemming from a natural subdivision of the dual oriented matroid complex. This yields a concrete finite regular CW complex which is homotopy equivalent to the Milnor fiber of the complexified real arrangement and implies that the homotopy type of the Milnor fiber of a complexified real arrangement only depends on the underlying combinatorial structure given by its oriented matroid. Moreover, our construction works for any oriented matroid, disregarding realizability, so we obtain a notion of a combinatorial Milnor fibration for any oriented matroid.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:49:53 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mücksch",
"Paul",
""
],
[
"Yoshinaga",
"Masahiko",
""
]
] |
2508.15332
|
Yasuhito Narita
|
Eckart Marsch and Yasuhito Narita
|
On the Lagrangian and fermion mass of the unified $SU(2) \otimes SU(4)$ gauge field theory
| null | null | null | null |
hep-ph hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
{In the Standard Model of elementary particles the fermions are assumed to be intrinsically massless. Here we propose a new theoretical idea of fermion mass generation (other than by the Higgs mechanism) through the coupling with the vector gauge fields of the unified $SU(2) \otimes SU(4)$ gauge symmetry, especially with the $Z$ boson of the weak interaction that affects all elementary fermions. The resulting small masses are suggested to be proportional to the self-energy of the $Z$ field as described by a Yukawa potential. Thereby the electrically neutral neutrino just gets a tiny mass through its $Z$-field coupling. In contrast, the electrically charged electron and quarks can become more massive by the inertia induced through the Coulomb energy of the electrostatic fields surrounding them in their rest frames.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:55:11 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Marsch",
"Eckart",
""
],
[
"Narita",
"Yasuhito",
""
]
] |
2508.15333
|
Ulises Torrella
|
Francesco Dagnino, Paola Giannini, Violet Ka I Pun and Ulises Torrella
|
Fair Termination for Resource-Aware Active Objects
|
18 pages, 12 pages of appendix, 12 figures, APLAS 2025
| null | null | null |
cs.PL
|
http://creativecommons.org/licenses/by/4.0/
|
Active object systems are a model of distributed computation that has been adopted for modelling distributed systems and business process workflows. This field of modelling is, in essence, concurrent and resource-aware, motivating the development of resource-aware formalisations on the active object model. The contributions of this work are the development of a core calculus for resource-aware active objects together with a type system ensuring that well-typed programs are fairly terminating, i.e., they can always eventually terminate. To achieve this, we combine techniques from graded semantics and type systems, which are quite well understood for sequential programs, with those for fair termination, which have been developed for synchronous~sessions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 07:56:23 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Dagnino",
"Francesco",
""
],
[
"Giannini",
"Paola",
""
],
[
"Pun",
"Violet Ka I",
""
],
[
"Torrella",
"Ulises",
""
]
] |
2508.15334
|
Guirui Zhong
|
Guirui Zhong, Qing Wang, Jun Du, Lei Wang, Mingqi Cai, and Xin Fang
|
An Enhanced Audio Feature Tailored for Anomalous Sound Detection Based on Pre-trained Models
|
13 pages, 3 figures, accepted by ICANN2025
| null | null | null |
cs.SD cs.LG eess.AS
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Anomalous Sound Detection (ASD) aims at identifying anomalous sounds from machines and has gained extensive research interests from both academia and industry. However, the uncertainty of anomaly location and much redundant information such as noise in machine sounds hinder the improvement of ASD system performance. This paper proposes a novel audio feature of filter banks with evenly distributed intervals, ensuring equal attention to all frequency ranges in the audio, which enhances the detection of anomalies in machine sounds. Moreover, based on pre-trained models, this paper presents a parameter-free feature enhancement approach to remove redundant information in machine audio. It is believed that this parameter-free strategy facilitates the effective transfer of universal knowledge from pre-trained tasks to the ASD task during model fine-tuning. Evaluation results on the Detection and Classification of Acoustic Scenes and Events (DCASE) 2024 Challenge dataset demonstrate significant improvements in ASD performance with our proposed methods.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:04:08 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhong",
"Guirui",
""
],
[
"Wang",
"Qing",
""
],
[
"Du",
"Jun",
""
],
[
"Wang",
"Lei",
""
],
[
"Cai",
"Mingqi",
""
],
[
"Fang",
"Xin",
""
]
] |
2508.15335
|
Bin Deng
|
Bin Deng, Yizhe Feng, Zeming Liu, Qing Wei, Xiangrong Zhu, Shuai Chen, Yuanfang Guo, Yunhong Wang
|
RETAIL: Towards Real-world Travel Planning for Large Language Models
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Although large language models have enhanced automated travel planning abilities, current systems remain misaligned with real-world scenarios. First, they assume users provide explicit queries, while in reality requirements are often implicit. Second, existing solutions ignore diverse environmental factors and user preferences, limiting the feasibility of plans. Third, systems can only generate plans with basic POI arrangements, failing to provide all-in-one plans with rich details. To mitigate these challenges, we construct a novel dataset \textbf{RETAIL}, which supports decision-making for implicit queries while covering explicit queries, both with and without revision needs. It also enables environmental awareness to ensure plan feasibility under real-world scenarios, while incorporating detailed POI information for all-in-one travel plans. Furthermore, we propose a topic-guided multi-agent framework, termed TGMA. Our experiments reveal that even the strongest existing model achieves merely a 1.0% pass rate, indicating real-world travel planning remains extremely challenging. In contrast, TGMA demonstrates substantially improved performance 2.72%, offering promising directions for real-world travel planning.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:08:38 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Deng",
"Bin",
""
],
[
"Feng",
"Yizhe",
""
],
[
"Liu",
"Zeming",
""
],
[
"Wei",
"Qing",
""
],
[
"Zhu",
"Xiangrong",
""
],
[
"Chen",
"Shuai",
""
],
[
"Guo",
"Yuanfang",
""
],
[
"Wang",
"Yunhong",
""
]
] |
2508.15336
|
Subhasis Dasgupta Dr.
|
Subhasis Dasgupta, Preetam Saha, Agniva Roy, Jaydip Sen
|
Predicting Road Crossing Behaviour using Pose Detection and Sequence Modelling
|
This is a pre-print version of the original paper accepted in the IEEE conference INDISCON 2025. It contains 8 figures and 1 table. The length of the paper is 7 pages
| null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The world is constantly moving towards AI based systems and autonomous vehicles are now reality in different parts of the world. These vehicles require sensors and cameras to detect objects and maneuver according to that. It becomes important to for such vehicles to also predict from a distant if a person is about to cross a road or not. The current study focused on predicting the intent of crossing the road by pedestrians in an experimental setup. The study involved working with deep learning models to predict poses and sequence modelling for temporal predictions. The study analysed three different sequence modelling to understand the prediction behaviour and it was found out that GRU was better in predicting the intent compared to LSTM model but 1D CNN was the best model in terms of speed. The study involved video analysis, and the output of pose detection model was integrated later on to sequence modelling techniques for an end-to-end deep learning framework for predicting road crossing intents.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:08:50 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Dasgupta",
"Subhasis",
""
],
[
"Saha",
"Preetam",
""
],
[
"Roy",
"Agniva",
""
],
[
"Sen",
"Jaydip",
""
]
] |
2508.15337
|
M. Paola Vaccaro
|
Maria Paola Vaccaro
|
Hierarchical Black Hole Mergers in AGN Disks: Tracing Massive Black Hole Growth Across Cosmic Time
|
4 pages, 1 figure. Proceedings contribution to IAU Symposium 398 (MODEST-25, Seoul, June 2025); to appear in the IAU Proceedings Series
| null | null | null |
astro-ph.HE astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
The dense and gaseous environments of active galactic nuclei (AGNs) can catalyze repeated mergers of stellar-mass black holes (BHs), potentially explaining the high-mass tail of binary black hole (BBH) mergers observed by LIGO-Virgo-KAGRA (LVK). We present a semi-analytical population synthesis framework that captures key physical processes in AGN disks, including gas capture, migration, binary pair-up, gas hardening, and dynamical binary-single interactions. Our simulations show that AGN disks can produce hierarchical mergers, especially near migration traps, and may contribute to the high-mass, high-spin BBH population. This work opens prospects for constraining both the AGN channel and supermassive black hole growth with future gravitational-wave detections.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:09:16 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Vaccaro",
"Maria Paola",
""
]
] |
2508.15338
|
Jinning Yang
|
Jinning Yang and Wen Shi
|
DiagECG: An LLM-Driven Framework for Diagnostic Reasoning via Discretized ECG Tokenization
| null | null | null | null |
cs.AI cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Electrocardiography plays a central role in cardiovascular diagnostics, yet existing automated approaches often struggle to generalize across clinical tasks and offer limited support for open-ended reasoning. We present DiagECG, a novel framework that integrates time-series and language modeling by enabling large language models to process 12-lead ECG signals for clinical text generation tasks. Our approach discretizes continuous ECG embeddings into symbolic tokens using a lead-independent encoder and quantization module. These tokens are then used to extend the vocabulary of LLM, allowing the model to handle both ECG and natural language inputs in a unified manner. To bridge the modality gap, we pretrain the model on an autoregressive ECG forecasting task, enabling the LLM to model temporal dynamics using its native language modeling capabilities. Finally, we perform instruction tuning on both ECG question answering and diagnostic report generation. Without modifying the core model, DiagECG achieves strong performance across tasks while maintaining generalization to out-of-distribution settings. Extensive experiments demonstrate the effectiveness of each component and highlight the potential of integrating symbolic ECG representations into LLMs for medical reasoning.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:13:37 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yang",
"Jinning",
""
],
[
"Shi",
"Wen",
""
]
] |
2508.15339
|
Jean-Marc Schlenker
|
Jean-Marc Schlenker
|
Projective rigidity of circle patterns and polyhedral surfaces in hyperbolic ends
|
24 pages, no picture
| null | null | null |
math.GT math.DG
|
http://creativecommons.org/licenses/by/4.0/
|
Let $S$ be a closed, orientable surface of genus $g\geq 2$. We consider Delaunay circle patterns on $S$ equipped with a complex projective structure. We prove that the space of complex projective structures on $S$ equipped with a Delaunay circle pattern of prescribed combinatorics and intersection angles is a manifold of dimension $6g-6$, and that the forgetful map to the space $\cC_S$ of $\CP^1$-structures on $S$ is a Lagrangian immersion. This extends a recent result of Bonsante and Wolf for circle packings.
This statement, and its proof, are more conveniently stated in terms of ideal polyhedral surfaces (surfaces with vertices at infinity) in hyperbolic ends, with the angles between the circles corresponding to the dihedral angles. Seen from this angle, we extend the statement to ideal polyhedral surfaces with prescribed edge lengths (or induced metrics), and to other types of polyhedral surfaces, either compact or hyperideal.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:13:50 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Schlenker",
"Jean-Marc",
""
]
] |
2508.15340
|
Anna Feh\'er
|
Anna Feh\'er, D\'aniel Cseh, R\'obert Kov\'acs
|
Discovering correlations between metal foam thermal characteristics and non-Fourier behavior
| null | null | null | null |
physics.app-ph cond-mat.other
|
http://creativecommons.org/licenses/by/4.0/
|
Due to their low density and large specific surface area, metal foams are increasingly used as cellular materials that combine excellent structural and thermal properties. Their cellular structure makes them particularly suitable for use in heat exchangers, insulation, and fire protection layers. The heat transport that takes place within them is a complex phenomenon characterized by the simultaneous presence of heat conduction, heat transfer, and heat radiation, making their modeling a significant challenge. The aim of the research is to develop a one-dimensional, time-dependent, discrete numerical model capable of describing the effective thermal behavior of metal foams. The model takes into account heat conduction through the solid phase, conductive heat transfer in the closed cavities, thermal radiation between the pore walls, and by-passing heat conduction around the cavity. The results highlight that geometric features such as cavity size and arrangement have a significant impact on temperature distribution and confirm that classical Fourier-based models are not accurately applicable to porous materials, but the Guyer--Krumhansl equation is an adequate candidate. Furthermore, we correlated the static and dynamic thermal diffusivity with the metal foam parameters, showing a significant sensitivity to the cavity thermal attributes.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:14:54 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Fehér",
"Anna",
""
],
[
"Cseh",
"Dániel",
""
],
[
"Kovács",
"Róbert",
""
]
] |
2508.15341
|
Sanjeet Kaur
|
Sanjeet Kaur, Daniel R. Mason, Prashanth Srinivasan, Stephen Dixon, Sid Mungale, Teresa Orr, Mikhail Yu. Lavrentiev, Duc Nguyen-Manh
|
The effect of multi-occupancy traps on the diffusion and retention of multiple hydrogen isotopes in irradiated tungsten and vanadium
|
16 pages, 9 figures
| null | null | null |
cond-mat.mtrl-sci
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a computational scheme for the diffusion and retention of multiple hydrogen isotopes (HI) with multi-occupancy traps parameterized by first principles calculations. We show that it is often acceptable to reduce the complexity of the coupled differential equations for gas evolution by taking the dynamic steady state, a generalisation of the Oriani equilibrium for multiple isotopes and multi-occupancy traps. The effective gas diffusivity varies most with mobile fraction when the total gas concentration approximates the trap density. We show HI binding to a monovacancy in vanadium produces a non-monotonic dependence between diffusivity and gas concentration, unlike the tungsten system. We demonstrate the difference between multiple single occupancy traps and multi-occupancy traps in long-term diffusion dynamics. The applicability of the multi-occupancy, multi-isotope model in steady state is assessed by comparison to an isotope exchange experiment between hydrogen and deuterium in self-ion irradiated tungsten. The vacancy distribution is estimated with molecular dynamics, and the retention across sample depth shows good agreement with experiment using no fitting parameters.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:14:56 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kaur",
"Sanjeet",
""
],
[
"Mason",
"Daniel R.",
""
],
[
"Srinivasan",
"Prashanth",
""
],
[
"Dixon",
"Stephen",
""
],
[
"Mungale",
"Sid",
""
],
[
"Orr",
"Teresa",
""
],
[
"Lavrentiev",
"Mikhail Yu.",
""
],
[
"Nguyen-Manh",
"Duc",
""
]
] |
2508.15342
|
Sandra Albrechtsen
|
Sandra Albrechtsen and James Davies
|
Counterexample to the conjectured coarse grid theorem
| null | null | null | null |
math.CO math.MG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We show that for every $M,A,n \in \mathbb{N}$ there exists a graph $G$ that does not contain the $(154\times 154)$-grid as a $3$-fat minor and is not $(M,A)$-quasi-isometric to a graph with no $K_n$ minor. This refutes the conjectured coarse grid theorem by Georgakopoulos and Papasoglu and the weak fat minor conjecture of Davies, Hickingbotham, Illingworth, and McCarty.
Our construction is a slight modification of the recent counterexample to the weak coarse Menger conjecture from Nguyen, Scott and Seymour.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:16:55 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Albrechtsen",
"Sandra",
""
],
[
"Davies",
"James",
""
]
] |
2508.15343
|
Sundararajan Natarajan
|
Rishi Mishra and Smriti and Ganapathy Krishnamurthi and Balaji Srinivasan and Sundararajan Natarajan
|
Eig-PIELM: A Mesh-Free Approach for Efficient Eigen-Analysis with Physics-Informed Extreme Learning Machines
| null | null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
In this work, a novel Eig-PIELM framework is proposed that extends physics-informed extreme learning machine for an efficient and accurate solution of linear eigenvalue problems. The method reformulates the governing differential equations into a compact algebraic system solvable in a single step. Boundary conditions are enforced exactly via an algebraic projection onto the boundary-admissible subspace, eliminating the computational overhead of penalty parameters, and backpropagation while preserving the computational advantages of extreme learning machines. The proposed framework is mesh-free and yields both eigenvalues and mode shapes simultaneously in one linear solve. The robustness and accuracy of the proposed framework is demonstrated through a range of benchmark problems. We believe that the mesh-free nature, solution structure and accuracy of Eig-PIELM makes it particularly valuable for parametric studies in mechanical, acoustic, and electromechanical systems where rapid frequency spectrum analysis is critical.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:19:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mishra",
"Rishi",
""
],
[
"Smriti",
"",
""
],
[
"Krishnamurthi",
"Ganapathy",
""
],
[
"Srinivasan",
"Balaji",
""
],
[
"Natarajan",
"Sundararajan",
""
]
] |
2508.15344
|
Manabu Moritsu Dr.
|
Manabu Moritsu, Hai-Bo Li, Tianyu Xing, Ye Yuan, Yao Zhang
|
A study to suppress a sneaking cosmic muon background in the COMET experiment
|
8 pages, 5 figures, Proceedings of the 4th J-PARC Symposium 2024
| null | null | null |
hep-ex physics.ins-det
|
http://creativecommons.org/licenses/by/4.0/
|
The COMET experiment, conducted at J-PARC, aims to search for muon-to-electron conversion with an unprecedentedly high sensitivity. One of the severest backgrounds in the Phase-I experiment originates from cosmic-ray muons. A cosmic muon sneaks into a detector solenoid magnet from a loophole, scattered and leaving a track in a cylindrical drift chamber. Among them, a positive muon track with reverse direction may mimic a signal electron of 105 MeV/$c$. In order to suppress the sneaking cosmic positive muon background, we developed a method to discriminate the track direction by using track-fitting quality. We demonstrated that the positive muon background can be reduced by an order of magnitude. In this paper, we will report the methodology, a Monte Carlo simulation and results with prospects.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:21:12 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Moritsu",
"Manabu",
""
],
[
"Li",
"Hai-Bo",
""
],
[
"Xing",
"Tianyu",
""
],
[
"Yuan",
"Ye",
""
],
[
"Zhang",
"Yao",
""
]
] |
2508.15345
|
Bj\"orn Volkmann
|
Bj\"orn Volkmann, Jan-Hendrik Ewering, Michael Meindl, Simon F. G. Ehlers, Thomas Seel
|
Bayesian Inference and Learning in Nonlinear Dynamical Systems: A Framework for Incorporating Explicit and Implicit Prior Knowledge
|
16 pages, Preprint submitted to Automatica
| null | null | null |
stat.ML cs.LG
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Accuracy and generalization capabilities are key objectives when learning dynamical system models. To obtain such models from limited data, current works exploit prior knowledge and assumptions about the system. However, the fusion of diverse prior knowledge, e. g. partially known system equations and smoothness assumptions about unknown model parts, with information contained in the data remains a challenging problem, especially in input-output settings with latent system state. In particular, learning functions that are nested inside known system equations can be a laborious and error-prone expert task. This paper considers inference of latent states and learning of unknown model parts for fusion of data information with different sources of prior knowledge. The main contribution is a general-purpose system identification tool that, for the first time, provides a consistent solution for both, online and offline Bayesian inference and learning while allowing to incorporate explicit and implicit prior system knowledge. We propose a novel interface for combining known dynamics functions with a learning-based approximation of unknown system parts. Based on the proposed model structure, closed-form densities for efficient parameter marginalization are derived. No user-tailored coordinate transformations or model inversions are needed, making the presented framework a general-purpose tool for inference and learning. The broad applicability of the devised framework is illustrated in three distinct case studies, including an experimental data set.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:23:14 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Volkmann",
"Björn",
""
],
[
"Ewering",
"Jan-Hendrik",
""
],
[
"Meindl",
"Michael",
""
],
[
"Ehlers",
"Simon F. G.",
""
],
[
"Seel",
"Thomas",
""
]
] |
2508.15346
|
Ting Lu
|
Ting Lu
|
The Haar State values of monomials and a method to pick orthonormal bases on $O(U_q(3))$
| null | null | null | null |
math.QA math-ph math.MP
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we investigate the evaluation problem of the Haar state on the quantum group $O(U_q(n))$ ($n\ge 3$) which is a $q$-deformation of the Haar measure on the Lie group $U(n)$. The relation between the Haar state values of monomials on $O(U_q(n))$ is studied. On $O(U_q(3))$, the Haar state values of monomials are explicitly computed and these values are expressed as a finite summation of rational polynomials in $q$. As an application, we compute the Gram matrices of the irreducible co-representations of $O(U_q(3))$ which is essential to the method of constructing orthonormal bases on $O(U_q(3))$ proposed by Noumi, Yamada, and Mimachi. New connections between the Haar state values of monomials and basic hypergeometric multi-summations are found during our computation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:24:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lu",
"Ting",
""
]
] |
2508.15347
|
Takahiro Morimoto
|
Takahiro Morimoto, Naoto Nagaosa
|
Role of Ward-Takahashi identity in an electron-phonon coupled system -- Revisiting phonon shift current
|
11 pages, 2 figures
| null | null | null |
cond-mat.str-el
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study bulk photovoltaic effects in electron-phonon coupled systems. The conservation of current or gauge invariance, manifested as the Ward-Takahashi identity, plays an essential role in the analysis of the Feynman diagrams, and the leading order contribution to the phonon shift current is identified accordingly. The leading order contribution essentially arises from the electric polarization carried by optically excited phonons, where the shift current is generated due to a change of electric polarization in the steady state under the optical excitation of phonons.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:24:20 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Morimoto",
"Takahiro",
""
],
[
"Nagaosa",
"Naoto",
""
]
] |
2508.15348
|
Markus Dannem\"uller
|
Markus Dannem\"uller, Tim Netzer
|
Lifts of Operator Systems
| null | null | null | null |
math.OA
|
http://creativecommons.org/licenses/by/4.0/
|
We transfer the theory of slack operators and sums-of-squares-criteria for lifts from convex cones to operator systems. These allow to study the following question, among others: Given an abstract operator system, is its enveloping $C^*$-algebra finite-dimensional, or is it the linear image of a system with a finite-dimensional enveloping algebra?
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:25:06 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Dannemüller",
"Markus",
""
],
[
"Netzer",
"Tim",
""
]
] |
2508.15349
|
Wouter Bos
|
Wouter J.T. Bos (LMFA), Xuan Shao, Tong Wu, Le Fang
|
The Role of Vortex Stretching in Drag Reduction of Polymer-Laden Turbulent Flow
|
Journal of Fluid Mechanics, In press
| null | null |
PMCS2I
|
physics.flu-dyn
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
An addition of polymers can significantly reduce drag in wall-bounded turbulent flows, such as pipes or channels. This phenomenon is accompanied by a noticeable modification of the mean velocity profile. Starting from the premise that polymers reduce vortex-stretching, we derive a theoretical prediction for the mean-velocity profile. After assessing this prediction by numerical experiments of turbulence with reduced vortex stretching, we show that the theory successfully describes experimental measurements of drag-reduction in pipe-flow.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:25:39 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bos",
"Wouter J. T.",
"",
"LMFA"
],
[
"Shao",
"Xuan",
""
],
[
"Wu",
"Tong",
""
],
[
"Fang",
"Le",
""
]
] |
2508.15350
|
Ajit Mehta
|
Ajit Kumar Mehta, Digvijay Wadekar, Isha Anantpurkar, Javier Roulet, Tejaswi Venumadhav, Tousif Islam, Jonathan Mushkin, Barak Zackay, and Matias Zaldarriaga
|
Binary black hole population inference combining confident and marginal events from the $\tt{IAS\text{-}HM}$ search pipeline
|
13 pages, 5 figures
| null | null | null |
gr-qc astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
We present the population properties of binary black hole mergers identified by the $\tt{IAS\text{-}HM}$ pipeline (which incorporates higher-order modes in the search templates) during the third observing run (O3) of the LIGO, Virgo, and KAGRA (LVK) detectors. In our population inference analysis, instead of only using events above a sharp cut based on a particular detection threshold (e.g., false alarm rate), we use a Bayesian framework to consistently include both marginal and confident events. We find that our inference based solely on highly significant events ($p_{\mathrm{astro}} \sim 1$) is broadly consistent with the GWTC-3 population analysis performed by the LVK collaboration. However, incorporating marginal events into the analysis leads to a preference for stronger redshift evolution in the merger rate and an increased density of asymmetric mass-ratio mergers relative to the GWTC-3 analysis, while remaining within its allowed parameter ranges. Using simple parametric models to describe the binary black hole population, we estimate a merger rate density of $32.4^{+18.5}_{-12.2}\ \mathrm{Gpc}^{-3}\,\mathrm{yr}^{-1}$ at redshift $z = 0.2$, and a redshift evolution parameter of $\kappa = 4.4^{+1.9}_{-2.0}$. Assuming a power-law form for the mass ratio distribution ($\propto q^{\beta}$), we infer $\beta = 0.1^{+1.9}_{-1.4}$, indicating a relatively flat distribution. These results highlight the potential impact of marginal events on population inferences and motivate future analyses with data from upcoming observing runs.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:28:27 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mehta",
"Ajit Kumar",
""
],
[
"Wadekar",
"Digvijay",
""
],
[
"Anantpurkar",
"Isha",
""
],
[
"Roulet",
"Javier",
""
],
[
"Venumadhav",
"Tejaswi",
""
],
[
"Islam",
"Tousif",
""
],
[
"Mushkin",
"Jonathan",
""
],
[
"Zackay",
"Barak",
""
],
[
"Zaldarriaga",
"Matias",
""
]
] |
2508.15351
|
Cynthia Marcelino
|
Cynthia Marcelino, Leonard Guelmino, Thomas Pusztai, Stefan Nastic
|
Databelt: A Continuous Data Path for Serverless Workflows in the 3D Compute Continuum
| null | null | null | null |
cs.DC
|
http://creativecommons.org/licenses/by/4.0/
|
Typically, serverless functions rely on remote storage services for managing state, which can result in increased latency and network communication overhead. In a dynamic environment such as the 3D (Edge-Cloud-Space) Compute Continuum, serverless functions face additional challenges due to frequent changes in network topology. As satellites move in and out of the range of ground stations, functions must make multiple hops to access cloud services, leading to high-latency state access and unnecessary data transfers. In this paper, we present Databelt, a state management framework for serverless workflows designed for the dynamic environment of the 3D Compute Continuum. Databelt introduces an SLO-aware state propagation mechanism that enables the function state to move continuously in orbit. Databelt proactively offloads function states to the most suitable node, such that when functions execute, the data is already present on the execution node or nearby, thus minimizing state access latency and reducing the number of network hops. Additionally, Databelt introduces a function state fusion mechanism that abstracts state management for functions sharing the same serverless runtime. When functions are fused, Databelt seamlessly retrieves their state as a group, reducing redundant network and storage operations and improving overall workflow efficiency. Our experimental results show that Databelt reduces workflow execution time by up to 66% and increases throughput by 50% compared to the baselines. Furthermore, our results show that Databelt function state fusion reduces storage operations latency by up to 20%, by reducing repetitive storage requests for functions within the same runtime, ensuring efficient execution of serverless workflows in highly dynamic network environments such as the 3D Continuum.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:29:01 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Marcelino",
"Cynthia",
""
],
[
"Guelmino",
"Leonard",
""
],
[
"Pusztai",
"Thomas",
""
],
[
"Nastic",
"Stefan",
""
]
] |
2508.15352
|
Sang Kyu Kim
|
Sang Kyu Kim, Eduardo Zubizarreta Casalengua, Yeji Sim, Friedrich Sbresny, Carolin Calcagno, Hubert Riedl, Jonathan J. Finley, Elena del Valle, Carlos Ant\'on-Solanas, Kai M\"uller, and Lukas Hanschke
|
Deterministic Control of Photon-Number Probabilities via Phase-Controlled Quantum Interference
|
14 pages, 11 figures
| null | null | null |
quant-ph cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
Deterministically tailoring optical Fock states beyond the single-photon level is crucial for boson sampling, loss-tolerant photonic qubits, and quantum-enhanced sensing, however has yet remained elusive. Here, we report an all-linear-optical protocol that converts a resonantly driven single-photon emitter into a deterministic generator of vacuum--single-photon--two-photon states. A phase-stabilized, path-unbalanced Mach-Zehnder interferometer combines vacuum--single-photon interference and Hong-Ou-Mandel effect, providing two knobs to shape photon-number probabilities. By tuning these knobs, we observe a dynamic transition from antibunching to strong bunching in correlation measurements. A fully quantum-mechanical, discrete time-bin model maps these results onto the tailored photon statistics. The same framework predicts that two indistinguishable emitters would extend the accessible space to deterministic NOON states and single-photon filtering. This protocol relying on linear optics and available single-photon sources provides a scalable, chip-compatible, and platform-independent route to on-demand and deterministic few-photon resources for quantum metrology, photonic computing, as well as long-distance quantum networks.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:32:03 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kim",
"Sang Kyu",
""
],
[
"Casalengua",
"Eduardo Zubizarreta",
""
],
[
"Sim",
"Yeji",
""
],
[
"Sbresny",
"Friedrich",
""
],
[
"Calcagno",
"Carolin",
""
],
[
"Riedl",
"Hubert",
""
],
[
"Finley",
"Jonathan J.",
""
],
[
"del Valle",
"Elena",
""
],
[
"Antón-Solanas",
"Carlos",
""
],
[
"Müller",
"Kai",
""
],
[
"Hanschke",
"Lukas",
""
]
] |
2508.15353
|
Olga Matykina
|
Olga Matykina and Dmitry Yudin
|
RCDINO: Enhancing Radar-Camera 3D Object Detection with DINOv2 Semantic Features
|
Accepted for publication in Optical Memory and Neural Networks, 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Three-dimensional object detection is essential for autonomous driving and robotics, relying on effective fusion of multimodal data from cameras and radar. This work proposes RCDINO, a multimodal transformer-based model that enhances visual backbone features by fusing them with semantically rich representations from the pretrained DINOv2 foundation model. This approach enriches visual representations and improves the model's detection performance while preserving compatibility with the baseline architecture. Experiments on the nuScenes dataset demonstrate that RCDINO achieves state-of-the-art performance among radar-camera models, with 56.4 NDS and 48.1 mAP. Our implementation is available at https://github.com/OlgaMatykina/RCDINO.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:33:36 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Matykina",
"Olga",
""
],
[
"Yudin",
"Dmitry",
""
]
] |
2508.15354
|
Chaoran Xiong
|
Chaoran Xiong, Yulong Huang, Fangwen Yu, Changhao Chen, Yue Wang, Songpengchen Xia, Ling Pei
|
Sensing, Social, and Motion Intelligence in Embodied Navigation: A Comprehensive Survey
| null | null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Embodied navigation (EN) advances traditional navigation by enabling robots to perform complex egocentric tasks through sensing, social, and motion intelligence. In contrast to classic methodologies that rely on explicit localization and pre-defined maps, EN leverages egocentric perception and human-like interaction strategies. This survey introduces a comprehensive EN formulation structured into five stages: Transition, Observation, Fusion, Reward-policy construction, and Action (TOFRA). The TOFRA framework serves to synthesize the current state of the art, provide a critical review of relevant platforms and evaluation metrics, and identify critical open research challenges. A list of studies is available at https://github.com/Franky-X/Awesome-Embodied-Navigation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:33:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xiong",
"Chaoran",
""
],
[
"Huang",
"Yulong",
""
],
[
"Yu",
"Fangwen",
""
],
[
"Chen",
"Changhao",
""
],
[
"Wang",
"Yue",
""
],
[
"Xia",
"Songpengchen",
""
],
[
"Pei",
"Ling",
""
]
] |
2508.15355
|
Wenyuan Li
|
Liyuan Cui, Wenyuan Li
|
Demand for catastrophe insurance under the path-dependent effects
| null | null | null | null |
q-fin.RM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper investigates optimal investment and insurance strategies under a mean-variance criterion with path-dependent effects. We use a rough volatility model and a Hawkes process with a power kernel to capture the path dependence of the market. By adding auxiliary state variables, we degenerate a non-Markovian problem to a Markovian problem. Next, an explicit solution is derived for a path-dependent extended Hamilton-Jacobi-Bellman (HJB) equation. Then, we derive the explicit solutions of the problem by extending the Functional Ito formula for fractional Brownian motion to the general path-dependent processes, which includes the Hawkes process. In addition, we use earthquake data from Sichuan Province, China, to estimate parameters for the Hawkes process. Our numerical results show that the individual becomes more risk-averse in trading when stock volatility is rough, while more risk-seeking when considering catastrophic shocks. Moreover, an individual's demand for catastrophe insurance increases with path-dependent effects. Our findings indicate that ignoring the path-dependent effect would lead to a significant underinsurance phenomenon and highlight the importance of the path-dependent effect in the catastrophe insurance pricing.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:34:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Cui",
"Liyuan",
""
],
[
"Li",
"Wenyuan",
""
]
] |
2508.15356
|
L\'eonard Brice
|
Ali Asadi, L\'eonard Brice, Krishnendu Chatterjee, K. S. Thejaswini
|
{\epsilon}-Stationary Nash Equilibria in Multi-player Stochastic Graph Games
| null | null | null | null |
cs.GT
|
http://creativecommons.org/licenses/by/4.0/
|
A strategy profile in a multi-player game is a Nash equilibrium if no player can unilaterally deviate to achieve a strictly better payoff. A profile is an $\epsilon$-Nash equilibrium if no player can gain more than $\epsilon$ by unilaterally deviating from their strategy. In this work, we use $\epsilon$-Nash equilibria to approximate the computation of Nash equilibria. Specifically, we focus on turn-based, multiplayer stochastic games played on graphs, where players are restricted to stationary strategies -- strategies that use randomness but not memory.
The problem of deciding the constrained existence of stationary Nash equilibria -- where each player's payoff must lie within a given interval -- is known to be $\exists\mathbb{R}$-complete in such a setting (Hansen and S{\o}lvsten, 2020). We extend this line of work to stationary $\epsilon$-Nash equilibria and present an algorithm that solves the following promise problem: given a game with a Nash equilibrium satisfying the constraints, compute an $\epsilon$-Nash equilibrium that $\epsilon$-satisfies those same constraints -- satisfies the constraints up to an $\epsilon$ additive error. Our algorithm runs in FNP^NP time.
To achieve this, we first show that if a constrained Nash equilibrium exists, then one exists where the non-zero probabilities are at least an inverse of a double-exponential in the input. We further prove that such a strategy can be encoded using floating-point representations, as in the work of Frederiksen and Miltersen (2013), which finally gives us our FNP^NP algorithm.
We further show that the decision version of the promise problem is NP-hard. Finally, we show a partial tightness result by proving a lower bound for such techniques: if a constrained Nash equilibrium exists, then there must be one that where the probabilities in the strategies are double-exponentially small.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:34:45 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Asadi",
"Ali",
""
],
[
"Brice",
"Léonard",
""
],
[
"Chatterjee",
"Krishnendu",
""
],
[
"Thejaswini",
"K. S.",
""
]
] |
2508.15357
|
Haji Gul
|
Haji Gul, Abul Ghani Naim, Ajaz Ahmad Bhat
|
KG-EDAS: A Meta-Metric Framework for Evaluating Knowledge Graph Completion Models
| null | null | null | null |
cs.CL cs.PF
|
http://creativecommons.org/licenses/by/4.0/
|
Knowledge Graphs (KGs) enable applications in various domains such as semantic search, recommendation systems, and natural language processing. KGs are often incomplete, missing entities and relations, an issue addressed by Knowledge Graph Completion (KGC) methods that predict missing elements. Different evaluation metrics, such as Mean Reciprocal Rank (MRR), Mean Rank (MR), and Hit@k, are commonly used to assess the performance of such KGC models. A major challenge in evaluating KGC models, however, lies in comparing their performance across multiple datasets and metrics. A model may outperform others on one dataset but underperform on another, making it difficult to determine overall superiority. Moreover, even within a single dataset, different metrics such as MRR and Hit@1 can yield conflicting rankings, where one model excels in MRR while another performs better in Hit@1, further complicating model selection for downstream tasks. These inconsistencies hinder holistic comparisons and highlight the need for a unified meta-metric that integrates performance across all metrics and datasets to enable a more reliable and interpretable evaluation framework. To address this need, we propose KG Evaluation based on Distance from Average Solution (EDAS), a robust and interpretable meta-metric that synthesizes model performance across multiple datasets and diverse evaluation criteria into a single normalized score ($M_i \in [0,1]$). Unlike traditional metrics that focus on isolated aspects of performance, EDAS offers a global perspective that supports more informed model selection and promotes fairness in cross-dataset evaluation. Experimental results on benchmark datasets such as FB15k-237 and WN18RR demonstrate that EDAS effectively integrates multi-metric, multi-dataset performance into a unified ranking, offering a consistent, robust, and generalizable framework for evaluating KGC models.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:37:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gul",
"Haji",
""
],
[
"Naim",
"Abul Ghani",
""
],
[
"Bhat",
"Ajaz Ahmad",
""
]
] |
2508.15358
|
Alberto Pozanco
|
Alberto Pozanco, Marianela Morales, Daniel Borrajo, Manuela Veloso
|
Planning with Minimal Disruption
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In many planning applications, we might be interested in finding plans that minimally modify the initial state to achieve the goals. We refer to this concept as plan disruption. In this paper, we formally introduce it, and define various planning-based compilations that aim to jointly optimize both the sum of action costs and plan disruption. Experimental results in different benchmarks show that the reformulated task can be effectively solved in practice to generate plans that balance both objectives.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:38:17 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Pozanco",
"Alberto",
""
],
[
"Morales",
"Marianela",
""
],
[
"Borrajo",
"Daniel",
""
],
[
"Veloso",
"Manuela",
""
]
] |
2508.15359
|
Qian Huang
|
Qian Huang and Christian Rohde
|
Statistical conservation laws for scalar model problems: Hierarchical evolution equations
| null | null | null | null |
math.AP cs.NA math.NA
|
http://creativecommons.org/licenses/by/4.0/
|
The probability density functions (PDFs) for the solution of the incompressible Navier-Stokes equation can be represented by a hierarchy of linear equations. This article develops new hierarchical evolution equations for PDFs of a scalar conservation law with random initial data as a model problem. Two frameworks are developed, including multi-point PDFs and single-point higher-order derivative PDFs. These hierarchies capture statistical correlations and guide closure strategies.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:41:03 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Huang",
"Qian",
""
],
[
"Rohde",
"Christian",
""
]
] |
2508.15360
|
Chenhui Gou
|
Chenhui Gou, Ziyu Ma, Zicheng Duan, Haoyu He, Feng Chen, Akide Liu, Bohan Zhuang, Jianfei Cai, Hamid Rezatofighi
|
An Empirical Study on How Video-LLMs Answer Video Questions
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Taking advantage of large-scale data and pretrained language models, Video Large Language Models (Video-LLMs) have shown strong capabilities in answering video questions. However, most existing efforts focus on improving performance, with limited attention to understanding their internal mechanisms. This paper aims to bridge this gap through a systematic empirical study. To interpret existing VideoLLMs, we adopt attention knockouts as our primary analytical tool and design three variants: Video Temporal Knockout, Video Spatial Knockout, and Language-to-Video Knockout. Then, we apply these three knockouts on different numbers of layers (window of layers). By carefully controlling the window of layers and types of knockouts, we provide two settings: a global setting and a fine-grained setting. Our study reveals three key findings: (1) Global setting indicates Video information extraction primarily occurs in early layers, forming a clear two-stage process -- lower layers focus on perceptual encoding, while higher layers handle abstract reasoning; (2) In the fine-grained setting, certain intermediate layers exert an outsized impact on video question answering, acting as critical outliers, whereas most other layers contribute minimally; (3) In both settings, we observe that spatial-temporal modeling relies more on language-guided retrieval than on intra- and inter-frame self-attention among video tokens, despite the latter's high computational cost. Finally, we demonstrate that these insights can be leveraged to reduce attention computation in Video-LLMs. To our knowledge, this is the first work to systematically uncover how Video-LLMs internally process and understand video content, offering interpretability and efficiency perspectives for future research.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:42:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Gou",
"Chenhui",
""
],
[
"Ma",
"Ziyu",
""
],
[
"Duan",
"Zicheng",
""
],
[
"He",
"Haoyu",
""
],
[
"Chen",
"Feng",
""
],
[
"Liu",
"Akide",
""
],
[
"Zhuang",
"Bohan",
""
],
[
"Cai",
"Jianfei",
""
],
[
"Rezatofighi",
"Hamid",
""
]
] |
2508.15361
|
Shiwen Ni
|
Shiwen Ni, Guhong Chen, Shuaimin Li, Xuanang Chen, Siyi Li, Bingli Wang, Qiyao Wang, Xingjian Wang, Yifan Zhang, Liyang Fan, Chengming Li, Ruifeng Xu, Le Sun, Min Yang
|
A Survey on Large Language Model Benchmarks
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In recent years, with the rapid development of the depth and breadth of large language models' capabilities, various corresponding evaluation benchmarks have been emerging in increasing numbers. As a quantitative assessment tool for model performance, benchmarks are not only a core means to measure model capabilities but also a key element in guiding the direction of model development and promoting technological innovation. We systematically review the current status and development of large language model benchmarks for the first time, categorizing 283 representative benchmarks into three categories: general capabilities, domain-specific, and target-specific. General capability benchmarks cover aspects such as core linguistics, knowledge, and reasoning; domain-specific benchmarks focus on fields like natural sciences, humanities and social sciences, and engineering technology; target-specific benchmarks pay attention to risks, reliability, agents, etc. We point out that current benchmarks have problems such as inflated scores caused by data contamination, unfair evaluation due to cultural and linguistic biases, and lack of evaluation on process credibility and dynamic environments, and provide a referable design paradigm for future benchmark innovation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:43:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ni",
"Shiwen",
""
],
[
"Chen",
"Guhong",
""
],
[
"Li",
"Shuaimin",
""
],
[
"Chen",
"Xuanang",
""
],
[
"Li",
"Siyi",
""
],
[
"Wang",
"Bingli",
""
],
[
"Wang",
"Qiyao",
""
],
[
"Wang",
"Xingjian",
""
],
[
"Zhang",
"Yifan",
""
],
[
"Fan",
"Liyang",
""
],
[
"Li",
"Chengming",
""
],
[
"Xu",
"Ruifeng",
""
],
[
"Sun",
"Le",
""
],
[
"Yang",
"Min",
""
]
] |
2508.15362
|
Will Johnson
|
Will Johnson
|
Largeness and generalized t-henselianity
|
17 pages. Addendum to arXiv:2508.10886 [math.LO]
| null | null | null |
math.LO math.AC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Let $K$ be a countable field. Then $K$ is large in the sense of Pop if and only if it admits a field topology which is "generalized t-henselian" (gt-henselian) in the sense of Dittmann, Walsberg, and Ye, meaning that the implicit function theorem holds for polynomials. Moreover, the \'etale open topology can be characterized in terms of the gt-henselian topologies on $K$: a subset $U \subseteq K^n$ is open in the \'etale open topology if and only if it is open with respect to every gt-henselian topology on $K$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:44:22 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Johnson",
"Will",
""
]
] |
2508.15363
|
Ninh Tran
|
Ninh Tran
|
An adaptive procedure for detecting replicated signals with $k$-family-wise error rate control
| null | null | null | null |
stat.ME
|
http://creativecommons.org/licenses/by/4.0/
|
Partial conjunction (PC) hypothesis testing is widely used to assess the replicability of scientific findings across multiple comparable studies. In high-throughput meta-analyses, testing a large number of PC hypotheses with k-family-wise error rate (k-FWER) control often suffers from low statistical power due to the multiplicity burden. The state-of-the-art AdaFilter-Bon procedure by Wang et al. (2022, Ann. Stat., 50(4), 1890-1909) alleviates this problem by filtering out hypotheses unlikely to be false before applying a rejection rule. However, a side effect of filtering is that it renders the rejection rule more stringent than necessary, leading to conservative k-FWER control. In this paper, we mitigate this conservativeness - and thereby improve the power of AdaFilter-Bon - by incorporating a post-filter null proportion estimate into the procedure. The resulting method, AdaFilter-AdaBon, has proven asymptotic k-FWER control under weak dependence and demonstrates empirical finite-sample control with higher power than the original AdaFilter-Bon in simulations.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:44:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Tran",
"Ninh",
""
]
] |
2508.15364
|
Saleh Afzoon
|
Saleh Afzoon, Amin Beheshti, Nabi Rezvani, Farshad Khunjush, Usman Naseem, John McMahon, Zahra Fathollahi, Mahdieh Labani, Wathiq Mansoor, Xuyun Zhang
|
ExBigBang: A Dynamic Approach for Explainable Persona Classification through Contextualized Hybrid Transformer Analysis
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
In user-centric design, persona development plays a vital role in understanding user behaviour, capturing needs, segmenting audiences, and guiding design decisions. However, the growing complexity of user interactions calls for a more contextualized approach to ensure designs align with real user needs. While earlier studies have advanced persona classification by modelling user behaviour, capturing contextual information, especially by integrating textual and tabular data, remains a key challenge. These models also often lack explainability, leaving their predictions difficult to interpret or justify. To address these limitations, we present ExBigBang (Explainable BigBang), a hybrid text-tabular approach that uses transformer-based architectures to model rich contextual features for persona classification. ExBigBang incorporates metadata, domain knowledge, and user profiling to embed deeper context into predictions. Through a cyclical process of user profiling and classification, our approach dynamically updates to reflect evolving user behaviours. Experiments on a benchmark persona classification dataset demonstrate the robustness of our model. An ablation study confirms the benefits of combining text and tabular data, while Explainable AI techniques shed light on the rationale behind the model's predictions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:45:04 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Afzoon",
"Saleh",
""
],
[
"Beheshti",
"Amin",
""
],
[
"Rezvani",
"Nabi",
""
],
[
"Khunjush",
"Farshad",
""
],
[
"Naseem",
"Usman",
""
],
[
"McMahon",
"John",
""
],
[
"Fathollahi",
"Zahra",
""
],
[
"Labani",
"Mahdieh",
""
],
[
"Mansoor",
"Wathiq",
""
],
[
"Zhang",
"Xuyun",
""
]
] |
2508.15365
|
Mitchell Griggs Dr
|
Mitchell T. Griggs, Kevin Burrage, Pamela M. Burrage
|
Implementation of Milstein Schemes for Stochastic Delay-Differential Equations with Arbitrary Fixed Delays
|
22 pages, 9 figures but 23 subfigures
| null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
This paper develops methods for numerically solving stochastic delay-differential equations (SDDEs) with multiple fixed delays that do not align with a uniform time mesh. We focus on numerical schemes of strong convergence orders $1/2$ and $1$, such as the Euler--Maruyama and Milstein schemes, respectively. Although numerical schemes for SDDEs with delays $\tau_1,\ldots,\tau_K$ are theoretically established, their implementations require evaluations at both present times such as $t_n$, and also at delayed times such as $t_n-\tau_k$ and $t_n-\tau_l-\tau_k$. As a result, previous simulations of these schemes have been largely restricted to the case of divisible delays. We develop simulation techniques for the general case of indivisible delays where delayed times such as $t_n-\tau_k$ are not restricted to a uniform time mesh. To achieve order of convergence (OoC) $1/2$, we implement the schemes with a fixed step size while using linear interpolation to approximate delayed scheme values. To achieve OoC $1$, we construct an augmented time mesh that includes all time points required to evaluate the schemes, which necessitates using a varying step size. We also introduce a technique to simulate delayed iterated stochastic integrals on the augmented time mesh, by extending an established method from the divisible-delays setting. We then confirm that the numerical schemes achieve their theoretical convergence orders with computational examples.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:50:15 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Griggs",
"Mitchell T.",
""
],
[
"Burrage",
"Kevin",
""
],
[
"Burrage",
"Pamela M.",
""
]
] |
2508.15366
|
Jayam Joshi
|
Jayam Joshi and Shradha Mishra
|
Analytical Theory of Chiral Active Particle Transport in a Fluctuating Density Field
| null | null | null | null |
cond-mat.stat-mech
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We develop a closed-form analytical theory for the transport of a chiral active Brownian particle (cABP) in three dimensions, moving through a fluctuating local density field that coarse-grains steric and dynamical interactions in a dense active medium. The density field is modeled as an Ornstein--Uhlenbeck process with finite correlation time $\tau$ and fluctuation strength $\sigma_\rho^2$, capturing both spatial variations and temporal memory. Within this framework, we derive exact expressions for the mean-squared displacement (MSD) and time-dependent diffusivity, showing how chirality and density coupling renormalize orientational persistence and generate dynamical crossovers. The theory predicts: (i) anomalously high initial diffusivity in denser regions, arising from a transient drift driven by local swim-pressure gradients; (ii) a finite crossover time $t_c$ for homogenizing density inhomogeneities, with steady-state diffusivity retaining memory of the initial environment; (iii) a non-monotonic $t_c(\Omega)$ with a global minimum at intermediate chirality, and a three-regime suppression of $D_\infty(\Omega)$ consistent with clustered phases in simulations; and (iv) a resonance-like peak in early-time oscillatory transport at an optimal chirality $\Omega^*$. The framework reproduces known scaling of $D_\infty$ with activity and chirality, while uncovering new memory effects and chirality-controlled optimal transport, offering predictive insight for biological circle swimmers and active metamaterials.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:50:37 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Joshi",
"Jayam",
""
],
[
"Mishra",
"Shradha",
""
]
] |
2508.15367
|
Jacinto Colan
|
Jacinto Colan, Ana Davila, Yasuhisa Hasegawa
|
Transfer learning optimization based on evolutionary selective fine tuning
|
Presented at the Workshop artiFicial And bio-inspIred netwoRked intelliGence foR cOnstrained aUtoNomous Devices (FAIRGROUND). 2025 International Joint Conference on Neural Networks (IJCNN)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Deep learning has shown substantial progress in image analysis. However, the computational demands of large, fully trained models remain a consideration. Transfer learning offers a strategy for adapting pre-trained models to new tasks. Traditional fine-tuning often involves updating all model parameters, which can potentially lead to overfitting and higher computational costs. This paper introduces BioTune, an evolutionary adaptive fine-tuning technique that selectively fine-tunes layers to enhance transfer learning efficiency. BioTune employs an evolutionary algorithm to identify a focused set of layers for fine-tuning, aiming to optimize model performance on a given target task. Evaluation across nine image classification datasets from various domains indicates that BioTune achieves competitive or improved accuracy and efficiency compared to existing fine-tuning methods such as AutoRGN and LoRA. By concentrating the fine-tuning process on a subset of relevant layers, BioTune reduces the number of trainable parameters, potentially leading to decreased computational cost and facilitating more efficient transfer learning across diverse data characteristics and distributions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:51:43 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Colan",
"Jacinto",
""
],
[
"Davila",
"Ana",
""
],
[
"Hasegawa",
"Yasuhisa",
""
]
] |
2508.15368
|
Merzuk Kaltak PhD
|
Merzuk Kaltak, Alexander Hampel, Martin Schlipf, Indukuru Ramesh Reddy, Bongae Kim, Georg Kresse
|
Constrained Random Phase Approximation: the spectral method
|
12 pages, 5 figures, to be published in Physical Review B, Poster presented at PSI-K, Lausanne (August 2025). Talk given at Workshop "The determination of Hubbard parameters: progress, pitfalls, and prospects", Gandia (September 2025)
| null | null | null |
cond-mat.str-el cond-mat.other physics.atm-clus physics.comp-ph quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a new constrained Random Phase Approximation (cRPA) method, termed spectral cRPA (s-cRPA), and compare it to established cRPA approaches for Scandium and Copper by varying the 3d shell filling. The s-cRPA consistently yields larger Hubbard $U$ interaction values. Applied to the realistic system CaFeO$_3$, s-cRPA produces interaction parameters significantly closer to those required within DFT+$U$ to induce the experimentally observed insulating state, overcoming the metallic behavior predicted by standard density functionals. We also address the issue of negative interaction values found in the projector cRPA method for filled d-shells, demonstrating that s-cRPA offers superior numerical stability through electron conversion. Overall, s-cRPA is more robust and effectively overcomes the known underestimation of $U$ in standard cRPA, making it a promising tool for the community. Additionally, we have enhanced our implementation with features for computing multi-center interactions to analyze spatial decay and developed a low-scaling variant with a compressed Matsubara grid to efficiently obtain full frequency-dependent interactions.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:51:49 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kaltak",
"Merzuk",
""
],
[
"Hampel",
"Alexander",
""
],
[
"Schlipf",
"Martin",
""
],
[
"Reddy",
"Indukuru Ramesh",
""
],
[
"Kim",
"Bongae",
""
],
[
"Kresse",
"Georg",
""
]
] |
2508.15369
|
Yonathan Guttel
|
Yonathan Guttel, Orit Moradov, Nachi Lieder, Asnat Greenstein-Messica
|
Enhancing Forecasting with a 2D Time Series Approach for Cohort-Based Data
|
Accepted at IEEE CiFer Companion 2025. 5 pages, 3 figures, 2 tables
|
2025 IEEE Symposium on Computational Intelligence for Financial Engineering and Economics Companion (CiFer Companion)
|
10.1109/CiFerCompanion65204.2025.10980398
| null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper introduces a novel two-dimensional (2D) time series forecasting model that integrates cohort behavior over time, addressing challenges in small data environments. We demonstrate its efficacy using multiple real-world datasets, showcasing superior performance in accuracy and adaptability compared to reference models. The approach offers valuable insights for strategic decision-making across industries facing financial and marketing forecasting challenges.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 08:53:40 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Guttel",
"Yonathan",
""
],
[
"Moradov",
"Orit",
""
],
[
"Lieder",
"Nachi",
""
],
[
"Greenstein-Messica",
"Asnat",
""
]
] |
2508.15370
|
Yichi Zhang
|
Yichi Zhang, Yao Huang, Yifan Wang, Yitong Sun, Chang Liu, Zhe Zhao, Zhengwei Fang, Huanran Chen, Xiao Yang, Xingxing Wei, Hang Su, Yinpeng Dong, Jun Zhu
|
Unveiling Trust in Multimodal Large Language Models: Evaluation, Analysis, and Mitigation
|
For Appendix, please refer to arXiv:2406.07057
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by-sa/4.0/
|
The trustworthiness of Multimodal Large Language Models (MLLMs) remains an intense concern despite the significant progress in their capabilities. Existing evaluation and mitigation approaches often focus on narrow aspects and overlook risks introduced by the multimodality. To tackle these challenges, we propose MultiTrust-X, a comprehensive benchmark for evaluating, analyzing, and mitigating the trustworthiness issues of MLLMs. We define a three-dimensional framework, encompassing five trustworthiness aspects which include truthfulness, robustness, safety, fairness, and privacy; two novel risk types covering multimodal risks and cross-modal impacts; and various mitigation strategies from the perspectives of data, model architecture, training, and inference algorithms. Based on the taxonomy, MultiTrust-X includes 32 tasks and 28 curated datasets, enabling holistic evaluations over 30 open-source and proprietary MLLMs and in-depth analysis with 8 representative mitigation methods. Our extensive experiments reveal significant vulnerabilities in current models, including a gap between trustworthiness and general capabilities, as well as the amplification of potential risks in base LLMs by both multimodal training and inference. Moreover, our controlled analysis uncovers key limitations in existing mitigation strategies that, while some methods yield improvements in specific aspects, few effectively address overall trustworthiness, and many introduce unexpected trade-offs that compromise model utility. These findings also provide practical insights for future improvements, such as the benefits of reasoning to better balance safety and performance. Based on these insights, we introduce a Reasoning-Enhanced Safety Alignment (RESA) approach that equips the model with chain-of-thought reasoning ability to discover the underlying risks, achieving state-of-the-art results.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:00:01 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Yichi",
""
],
[
"Huang",
"Yao",
""
],
[
"Wang",
"Yifan",
""
],
[
"Sun",
"Yitong",
""
],
[
"Liu",
"Chang",
""
],
[
"Zhao",
"Zhe",
""
],
[
"Fang",
"Zhengwei",
""
],
[
"Chen",
"Huanran",
""
],
[
"Yang",
"Xiao",
""
],
[
"Wei",
"Xingxing",
""
],
[
"Su",
"Hang",
""
],
[
"Dong",
"Yinpeng",
""
],
[
"Zhu",
"Jun",
""
]
] |
2508.15371
|
Jaydip Sen
|
Jaydip Sen, Subhasis Dasgupta, Hetvi Waghela
|
Confidence-Modulated Speculative Decoding for Large Language Models
|
This is the preprint of the paper, which has been accepted for oral presentation and publication in the proceedings of IEEE INDISCON 2025. The conference will be organized at the National Institute of Technology, Rourkela, India, from August 21 to 23, 2025. The paper is 10 pages long, and it contains 2 figures and 5 tables
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Speculative decoding has emerged as an effective approach for accelerating autoregressive inference by parallelizing token generation through a draft-then-verify paradigm. However, existing methods rely on static drafting lengths and rigid verification criteria, limiting their adaptability across varying model uncertainties and input complexities. This paper proposes an information-theoretic framework for speculative decoding based on confidence-modulated drafting. By leveraging entropy and margin-based uncertainty measures over the drafter's output distribution, the proposed method dynamically adjusts the number of speculatively generated tokens at each iteration. This adaptive mechanism reduces rollback frequency, improves resource utilization, and maintains output fidelity. Additionally, the verification process is modulated using the same confidence signals, enabling more flexible acceptance of drafted tokens without sacrificing generation quality. Experiments on machine translation and summarization tasks demonstrate significant speedups over standard speculative decoding while preserving or improving BLEU and ROUGE scores. The proposed approach offers a principled, plug-in method for efficient and robust decoding in large language models under varying conditions of uncertainty.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:06:31 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Sen",
"Jaydip",
""
],
[
"Dasgupta",
"Subhasis",
""
],
[
"Waghela",
"Hetvi",
""
]
] |
2508.15372
|
Xinshuang Liu
|
Xinshuang Liu, Runfa Blark Li, Keito Suzuki, Truong Nguyen
|
Image-Conditioned 3D Gaussian Splat Quantization
| null | null | null | null |
cs.CV cs.AI cs.GR
|
http://creativecommons.org/licenses/by/4.0/
|
3D Gaussian Splatting (3DGS) has attracted considerable attention for enabling high-quality real-time rendering. Although 3DGS compression methods have been proposed for deployment on storage-constrained devices, two limitations hinder archival use: (1) they compress medium-scale scenes only to the megabyte range, which remains impractical for large-scale scenes or extensive scene collections; and (2) they lack mechanisms to accommodate scene changes after long-term archival. To address these limitations, we propose an Image-Conditioned Gaussian Splat Quantizer (ICGS-Quantizer) that substantially enhances compression efficiency and provides adaptability to scene changes after archiving. ICGS-Quantizer improves quantization efficiency by jointly exploiting inter-Gaussian and inter-attribute correlations and by using shared codebooks across all training scenes, which are then fixed and applied to previously unseen test scenes, eliminating the overhead of per-scene codebooks. This approach effectively reduces the storage requirements for 3DGS to the kilobyte range while preserving visual fidelity. To enable adaptability to post-archival scene changes, ICGS-Quantizer conditions scene decoding on images captured at decoding time. The encoding, quantization, and decoding processes are trained jointly, ensuring that the codes, which are quantized representations of the scene, are effective for conditional decoding. We evaluate ICGS-Quantizer on 3D scene compression and 3D scene updating. Experimental results show that ICGS-Quantizer consistently outperforms state-of-the-art methods in compression efficiency and adaptability to scene changes. Our code, model, and data will be publicly available on GitHub.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:07:26 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Xinshuang",
""
],
[
"Li",
"Runfa Blark",
""
],
[
"Suzuki",
"Keito",
""
],
[
"Nguyen",
"Truong",
""
]
] |
2508.15373
|
J. Fernando Barbero G.
|
J. Fernando Barbero G., Juan Margalef-Bentabol, Aitor Vicente-Cano, Eduardo J.S. Villase\~nor
|
Husain-Kucha\v{r} model as the Carrollian limit of the Holst term
|
13 pages
| null | null | null |
gr-qc hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We show how the Husain-Kucha\v{r} model can be understood as a Carrollian limit of the Holst term in the context of background-independent field theories described in terms of coframes and spin connections. We also discuss the footprint of the Carrollian symmetry in the Hamiltonian formulation of the Husain-Kucha\v{r} action.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:08:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"G.",
"J. Fernando Barbero",
""
],
[
"Margalef-Bentabol",
"Juan",
""
],
[
"Vicente-Cano",
"Aitor",
""
],
[
"Villaseñor",
"Eduardo J. S.",
""
]
] |
2508.15374
|
Omri Ben-Dov
|
Omri Ben-Dov, Samira Samadi, Amartya Sanyal, Alexandru \c{T}ifrea
|
Fairness for the People, by the People: Minority Collective Action
| null | null | null | null |
cs.LG cs.CY
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Machine learning models often preserve biases present in training data, leading to unfair treatment of certain minority groups. Despite an array of existing firm-side bias mitigation techniques, they typically incur utility costs and require organizational buy-in. Recognizing that many models rely on user-contributed data, end-users can induce fairness through the framework of Algorithmic Collective Action, where a coordinated minority group strategically relabels its own data to enhance fairness, without altering the firm's training process. We propose three practical, model-agnostic methods to approximate ideal relabeling and validate them on real-world datasets. Our findings show that a subgroup of the minority can substantially reduce unfairness with a small impact on the overall prediction error.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:09:39 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ben-Dov",
"Omri",
""
],
[
"Samadi",
"Samira",
""
],
[
"Sanyal",
"Amartya",
""
],
[
"Ţifrea",
"Alexandru",
""
]
] |
2508.15375
|
Hanwen Hu
|
Hanwen Hu, Jiancheng An, Lu Gan, and Chau Yuen
|
Performance Analysis of RIS-Aided High-Mobility Wireless Systems
|
This paper has been accepted for publication in 2025 IEEE 102nd Vehicular Technology Conference
| null | null | null |
eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Reconfigurable intelligent surface (RIS) technology holds immense potential for increasing the performance of wireless networks. Therefore, RIS is also regarded as one of the solutions to address communication challenges in high-mobility scenarios, such as Doppler shift and fast fading. This paper investigates a high-speed train (HST) multiple-input single-output (MISO) communication system aided by a RIS. We propose a block coordinate descent (BCD) algorithm to jointly optimize the RIS phase shifts and the transmit beamforming vectors to maximize the channel gain. Numerical results are provided to demonstrate that the proposed algorithm significantly enhances the system performance, achieving an average channel gain improvement of 15 dB compared to traditional schemes. Additionally, the introduction of RIS eliminates outage probability and improves key performance metrics such as achievable rate, channel capacity, and bit error rate (BER). These findings highlight the critical role of RIS in enhancing HST communication systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:09:47 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Hu",
"Hanwen",
""
],
[
"An",
"Jiancheng",
""
],
[
"Gan",
"Lu",
""
],
[
"Yuen",
"Chau",
""
]
] |
2508.15376
|
Cong Wang
|
Cong Wang, Xianda Guo, Wenbo Xu, Wei Tian, Ruiqi Song, Chenming Zhang, Lingxi Li, Long Chen
|
DriveSplat: Decoupled Driving Scene Reconstruction with Geometry-enhanced Partitioned Neural Gaussians
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
In the realm of driving scenarios, the presence of rapidly moving vehicles, pedestrians in motion, and large-scale static backgrounds poses significant challenges for 3D scene reconstruction. Recent methods based on 3D Gaussian Splatting address the motion blur problem by decoupling dynamic and static components within the scene. However, these decoupling strategies overlook background optimization with adequate geometry relationships and rely solely on fitting each training view by adding Gaussians. Therefore, these models exhibit limited robustness in rendering novel views and lack an accurate geometric representation. To address the above issues, we introduce DriveSplat, a high-quality reconstruction method for driving scenarios based on neural Gaussian representations with dynamic-static decoupling. To better accommodate the predominantly linear motion patterns of driving viewpoints, a region-wise voxel initialization scheme is employed, which partitions the scene into near, middle, and far regions to enhance close-range detail representation. Deformable neural Gaussians are introduced to model non-rigid dynamic actors, whose parameters are temporally adjusted by a learnable deformation network. The entire framework is further supervised by depth and normal priors from pre-trained models, improving the accuracy of geometric structures. Our method has been rigorously evaluated on the Waymo and KITTI datasets, demonstrating state-of-the-art performance in novel-view synthesis for driving scenarios.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:14:50 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Cong",
""
],
[
"Guo",
"Xianda",
""
],
[
"Xu",
"Wenbo",
""
],
[
"Tian",
"Wei",
""
],
[
"Song",
"Ruiqi",
""
],
[
"Zhang",
"Chenming",
""
],
[
"Li",
"Lingxi",
""
],
[
"Chen",
"Long",
""
]
] |
2508.15377
|
Tim Mauch
|
Tim Mauch, Thilo Gross
|
Diffusion-driven pattern formation in an opinion dynamical network model
|
10 pages, 8 figures
| null | null | null |
physics.soc-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The spatial organization of individuals and their interactions in communities are important factors known to preserve diversity in many complex systems. Inspired by metapopulation models from ecology, we study opinion formation using a network-based approach in which nodes represent communities of interacting agents holding one of two competing opinions, and links represent avenues of migration. Agents adapt to the dominant opinion within a community or migrate toward communities with similar views. Using a master stability function approach, we analytically derive conditions for diffusion-driven pattern formation and identify structural features of the community network that sustain opinion diversity. Our model shows that even under minimal opinion rules, the interaction between local dynamics and community structure generates spatial patterns that allow minority opinions to persist by gaining local dominance.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:15:32 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Mauch",
"Tim",
""
],
[
"Gross",
"Thilo",
""
]
] |
2508.15378
|
Haodi Zhong
|
Haodi Zhong, Liuxin Zou, Di Wang, Bo Wang, Zhenxing Niu, Quan Wang
|
EvoFormer: Learning Dynamic Graph-Level Representations with Structural and Temporal Bias Correction
| null |
ACM International Conference on Information and Knowledge Management 2025
|
10.1145/3746252.3761134
| null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Dynamic graph-level embedding aims to capture structural evolution in networks, which is essential for modeling real-world scenarios. However, existing methods face two critical yet under-explored issues: Structural Visit Bias, where random walk sampling disproportionately emphasizes high-degree nodes, leading to redundant and noisy structural representations; and Abrupt Evolution Blindness, the failure to effectively detect sudden structural changes due to rigid or overly simplistic temporal modeling strategies, resulting in inconsistent temporal embeddings. To overcome these challenges, we propose EvoFormer, an evolution-aware Transformer framework tailored for dynamic graph-level representation learning. To mitigate Structural Visit Bias, EvoFormer introduces a Structure-Aware Transformer Module that incorporates positional encoding based on node structural roles, allowing the model to globally differentiate and accurately represent node structures. To overcome Abrupt Evolution Blindness, EvoFormer employs an Evolution-Sensitive Temporal Module, which explicitly models temporal evolution through a sequential three-step strategy: (I) Random Walk Timestamp Classification, generating initial timestamp-aware graph-level embeddings; (II) Graph-Level Temporal Segmentation, partitioning the graph stream into segments reflecting structurally coherent periods; and (III) Segment-Aware Temporal Self-Attention combined with an Edge Evolution Prediction task, enabling the model to precisely capture segment boundaries and perceive structural evolution trends, effectively adapting to rapid temporal shifts. Extensive evaluations on five benchmark datasets confirm that EvoFormer achieves state-of-the-art performance in graph similarity ranking, temporal anomaly detection, and temporal segmentation tasks, validating its effectiveness in correcting structural and temporal biases.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:19:54 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhong",
"Haodi",
""
],
[
"Zou",
"Liuxin",
""
],
[
"Wang",
"Di",
""
],
[
"Wang",
"Bo",
""
],
[
"Niu",
"Zhenxing",
""
],
[
"Wang",
"Quan",
""
]
] |
2508.15379
|
Mingduo Xie
|
Jinliang Yu, Mingduo Xie, Yue Wang, Tianfan Fu, Xianglai Xu, Jiajun Wang
|
Bladder Cancer Diagnosis with Deep Learning: A Multi-Task Framework and Online Platform
| null | null | null | null |
eess.IV cs.AI cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Clinical cystoscopy, the current standard for bladder cancer diagnosis, suffers from significant reliance on physician expertise, leading to variability and subjectivity in diagnostic outcomes. There is an urgent need for objective, accurate, and efficient computational approaches to improve bladder cancer diagnostics.
Leveraging recent advancements in deep learning, this study proposes an integrated multi-task deep learning framework specifically designed for bladder cancer diagnosis from cystoscopic images. Our framework includes a robust classification model using EfficientNet-B0 enhanced with Convolutional Block Attention Module (CBAM), an advanced segmentation model based on ResNet34-UNet++ architecture with self-attention mechanisms and attention gating, and molecular subtyping using ConvNeXt-Tiny to classify molecular markers such as HER-2 and Ki-67. Additionally, we introduce a Gradio-based online diagnostic platform integrating all developed models, providing intuitive features including multi-format image uploads, bilingual interfaces, and dynamic threshold adjustments.
Extensive experimentation demonstrates the effectiveness of our methods, achieving outstanding accuracy (93.28%), F1-score (82.05%), and AUC (96.41%) for classification tasks, and exceptional segmentation performance indicated by a Dice coefficient of 0.9091. The online platform significantly improved the accuracy, efficiency, and accessibility of clinical bladder cancer diagnostics, enabling practical and user-friendly deployment. The code is publicly available.
Our multi-task framework and integrated online tool collectively advance the field of intelligent bladder cancer diagnosis by improving clinical reliability, supporting early tumor detection, and enabling real-time diagnostic feedback. These contributions mark a significant step toward AI-assisted decision-making in urology.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:20:03 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yu",
"Jinliang",
""
],
[
"Xie",
"Mingduo",
""
],
[
"Wang",
"Yue",
""
],
[
"Fu",
"Tianfan",
""
],
[
"Xu",
"Xianglai",
""
],
[
"Wang",
"Jiajun",
""
]
] |
2508.15380
|
Vishwa Prakash Hv
|
Vishwa Prakash HV, Ruta Mehta, Prajakta Nimbhorkar
|
Almost and Approximate EFX for Few Types of Agents
| null | null | null | null |
cs.GT cs.DS
|
http://creativecommons.org/licenses/by-sa/4.0/
|
We study the problem of fair allocation of a set of indivisible goods among $n$ agents with $k$ distinct additive valuations, with the goal of achieving approximate envy-freeness up to any good ($\alpha-\mathrm{EFX}$).
It is known that EFX allocations exist for $n$ agents when there are at most three distinct valuations due to HV et al. Furthermore, Amanatidis et al. showed that a $\frac{2}{3}-\mathrm{EFX}$ allocation is guaranteed to exist when number of agents is at most seven. In this paper, we show that a $\frac{2}{3}-\mathrm{EFX}$ allocation exists for any number of agents when there are at most four distinct valuations.
Secondly, we consider a relaxation called $\mathrm{EFX}$ with charity, where some goods remain unallocated such that no agent envies the set of unallocated goods. Akrami et al. showed that for $n$ agents and any $\varepsilon \in \left(0, \frac{1}{2}\right]$, there exists a $(1-\varepsilon)-\mathrm{EFX}$ allocation with at most $\tilde{\mathcal{O}}((n/\varepsilon)^{\frac{1}{2}})$ goods to charity. In this paper, we show that a $(1-\varepsilon)-\mathrm{EFX}$ allocation with a $\tilde{\mathcal{O}}(k/\varepsilon)^{\frac{1}{2}}$ charity exists for any number of agents when there are at most $k$ distinct valuations.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:20:06 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"HV",
"Vishwa Prakash",
""
],
[
"Mehta",
"Ruta",
""
],
[
"Nimbhorkar",
"Prajakta",
""
]
] |
2508.15381
|
Zhi Zhou
|
Siyu Cen, Bangti Jin, Qimeng Quan, Zhi Zhou
|
Numerical Analysis of Unsupervised Learning Approaches for Parameter Identification in PDEs
|
49 pages
| null | null | null |
math.NA cs.NA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Identifying parameters in partial differential equations (PDEs) represents a very broad class of applied inverse problems. In recent years, several unsupervised learning approaches using (deep) neural networks have been developed to solve PDE parameter identifications. These approaches employ neural networks as ansatz functions to approximate the parameters and / or the states, and have demonstrated impressive empirical performance. In this paper, we provide a comprehensive survey on these unsupervised learning techniques on one model problem, diffusion coefficient identification, from the classical numerical analysis perspective, and outline a general framework for deriving rigorous error bounds on the discrete approximations obtained using the Galerkin finite element method, hybrid method and deep neural networks. Throughout we highlight the crucial role of conditional stability estimates in the error analysis.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:20:15 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Cen",
"Siyu",
""
],
[
"Jin",
"Bangti",
""
],
[
"Quan",
"Qimeng",
""
],
[
"Zhou",
"Zhi",
""
]
] |
2508.15382
|
Florian Bouville
|
Ruxue Yang, Temesgen Tadeyos Zate, Soumyajit Mojumder, Oriol Gavalda-Diaz, Zihe Li, Ajeet Kumar, James Roscow, Hamideh Khanbareh, Astri Bj{\o}rnetun Haugen, Florian Bouville
|
Strong lead-free bioinspired piezoceramics for durable energy transducers
| null | null | null | null |
cond-mat.mtrl-sci physics.app-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Durable, high-performance and eco-friendly lead-free piezoceramics are essential for next-generation sustainable energy transducers and electromechanical systems. While significant performance enhancements have been made, through chemical composition, texture, or crystal defects, piezoceramics are intrinsically weak mechanically, which negatively impact their working conditions and durability. What's more, improving comprehensive mechanical durability without sacrificing piezoelectric performance remains a key challenge. Here, we design bioinspired Bi0.5Na0.5TiO3 (BNT) ceramics using a scalable colloidal process that enables multiscale control over the microstructure. The design comprises plate-like monocrystalline BNT bricks stacked to induce a crystallographic texture along the poling direction, bonded together by a silica-based mortar, forming the brick-and-mortar phase. This deliberate microstructure design yields 2- to 3-fold increase in flexural strength, and 1.6- to 2-fold increase in fracture toughness compared with a BNT synthesized conventionally, comparable to common structural ceramics, without sacrificing the piezoelectric performance. In addition, the bioinspired BNT exhibit dramatically enhanced ferroelectric fatigue resistance, with a 40- to 600-folds improvement in the number of field-induced electromechanical cycles before failure. These gains stem from residual stress fields generated at the interface between silica pockets and BNT bricks, which delay crack initiation. Furthermore, we demonstrated enhanced transducing capability and electromechanical fatigue resistance using a cantilever beam-based piezoelectric transducer under bending mode. Given its non-chemical-compositional origin, this bioinspired strategy could be broadly applicable to other piezoelectric material systems for applications where both functional and structural performance are critical.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:20:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Yang",
"Ruxue",
""
],
[
"Zate",
"Temesgen Tadeyos",
""
],
[
"Mojumder",
"Soumyajit",
""
],
[
"Gavalda-Diaz",
"Oriol",
""
],
[
"Li",
"Zihe",
""
],
[
"Kumar",
"Ajeet",
""
],
[
"Roscow",
"James",
""
],
[
"Khanbareh",
"Hamideh",
""
],
[
"Haugen",
"Astri Bjørnetun",
""
],
[
"Bouville",
"Florian",
""
]
] |
2508.15383
|
Ernest Y.-Z. Tan
|
Ernest Y.-Z. Tan and Shlok Nahar
|
Incorporating device characterization into security proofs
|
16 pages + appendices
| null | null | null |
quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Typical security proofs for quantum key distribution (QKD) rely on having some model for the devices, with the security guarantees implicitly relying on the values of various parameters of the model, such as dark count rates or detector efficiencies. Hence to deploy QKD in practice, we must establish how to certify or characterize the model parameters of a manufacturer's QKD devices. We present a rigorous framework for analyzing such procedures, laying out concrete requirements for both the security proofs and the certification or characterization procedures. In doing so, we describe various forms of conclusions that can and cannot be validly drawn from such procedures, addressing some potential misconceptions. We also discuss connections to composable security frameworks and some technical aspects that remain to be resolved in that direction.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:21:12 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Tan",
"Ernest Y. -Z.",
""
],
[
"Nahar",
"Shlok",
""
]
] |
2508.15384
|
Jaewon Lee
|
Jaewon Lee and O\u{g}uz \c{S}avk
|
On homology spheres of the trivial local equivalence class
|
12 pages, 4 figures. Comments welcome!
| null | null | null |
math.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In the homology cobordism group $\Theta_{\mathbb{Z}}^3$, it is not known if there is a non-trivial linear dependence between Seifert fibered spheres. Based on involutive Heegaard Floer theory, Hendricks, Manolescu, and Zemke introduced the local equivalence group $\mathfrak{I}$ with the homomorphism $h:\Theta_{\mathbb{Z}}^3 \rightarrow \mathfrak{I}$. And it is possible to find a non-trivial linear dependence between the image of Seifert fibered spheres in $\mathfrak{I}$ using the work of Dai and Stoffregen. Therefore, it is interesting to ask if such a linear dependence in $\mathfrak{I}$ originates from $\Theta_{\mathbb{Z}}^3$. In this paper, we prove that this is not the case in general. We show that infinitely many homology spheres linearly dependent in $\mathfrak{I}$ are linearly independent even in the rational homology cobordism group $\Theta_{\mathbb{Q}}^3$, by employing the $r_s$-invariants from the filtered instanton Floer homology developed by Nozaki, Sato, and Taniguchi. Moreover, these homology spheres include homology spheres of the trivial local equivalence class of $\operatorname{Pin(2)}$-equivariant Seiberg--Witten Floer stable homotopy type.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:21:54 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Lee",
"Jaewon",
""
],
[
"Şavk",
"Oğuz",
""
]
] |
2508.15385
|
You-Ting Huang
|
Ponder Liu, Hao-Cheng Hung, You-Ting Huang, Jia-Cheng Li, Carmine Ortix, Ching-Hao Chang
|
Bending nanoribbon to induce large anisotropic magnetoconductance
| null | null | null | null |
cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
When a nanoribbon is bent under a homogeneous external magnetic field, the effective magnetic field inside becomes either homogeneous or inhomogeneous, depending on the direction of the field. This enables the selective creation of bulk, interface, and edge magnetic states in the bent structure, for a magnetic field with a strength. We establish theoretically that these tuneable states lead to a strong geometry-induced anisotropic magnetoconductance (GAMC) in perpendicularly bent nanoribbon, which can reach up to 100\%. Moreover, the GAMC can be further enhanced to 200\%, 300\%, or even higher by either further bending or tuning the bending angle. The potential of this phenomenon for practical applications is demonstrated by its stable anisotropy, which remains consistent across a wide range of Fermi energies, can be observed even at weak magnetic fields and room temperature, and occurs in various systems such as two-dimensional electron gas (2DEG) and graphene.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:22:40 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Ponder",
""
],
[
"Hung",
"Hao-Cheng",
""
],
[
"Huang",
"You-Ting",
""
],
[
"Li",
"Jia-Cheng",
""
],
[
"Ortix",
"Carmine",
""
],
[
"Chang",
"Ching-Hao",
""
]
] |
2508.15386
|
Sabine Houy
|
Sabine Houy, Bruno Kreyssig, Timothee Riom, Alexandre Bartel, Patrick McDaniel
|
A Practical Guideline and Taxonomy to LLVM's Control Flow Integrity
| null | null | null | null |
cs.CR cs.SE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Memory corruption vulnerabilities remain one of the most severe threats to software security. They often allow attackers to achieve arbitrary code execution by redirecting a vulnerable program's control flow. While Control Flow Integrity (CFI) has gained traction to mitigate this exploitation path, developers are not provided with any direction on how to apply CFI to real-world software. In this work, we establish a taxonomy mapping LLVM's forward-edge CFI variants to memory corruption vulnerability classes, offering actionable guidance for developers seeking to deploy CFI incrementally in existing codebases. Based on the Top 10 Known Exploited Vulnerabilities (KEV) list, we identify four high-impact vulnerability categories and select one representative CVE for each. We evaluate LLVM's CFI against each CVE and explain why CFI blocks exploitation in two cases while failing in the other two, illustrating its potential and current limitations. Our findings support informed deployment decisions and provide a foundation for improving the practical use of CFI in production systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:23:24 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Houy",
"Sabine",
""
],
[
"Kreyssig",
"Bruno",
""
],
[
"Riom",
"Timothee",
""
],
[
"Bartel",
"Alexandre",
""
],
[
"McDaniel",
"Patrick",
""
]
] |
2508.15387
|
Beiming Yuan
|
Ruizhuo Song, Beiming Yuan
|
DIO: Refining Mutual Information and Causal Chain to Enhance Machine Abstract Reasoning Ability
|
15 pages, 9 figures, 8 tables
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Despite the outstanding performance of current deep learning models across various domains, their fundamental bottleneck in abstract reasoning remains unresolved. To address this challenge, the academic community has introduced Raven's Progressive Matrices (RPM) problems as an authoritative benchmark for evaluating the abstract reasoning capabilities of deep learning algorithms, with a focus on core intelligence dimensions such as abstract reasoning, pattern recognition, and complex problem-solving. Therefore, this paper centers on solving RPM problems, aiming to contribute to enhancing the abstract reasoning abilities of machine intelligence. Firstly, this paper adopts a ``causal chain modeling'' perspective to systematically analyze the complete causal chain in RPM tasks: image $\rightarrow$ abstract attributes $\rightarrow$ progressive attribute patterns $\rightarrow$ pattern consistency $\rightarrow$ correct answer. Based on this analysis, the network architecture of the baseline model DIO is designed. However, experiments reveal that the optimization objective formulated for DIO, namely maximizing the variational lower bound of mutual information between the context and the correct option, fails to enable the model to genuinely acquire the predefined human reasoning logic. This is attributed to two main reasons: the tightness of the lower bound significantly impacts the effectiveness of mutual information maximization, and mutual information, as a statistical measure, does not capture the causal relationship between subjects and objects. To overcome these limitations, this paper progressively proposes three improvement methods:
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:23:51 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Song",
"Ruizhuo",
""
],
[
"Yuan",
"Beiming",
""
]
] |
2508.15388
|
Yu Xia
|
Yu Xia, Rui Zhong, Zeyu Song, Wei Yang, Junchen Wan, Qingpeng Cai, Chi Lu, Peng Jiang
|
TrackRec: Iterative Alternating Feedback with Chain-of-Thought via Preference Alignment for Recommendation
| null | null | null | null |
cs.IR
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The extensive world knowledge and powerful reasoning capabilities of large language models (LLMs) have attracted significant attention in recommendation systems (RS). Specifically, The chain of thought (CoT) has been shown to improve the performance of LLMs on complex reasoning tasks for RS. However, due to the fact that LLMs often suffer from hallucination issues, there is no guarantee that their reasoning CoT is effective. A key challenge is to further enhance the recommendation capabilities of LLMs through effective CoT reasonings. Therefore, we propose \textbf{TrackRec}, a framework designed to enhance reasoning capabilities of LLMs for RS. TrackRec specifically focuses on accurately inferring recommendation CoT \textbf{(RecCoT)} for user preference using the knowledge from LLMs. This RecCoT can serve both as an explanation for the LLM's completion of recommendation tasks and as auxiliary features to assist recommendation models in accomplishing recommendation tasks. TrackRec consists of a RecCoT generator $(G)$ and a RecCoT validator $(V)$. Furthermore, we design alternating feedback learning mechanism that $G$ undergoes direct preference optimization via feedback from $V$ to produce increasingly accurate RecCoT aligned with $V$'s standards. Meanwhile, $V$ is fine-tuned using the inference feedback from $G$ to enhance its validation capabilities in alignment with recommendation tasks. Through iterative alternating feedback learning between $G$ and $V$, TrackRec continuously improves the user preference analysis capability of $G$ and the validation capacity of $V$. Extensive experiments demonstrate the effectiveness of our approach, showing that it surpasses state-of-the-art methods. Moreover, TrackRec has been deployed on a lagre advertising platform with hundreds of millions of users, achieving substantial gains.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:25:31 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xia",
"Yu",
""
],
[
"Zhong",
"Rui",
""
],
[
"Song",
"Zeyu",
""
],
[
"Yang",
"Wei",
""
],
[
"Wan",
"Junchen",
""
],
[
"Cai",
"Qingpeng",
""
],
[
"Lu",
"Chi",
""
],
[
"Jiang",
"Peng",
""
]
] |
2508.15389
|
Wenrui Li
|
Wenrui Li, Wei Han, Liang-Jian Deng, Ruiqin Xiong, Xiaopeng Fan
|
Spiking Variational Graph Representation Inference for Video Summarization
|
Accepted by IEEE TIP
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
With the rise of short video content, efficient video summarization techniques for extracting key information have become crucial. However, existing methods struggle to capture the global temporal dependencies and maintain the semantic coherence of video content. Additionally, these methods are also influenced by noise during multi-channel feature fusion. We propose a Spiking Variational Graph (SpiVG) Network, which enhances information density and reduces computational complexity. First, we design a keyframe extractor based on Spiking Neural Networks (SNN), leveraging the event-driven computation mechanism of SNNs to learn keyframe features autonomously. To enable fine-grained and adaptable reasoning across video frames, we introduce a Dynamic Aggregation Graph Reasoner, which decouples contextual object consistency from semantic perspective coherence. We present a Variational Inference Reconstruction Module to address uncertainty and noise arising during multi-channel feature fusion. In this module, we employ Evidence Lower Bound Optimization (ELBO) to capture the latent structure of multi-channel feature distributions, using posterior distribution regularization to reduce overfitting. Experimental results show that SpiVG surpasses existing methods across multiple datasets such as SumMe, TVSum, VideoXum, and QFVS. Our codes and pre-trained models are available at https://github.com/liwrui/SpiVG.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:25:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Li",
"Wenrui",
""
],
[
"Han",
"Wei",
""
],
[
"Deng",
"Liang-Jian",
""
],
[
"Xiong",
"Ruiqin",
""
],
[
"Fan",
"Xiaopeng",
""
]
] |
2508.15390
|
Woojin Chung
|
Woojin Chung, Jeonghoon Kim
|
Exploiting Vocabulary Frequency Imbalance in Language Model Pre-training
|
Preprint
| null | null | null |
cs.CL cs.LG
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Large language models are trained with tokenizers, and the resulting token distribution is highly imbalanced: a few words dominate the stream while most occur rarely. Recent practice favors ever-larger vocabularies, but the source of the benefit is unclear. We conduct a controlled study that scales the language model's vocabulary from 24K to 196K while holding data, compute, and optimization fixed. We first quantify the complexity of tokenized text, formalized via Kolmogorov complexity, and show that larger vocabularies reduce this complexity. Above 24K, every common word is already a single token, so further growth mainly deepens the relative token-frequency imbalance. A word-level loss decomposition shows that larger vocabularies reduce cross-entropy almost exclusively by lowering uncertainty on the 2,500 most frequent words, even though loss on the rare tail rises. Constraining input and output embedding norms to attenuate the effect of token-frequency imbalance reverses the gain, directly showing that the model exploits rather than suffers from imbalance. Because the same frequent words cover roughly 77% of tokens in downstream benchmarks, this training advantage transfers intact. We also show that enlarging model parameters with a fixed vocabulary yields the same frequent-word benefit. Our results reframe "bigger vocabularies help" as "lowering the complexity of tokenized text helps," providing a simple, principled lever for tokenizer-model co-design and clarifying the loss dynamics that govern language-model scaling in pre-training.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:26:48 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Chung",
"Woojin",
""
],
[
"Kim",
"Jeonghoon",
""
]
] |
2508.15391
|
Benjamin Kraner
|
Benjamin Kraner, Luca Pennella, Nicol\`o Vallarano and Claudio J. Tessone
|
Money in Motion: Micro-Velocity and Usage of Ethereums Liquid Staking Tokens
| null | null | null | null |
cs.ET
|
http://creativecommons.org/licenses/by/4.0/
|
We introduce a micro-velocity framework for analysing the on-chain circulation of Lidos liquid-staking tokens, stETH, and its wrapped ERC-20 form, wstETH. By reconstructing full transfer and share-based accounting histories, we compute address-level velocities and decompose them into behavioural components. Despite their growing importance, the micro-level monetary dynamics of LSTs remain largely unexplored. Our data reveal persistently high velocity for both tokens, reflecting intensive reuse within DeFi. Yet activity is highly concentrated: a small cohort of large addresses, likely institutional accounts, are responsible for most turnover, while the rest of the users remain largely passive. We also observe a gradual transition in user behavior, characterized by a shift toward wstETH, the non-rebasing variant of stETH. This shift appears to align with DeFi composability trends, as wstETH is more frequently deployed across protocols such as AAVE, Spark, Balancer, and SkyMoney.
To make the study fully reproducible, we release (i) an open-source pipeline that indexes event logs and historical contract state, and (ii) two public datasets containing every Transfer and TransferShares record for stETH and wstETH through 2024-11-08. This is the first large-scale empirical characterisation of liquid-staking token circulation. Our approach offers a scalable template for monitoring staking asset flows and provides new, open-access resources to the research community.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:27:57 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kraner",
"Benjamin",
""
],
[
"Pennella",
"Luca",
""
],
[
"Vallarano",
"Nicolò",
""
],
[
"Tessone",
"Claudio J.",
""
]
] |
2508.15392
|
Chenghao Zhang
|
Chenghao Zhang, Qingqing Long, Ludi Wang, Wenjuan Cui, Jianjun Yu, Yi Du
|
CITE: A Comprehensive Benchmark for Heterogeneous Text-Attributed Graphs on Catalytic Materials
|
23 pages, 4 figures,
| null | null | null |
cs.LG cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Text-attributed graphs(TAGs) are pervasive in real-world systems,where each node carries its own textual features. In many cases these graphs are inherently heterogeneous, containing multiple node types and diverse edge types. Despite the ubiquity of such heterogeneous TAGs, there remains a lack of large-scale benchmark datasets. This shortage has become a critical bottleneck, hindering the development and fair comparison of representation learning methods on heterogeneous text-attributed graphs. In this paper, we introduce CITE - Catalytic Information Textual Entities Graph, the first and largest heterogeneous text-attributed citation graph benchmark for catalytic materials. CITE comprises over 438K nodes and 1.2M edges, spanning four relation types. In addition, we establish standardized evaluation procedures and conduct extensive benchmarking on the node classification task, as well as ablation experiments on the heterogeneous and textual properties of CITE. We compare four classes of learning paradigms, including homogeneous graph models, heterogeneous graph models, LLM(Large Language Model)-centric models, and LLM+Graph models. In a nutshell, we provide (i) an overview of the CITE dataset, (ii) standardized evaluation protocols, and (iii) baseline and ablation experiments across diverse modeling paradigms.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:28:19 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Zhang",
"Chenghao",
""
],
[
"Long",
"Qingqing",
""
],
[
"Wang",
"Ludi",
""
],
[
"Cui",
"Wenjuan",
""
],
[
"Yu",
"Jianjun",
""
],
[
"Du",
"Yi",
""
]
] |
2508.15393
|
Miha O\v{z}bot
|
Miha O\v{z}bot, Igor \v{S}krjanc
|
Federated Learning based on Self-Evolving Gaussian Clustering
|
5 pages, in slovenian language, 3 figures. Published in the Proceedings of the 33rd International Electrotechnical and Computer Science Conference (ERK 2024), Portoroz, Slovenia, pp. 240-243. Indexed in COBISS (COBISS.SI-ID 212879107). Official version available at https://erk.fe.uni-lj.si/2024/papers/ozbot_federativno_ucenje.pdf
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
In this study, we present an Evolving Fuzzy System within the context of Federated Learning, which adapts dynamically with the addition of new clusters and therefore does not require the number of clusters to be selected apriori. Unlike traditional methods, Federated Learning allows models to be trained locally on clients' devices, sharing only the model parameters with a central server instead of the data. Our method, implemented using PyTorch, was tested on clustering and classification tasks. The results show that our approach outperforms established classification methods on several well-known UCI datasets. While computationally intensive due to overlap condition calculations, the proposed method demonstrates significant advantages in decentralized data processing.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:32:37 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Ožbot",
"Miha",
""
],
[
"Škrjanc",
"Igor",
""
]
] |
2508.15394
|
Jun Choi
|
Jun Choi, Chang-Ock Lee, Minam Moon
|
Hybrid Least Squares/Gradient Descent Methods for DeepONets
| null | null | null | null |
cs.LG cs.AI cs.NA math.NA
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
We propose an efficient hybrid least squares/gradient descent method to accelerate DeepONet training. Since the output of DeepONet can be viewed as linear with respect to the last layer parameters of the branch network, these parameters can be optimized using a least squares (LS) solve, and the remaining hidden layer parameters are updated by means of gradient descent form. However, building the LS system for all possible combinations of branch and trunk inputs yields a prohibitively large linear problem that is infeasible to solve directly. To address this issue, our method decomposes the large LS system into two smaller, more manageable subproblems $\unicode{x2014}$ one for the branch network and one for the trunk network $\unicode{x2014}$ and solves them separately. This method is generalized to a broader type of $L^2$ loss with a regularization term for the last layer parameters, including the case of unsupervised learning with physics-informed loss.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:34:06 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Choi",
"Jun",
""
],
[
"Lee",
"Chang-Ock",
""
],
[
"Moon",
"Minam",
""
]
] |
2508.15395
|
Can Chen
|
Yingqi Liu and Tianlu Pan and Jingjun Tan and Renxin Zhong and Can Chen
|
Integrated Take-off Management and Trajectory Optimization for Merging Control in Urban Air Mobility Corridors
|
31 pages
| null | null | null |
math.OC cs.SY eess.SY
|
http://creativecommons.org/licenses/by/4.0/
|
Urban Air Mobility (UAM) has the potential to revolutionize daily transportation, offering rapid and efficient aerial mobility services. Take-off and merging phases are critical for air corridor operations, requiring the coordination of take-off aircraft and corridor traffic while ensuring safety and seamless transition. This paper proposes an integrated take-off management and trajectory optimization for merging control in UAM corridors. We first introduce a novel take-off airspace design. To our knowledge, this paper is one of the first to propose a structured design for take-off airspace. Based on the take-off airspace design, we devise a hierarchical coordinated take-off and merging management (HCTMM) strategy. To be specific, the take-off airspace design can simplify aircraft dynamics and thus reduce the dimensionality of the trajectory optimization problem whilst mitigating obstacle avoidance complexities. The HCTMM strategy strictly ensures safety and improves the efficiency of take-off and merging operations. At the tactical level, a scheduling algorithm coordinates aircraft take-off times and selects dynamic merging points to reduce conflicts and ensure smooth take-off and merging processes. At the operational level, a trajectory optimization strategy ensures that each aircraft reaches the dynamic merging point efficiently while satisfying safety constraints. Simulation results show that, compared to representative strategies with fixed or dynamic merging points, the HCTMM strategy significantly improves operational efficiency and reduces computational burden, while ensuring safety under various corridor traffic conditions. Further results confirm the scalability of the HCTMM strategy and the computational efficiency enabled by the proposed take-off airspace design.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:36:13 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Liu",
"Yingqi",
""
],
[
"Pan",
"Tianlu",
""
],
[
"Tan",
"Jingjun",
""
],
[
"Zhong",
"Renxin",
""
],
[
"Chen",
"Can",
""
]
] |
2508.15396
|
Tobias Schreieder
|
Tobias Schreieder and Tim Schopf and Michael F\"arber
|
Attribution, Citation, and Quotation: A Survey of Evidence-based Text Generation with Large Language Models
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
The increasing adoption of large language models (LLMs) has been accompanied by growing concerns regarding their reliability and trustworthiness. As a result, a growing body of research focuses on evidence-based text generation with LLMs, aiming to link model outputs to supporting evidence to ensure traceability and verifiability. However, the field is fragmented due to inconsistent terminology, isolated evaluation practices, and a lack of unified benchmarks. To bridge this gap, we systematically analyze 134 papers, introduce a unified taxonomy of evidence-based text generation with LLMs, and investigate 300 evaluation metrics across seven key dimensions. Thereby, we focus on approaches that use citations, attribution, or quotations for evidence-based text generation. Building on this, we examine the distinctive characteristics and representative methods in the field. Finally, we highlight open challenges and outline promising directions for future work.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:36:35 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Schreieder",
"Tobias",
""
],
[
"Schopf",
"Tim",
""
],
[
"Färber",
"Michael",
""
]
] |
2508.15397
|
Michal Natorski
|
Michal Natorski
|
Multilateralism in the Global Governance of Artificial Intelligence
|
32 pages, 2 tables
| null | null | null |
cs.CY
|
http://creativecommons.org/licenses/by/4.0/
|
This chapter inquires how international multilateralism addresses the emergence of the general-purpose technology of Artificial Intelligence. In more detail, it analyses two key features of AI multilateralism: its generalized principles and the coordination of state relations in the realm of AI. Firstly, it distinguishes the generalized principles of AI multilateralism of epochal change, determinism, and dialectical understanding. In the second place, the adaptation of multilateralism to AI led to the integration of AI issues into the agendas of existing cooperation frameworks and the creation of new ad hoc frameworks focusing exclusively on AI issues. In both cases, AI multilateralism develops in the shadow of the state hierarchy in relations with other AI stakeholders. While AI multilateralism is multi-stakeholder, and the hierarchy between state and non-state actors may seem blurred, states preserve the competence as decisive decision-makers in agenda-setting, negotiation, and implementation of soft law international commitments.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:39:04 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Natorski",
"Michal",
""
]
] |
2508.15398
|
Takahiro Matsumoto
|
Takahiro Matsumoto, Masafumi Suzuki, Mariko Yamaguchi, Masakatsu Aoki, Shunsuke Konagai, and Kazuhiko Murasaki
|
A Low-Latency 3D Live Remote Visualization System for Tourist Sites Integrating Dynamic and Pre-captured Static Point Clouds
|
3 pages, 4 figures, submitted to IEEE ISMAR 2025 Posters
| null | null | null |
cs.MM
|
http://creativecommons.org/licenses/by/4.0/
|
Various real-time methods for capturing and transmitting dynamic 3D spaces have been proposed, including those based on RGB-D cameras and volumetric capture. However, applying existing methods to outdoor tourist sites remains difficult because maintenance and aesthetic constraints limit sensor placement, and daylight variability complicates processing. We propose a system that combines multiple LiDARs and cameras for live dynamic point cloud capture, and integrates them with pre-captured static point clouds for wide-area 3D visualization. The system sustains 30 fps across wide-area scenes while keeping latency below 100 ms. To mitigate lighting inconsistencies, static point-cloud colors are automatically adjusted to current lighting. The effectiveness of our system is demonstrated through real-world deployment in a tourist site.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:40:59 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Matsumoto",
"Takahiro",
""
],
[
"Suzuki",
"Masafumi",
""
],
[
"Yamaguchi",
"Mariko",
""
],
[
"Aoki",
"Masakatsu",
""
],
[
"Konagai",
"Shunsuke",
""
],
[
"Murasaki",
"Kazuhiko",
""
]
] |
2508.15399
|
Ruslan Viacheslavovich Skuratovskii
|
R.V. Skuratovskii
|
Involutive and minimal generating sets of Extended Special Linear group $ES{L_3}(\mathbb{Z})$, formulas of roots in GL$_2$($\mathbb{F}_p$), GL$_2(\mathbb{Z})$ and SL$_3(\mathbb{Z})$
|
This work was reported on 5 algebraic conferences. 05
| null | null | null |
math.GR
|
http://creativecommons.org/publicdomain/zero/1.0/
|
In this research we continue our previous investigation of wreath product normal structure \cite{SkuESL}. We generalize the group of unimodular matrices \cite{Amit} and find its structure. For this goal we propose one extension of the special linear group. Groups generated by three involutions, two of which are permutable, have long been of interest in the theory of matrix groups \cite{Maz}, for instance such generating set was researched for $S{{L}_{2}}({{\mathbb{Z}+ i\mathbb{Z}}})$. But for size of matrix 3 on 3 this is imposable for some groups. We research this question for $ES{{L}_{3}}({{\mathbb{Z}}})$. An analytical formula of root in $SL(3, \mathbb{Z}$) is found, recursive formula for $n$-th power root in $SL(2, \mathbb{Z}$) is found too.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:41:40 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Skuratovskii",
"R. V.",
""
]
] |
2508.15400
|
Andrea Merlo
|
Giacomo Del Nin and Andrea Merlo
|
Marstrand's density theorem for arbitrary norms in the plane
| null | null | null | null |
math.CA math.MG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We show that if a non-trivial measure in the plane admits, at almost every point, positive and finite $\alpha$-dimensional density with respect to some norm, then $\alpha$ must be an integer.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:41:57 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Del Nin",
"Giacomo",
""
],
[
"Merlo",
"Andrea",
""
]
] |
2508.15401
|
Yixuan Feng
|
Yixuan Feng, Xavier R. Advincula, Hongwei Fang, Christoph Schran
|
Clay Edges Are Dynamic Proton-conducting Networks Modulated by Structure and pH
| null | null | null | null |
physics.chem-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Montmorillonite, a ubiquitous clay mineral, plays a vital role in geochemical and environmental processes due to its chemically complex edge surfaces. However, the molecular-scale acid-base reactivity of these interfaces remains poorly understood due to the limitations of both experimental resolution and conventional simulations. Here, we employ machine learning potentials with first-principles accuracy to perform nanosecond-scale molecular dynamics simulations of montmorillonite nanoparticles across a range of pH. Our results reveal clear amphoteric behavior: edge sites undergo protonation in acidic environments and deprotonation in basic conditions. Even at neutral pH, spontaneous and directional proton transfer events are common, proceeding via both direct and solvent-mediated pathways. These findings demonstrate that montmorillonite edges are not static arrays of hydroxyl groups but dynamic, proton-conducting networks whose reactivity is modulated by local structure and solution conditions. This work offers a molecular-level framework for understanding proton transport and buffering in clay-water systems, with broad implications for catalysis, ion exchange, and environmental remediation.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:43:26 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Feng",
"Yixuan",
""
],
[
"Advincula",
"Xavier R.",
""
],
[
"Fang",
"Hongwei",
""
],
[
"Schran",
"Christoph",
""
]
] |
2508.15402
|
Antonio Casares
|
Antonio Casares
|
Transition-based vs stated-based acceptance for automata over infinite words
| null | null | null | null |
cs.FL cs.LO
|
http://creativecommons.org/licenses/by/4.0/
|
Automata over infinite objects are a well-established model with applications in logic and formal verification. Traditionally, acceptance in such automata is defined based on the set of states visited infinitely often during a run. However, there is a growing trend towards defining acceptance based on transitions rather than states.
In this survey, we analyse the reasons for this shift and advocate using transition-based acceptance in the context of automata over infinite words. We present a collection of problems where the choice of formalism has a major impact and discuss the causes of these differences.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:44:33 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Casares",
"Antonio",
""
]
] |
2508.15403
|
Yu-Xin Xiao
|
Yu-Xin Xiao, Qing-Fei Han, He-Xia Zhang, Hanzhong Zhang
|
System-size dependence of $\gamma$-jet modifications in heavy-ion collisions
| null | null |
10.1088/1674-1137/adef24
| null |
nucl-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Medium modifications of the $\gamma$-triggered jets are investigated with the Linear Boltzmann Transport (LBT) model in heavy-ion collisions with varying system sizes, focusing on centrality dependence in Pb+Pb and Xe+Xe collisions at the LHC. Our numerical results reveal that jets produced in central collisions exhibit a wider transverse asymmetry ($A_N^y$) distribution, a broader jet shape, and a more pronounced $\gamma$-jet transverse momentum imbalance ($X_{J\gamma}=p_T^{\rm jet}/p_T^\gamma$) compared to peripheral collisions. These effects arise from the longer path length and stronger jet-medium interactions in central collisions, leading to enhanced jet quenching and medium response. Our findings demonstrate that the magnitude of $\gamma$-jet modifications is sensitive to the size and centrality of the collision system, with larger systems inducing more significant alterations due to increased energy loss and medium feedback.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:48:42 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Xiao",
"Yu-Xin",
""
],
[
"Han",
"Qing-Fei",
""
],
[
"Zhang",
"He-Xia",
""
],
[
"Zhang",
"Hanzhong",
""
]
] |
2508.15404
|
Anthony Bisulco
|
Anthony Bisulco, Rahul Ramesh, Randall Balestriero, Pratik Chaudhari
|
From Linearity to Non-Linearity: How Masked Autoencoders Capture Spatial Correlations
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Masked Autoencoders (MAEs) have emerged as a powerful pretraining technique for vision foundation models. Despite their effectiveness, they require extensive hyperparameter tuning (masking ratio, patch size, encoder/decoder layers) when applied to novel datasets. While prior theoretical works have analyzed MAEs in terms of their attention patterns and hierarchical latent variable models, the connection between MAE hyperparameters and performance on downstream tasks is relatively unexplored. This work investigates how MAEs learn spatial correlations in the input image. We analytically derive the features learned by a linear MAE and show that masking ratio and patch size can be used to select for features that capture short- and long-range spatial correlations. We extend this analysis to non-linear MAEs to show that MAE representations adapt to spatial correlations in the dataset, beyond second-order statistics. Finally, we discuss some insights on how to select MAE hyper-parameters in practice.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:52:12 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Bisulco",
"Anthony",
""
],
[
"Ramesh",
"Rahul",
""
],
[
"Balestriero",
"Randall",
""
],
[
"Chaudhari",
"Pratik",
""
]
] |
2508.15405
|
Jan Skacel
|
J. Skacel, P. Parodi, G. Gangemi, F. Bariselli, Z. Bonaventura and T. Magin
|
Simulation of Space Platform Charging in Very Low Earth Orbit Using Stochastic Particle Methods
| null | null | null | null |
physics.space-ph
|
http://creativecommons.org/licenses/by/4.0/
|
The residual atmospheric drag on satellites in Very Low Earth Orbit (VLEO) has been recognized as a limiting factor for satellite lifetimes. This work focuses on one component of the drag, i.e., the ionospheric drag from charged particles, which was observed to influence the overall aerodynamics of satellites. Space platform charging processes are modeled using a stochastic particle method, Particle-in-Cell, with with the possibility to treat the electrons as a Boltzmann fluid, thus reducing the computational time while preserving the accuracy of the physics of charging. Part of the study involves examining non-equilibrium distribution functions, specifically the Kappa distribution, commonly used to describe space plasmas. It has been shown that such distributions can noticeably affect overall space platform charging, thereby enhancing ionospheric drag on satellites using the proposed hybrid models. Nevertheless, it is acknowledged that ionospheric drag might constitute only a fraction of the total drag caused by neutral atmospheric gas.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:54:49 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Skacel",
"J.",
""
],
[
"Parodi",
"P.",
""
],
[
"Gangemi",
"G.",
""
],
[
"Bariselli",
"F.",
""
],
[
"Bonaventura",
"Z.",
""
],
[
"Magin",
"T.",
""
]
] |
2508.15406
|
Zhi Zhou
|
Tianhao Hu, Xinchi Huang, Bangti Jin, Qimeng Quan, Zhi Zhou
|
Conditional Stability and Numerical Reconstruction of a Parabolic Inverse Source Problem Using Carleman Estimates
|
inverse source problem, Carleman estimate, stability estimate, finite element method, error estimate
| null | null | null |
math.NA cs.NA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this work we develop a new numerical approach for recovering a spatially dependent source component in a standard parabolic equation from partial interior measurements. We establish novel conditional Lipschitz stability and H\"{o}lder stability for the inverse problem with and without boundary conditions, respectively, using suitable Carleman estimates. Then we propose a numerical approach for solving the inverse problem using conforming finite element approximations in both time and space. Moreover, by utilizing the conditional stability estimates, we prove rigorous error bounds on the discrete approximation. We present several numerical experiments to illustrate the effectiveness of the approach.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:54:50 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Hu",
"Tianhao",
""
],
[
"Huang",
"Xinchi",
""
],
[
"Jin",
"Bangti",
""
],
[
"Quan",
"Qimeng",
""
],
[
"Zhou",
"Zhi",
""
]
] |
2508.15407
|
Cheng Wang
|
Cheng Wang, Gelei Deng, Xianglin Yang, Han Qiu, Tianwei Zhang
|
When Audio and Text Disagree: Revealing Text Bias in Large Audio-Language Models
|
Accepted by EMNLP 2025 Main
| null | null | null |
cs.CL cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Large Audio-Language Models (LALMs) are enhanced with audio perception capabilities, enabling them to effectively process and understand multimodal inputs that combine audio and text. However, their performance in handling conflicting information between audio and text modalities remains largely unexamined. This paper introduces MCR-BENCH, the first comprehensive benchmark specifically designed to evaluate how LALMs prioritize information when presented with inconsistent audio-text pairs. Through extensive evaluation across diverse audio understanding tasks, we reveal a concerning phenomenon: when inconsistencies exist between modalities, LALMs display a significant bias toward textual input, frequently disregarding audio evidence. This tendency leads to substantial performance degradation in audio-centric tasks and raises important reliability concerns for real-world applications. We further investigate the influencing factors of text bias, and explore mitigation strategies through supervised finetuning, and analyze model confidence patterns that reveal persistent overconfidence even with contradictory inputs. These findings underscore the need for improved modality balance during training and more sophisticated fusion mechanisms to enhance the robustness when handling conflicting multi-modal inputs. The project is available at https://github.com/WangCheng0116/MCR-BENCH.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 09:58:24 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Wang",
"Cheng",
""
],
[
"Deng",
"Gelei",
""
],
[
"Yang",
"Xianglin",
""
],
[
"Qiu",
"Han",
""
],
[
"Zhang",
"Tianwei",
""
]
] |
2508.15408
|
Mikihito Nishi
|
Mikihito Nishi
|
K-Means Panel Data Clustering in the Presence of Small Groups
| null | null | null | null |
econ.EM stat.ME
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We consider panel data models with group structure. We study the asymptotic behavior of least-squares estimators and information criterion for the number of groups, allowing for the presence of small groups that have an asymptotically negligible relative size. Our contributions are threefold. First, we derive sufficient conditions under which the least-squares estimators are consistent and asymptotically normal. One of the conditions implies that a longer sample period is required as there are smaller groups. Second, we show that information criteria for the number of groups proposed in earlier works can be inconsistent or perform poorly in the presence of small groups. Third, we propose modified information criteria (MIC) designed to perform well in the presence of small groups. A Monte Carlo simulation confirms their good performance in finite samples. An empirical application illustrates that K-means clustering paired with the proposed MIC allows one to discover small groups without producing too many groups. This enables characterizing small groups and differentiating them from the other large groups in a parsimonious group structure.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:02:49 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Nishi",
"Mikihito",
""
]
] |
2508.15409
|
Carlos Merino
|
L.L. Jenkovszky and C. Merino
|
Critical Phenomena in Hadronic and DIS Processes
|
9 pages, 2 figures
| null | null | null |
hep-ph nucl-th
|
http://creativecommons.org/licenses/by/4.0/
|
We compare the critical phenomena (e.g. phase transitions, crossover) in proton-proton, proton-nucleus scattering and in lepton-proton deep inelastic scattering (DIS) systems.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:03:08 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Jenkovszky",
"L. L.",
""
],
[
"Merino",
"C.",
""
]
] |
2508.15410
|
Chris Boldt
|
Jivesh Kaushal, Chris Boldt, Stefan Scheel, Athanasios Laliotis, and Paolo Pedri
|
Multipole expansion for dispersion forces -- watch this trace
|
6 pages, plus 4 pages supplementary material
| null |
10.1103/1q4p-stl6
| null |
quant-ph physics.atom-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Light-matter interaction models invariably rely on the multipole expansion of the electromagnetic potentials generated by complex charge distributions. These multipoles are typically taken to be traceless, however, for a correct evaluation of dispersion forces at all distances, the validity of this assumption has to be checked carefully. Here, we revisit the concept of dispersion forces on an atom near a dielectric surface from the perspective of macroscopic quantum electrodynamics and find that, beyond the quadrupole, the multipoles cannot always be taken as fully traceless. In particular, we show that the trace of the octupole moment contributes to Casimir-Polder interactions beyond the electrostatic regime.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:03:28 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kaushal",
"Jivesh",
""
],
[
"Boldt",
"Chris",
""
],
[
"Scheel",
"Stefan",
""
],
[
"Laliotis",
"Athanasios",
""
],
[
"Pedri",
"Paolo",
""
]
] |
2508.15411
|
Frederik Vandeputte
|
Frederik Vandeputte
|
Foundational Design Principles and Patterns for Building Robust and Adaptive GenAI-Native Systems
| null | null | null | null |
cs.SE cs.CL cs.LG cs.MA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Generative AI (GenAI) has emerged as a transformative technology, demonstrating remarkable capabilities across diverse application domains. However, GenAI faces several major challenges in developing reliable and efficient GenAI-empowered systems due to its unpredictability and inefficiency. This paper advocates for a paradigm shift: future GenAI-native systems should integrate GenAI's cognitive capabilities with traditional software engineering principles to create robust, adaptive, and efficient systems.
We introduce foundational GenAI-native design principles centered around five key pillars -- reliability, excellence, evolvability, self-reliance, and assurance -- and propose architectural patterns such as GenAI-native cells, organic substrates, and programmable routers to guide the creation of resilient and self-evolving systems. Additionally, we outline the key ingredients of a GenAI-native software stack and discuss the impact of these systems from technical, user adoption, economic, and legal perspectives, underscoring the need for further validation and experimentation. Our work aims to inspire future research and encourage relevant communities to implement and refine this conceptual framework.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:05:18 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Vandeputte",
"Frederik",
""
]
] |
2508.15412
|
Amit Te'eni
|
Amit Te'eni, Eliahu Cohen
|
Equivalence of mutually unbiased bases via orbits: general theory and a $d=4$ case study
|
33 pages
| null | null | null |
math-ph math.MP quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
In quantum mechanics, mutually unbiased bases (MUBs) represent orthonormal bases that are as "far apart" as possible, and their classification reveals rich underlying geometric structure. Given a complex inner product space, we construct the space of its orthonormal bases as a discrete quotient of the complete flag manifold. We introduce a metric on this space, which corresponds to the "MUBness" distance. This allows us to describe equivalence between sets of mutually unbiased bases in terms of the geometry of this space. The subspace of bases that are unbiased with respect to the standard basis decomposes into orbits under a certain group action, and this decomposition corresponds to the classification of complex Hadamard matrices. More generally, we consider a list of $k$ MUBs, that one wishes to extend. The candidates are points in the subspace comprising all bases which are unbiased with respect to the entire list. This space also decomposes into orbits under a group action, and we prove that points in distinct orbits yield inequivalent MUB lists. Thus, we generalize the relation between complex Hadamard matrices and MUBs. As an application, we identify new symmetries that reduce the parameter space of MUB triples in dimension $4$ by a factor of $4$.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:06:39 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Te'eni",
"Amit",
""
],
[
"Cohen",
"Eliahu",
""
]
] |
2508.15413
|
Sizhen Bian
|
Pixi Kang and Julian Moosmann and Mengxi Liu and Bo Zhou and Michele Magno and Paul Lukowicz and Sizhen Bian
|
Bridging Generalization and Personalization in Wearable Human Activity Recognition via On-Device Few-Shot Learning
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Human Activity Recognition (HAR) using wearable devices has advanced significantly in recent years, yet its generalization remains limited when models are deployed to new users. This degradation in performance is primarily due to user-induced concept drift (UICD), highlighting the importance of efficient personalization. In this paper, we present a hybrid framework that first generalizes across users and then rapidly adapts to individual users using few-shot learning directly on-device. By updating only the classifier layer with user-specific data, our method achieves robust personalization with minimal computational and memory overhead. We implement this framework on the energy-efficient RISC-V-based GAP9 microcontroller and validate it across three diverse HAR scenarios: RecGym, QVAR-Gesture, and Ultrasound-Gesture. Post-deployment adaptation yields consistent accuracy improvements of 3.73\%, 17.38\%, and 3.70\% respectively. These results confirm that fast, lightweight, and effective personalization is feasible on embedded platforms, paving the way for scalable and user-aware HAR systems in the wild \footnote{https://github.com/kangpx/onlineTiny2023}.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:08:20 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Kang",
"Pixi",
""
],
[
"Moosmann",
"Julian",
""
],
[
"Liu",
"Mengxi",
""
],
[
"Zhou",
"Bo",
""
],
[
"Magno",
"Michele",
""
],
[
"Lukowicz",
"Paul",
""
],
[
"Bian",
"Sizhen",
""
]
] |
2508.15414
|
Yu Zhang
|
Wei Teng, Sheng Nan Wang, Yu Zhang, Xian Zhi Zhao, Xi Deng, Xiao Tong Li
|
B(E2) anomaly and triaxial deformation within a two-fluid SU(3) symmetry
|
6 pages, 3 figures
|
Chin. Phys. C, 49, 084106 (2025)
|
10.1088/1674-1137/adcd4a
| null |
nucl-th
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The correlation between $B(E2)$ structure and triaxial deformation has been investigated within the framework of the proton-neutron boson model. The analysis reveals that the distinctive feature, characterized by $B(E2;4_1^+\rightarrow2_1^+)/B(E2;2_1^+\rightarrow0_1^+)<1.0$ along with $E(4_1^+)/E(2_1^+)>2.0$, can emerge from the triaxial SU(3) symmetry inherent in two-fluid boson systems, attributed to band-mixing effects. This suggests a symmetry-based understanding of the anomalous $E2$ transitions observed in experiments.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:14:25 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Teng",
"Wei",
""
],
[
"Wang",
"Sheng Nan",
""
],
[
"Zhang",
"Yu",
""
],
[
"Zhao",
"Xian Zhi",
""
],
[
"Deng",
"Xi",
""
],
[
"Li",
"Xiao Tong",
""
]
] |
2508.15415
|
Mao Ye
|
Dengyan Luo, Yanping Xiang, Hu Wang, Luping Ji.Shuai Li, Mao Ye
|
Bidirectional Temporal Information Propagation for Moving Infrared Small Target Detection
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Moving infrared small target detection is broadly adopted in infrared search and track systems, and has attracted considerable research focus in recent years. The existing learning-based multi-frame methods mainly aggregate the information of adjacent frames in a sliding window fashion to assist the detection of the current frame. However, the sliding-window-based methods do not consider joint optimization of the entire video clip and ignore the global temporal information outside the sliding window, resulting in redundant computation and sub-optimal performance. In this paper, we propose a Bidirectional temporal information propagation method for moving InfraRed small target Detection, dubbed BIRD. The bidirectional propagation strategy simultaneously utilizes local temporal information of adjacent frames and global temporal information of past and future frames in a recursive fashion. Specifically, in the forward and backward propagation branches, we first design a Local Temporal Motion Fusion (LTMF) module to model local spatio-temporal dependency between a target frame and its two adjacent frames. Then, a Global Temporal Motion Fusion (GTMF) module is developed to further aggregate the global propagation feature with the local fusion feature. Finally, the bidirectional aggregated features are fused and input into the detection head for detection. In addition, the entire video clip is jointly optimized by the traditional detection loss and the additional Spatio-Temporal Fusion (STF) loss. Extensive experiments demonstrate that the proposed BIRD method not only achieves the state-of-the-art performance but also shows a fast inference speed.
|
[
{
"version": "v1",
"created": "Thu, 21 Aug 2025 10:17:21 GMT"
}
] | 2025-08-22T00:00:00 |
[
[
"Luo",
"Dengyan",
""
],
[
"Xiang",
"Yanping",
""
],
[
"Wang",
"Hu",
""
],
[
"Li",
"Luping Ji. Shuai",
""
],
[
"Ye",
"Mao",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.