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
2025-08-29 00:00:00
| authors_parsed
listlengths 1
3.31k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2508.20337
|
Daniel Orr
|
Daniel Orr and Johnny Rivera Jr
|
Combinatorial proof of a permuted basement Macdonald polynomial identity
| null | null | null | null |
math.CO math.QA math.RT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A well-known and fundamental property of the Macdonald polynomials $P_\lambda(x;q,t)$ is their invariance under the transformation sending $(q,t)$ to $(q^{-1},t^{-1})$. Recently, Concha and Lapointe showed that this property extends in an interesting, nontrivial way to an identity for partially symmetric Macdonald polynomials. Their identity played a key role in the work of Bechtloff Weising and Orr linking partially symmetric Macdonald polynomials to parabolic flag Hilbert schemes. In this paper, we refine the Concha-Lapointe identity to a sub-family of Alexandersson's permuted basement Macdonald polynomials and give a combinatorial proof of the refined identity. We show also that the Concha-Lapointe identity is equivalent to the assertion that (normalized) partially symmetric Macdonald polynomials are fixed under the Kazhdan-Lusztig involution.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 00:45:45 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Orr",
"Daniel",
""
],
[
"Rivera",
"Johnny",
"Jr"
]
] |
2508.20338
|
Rahul Rao
|
Evan W. Muller, Aleksey Ruditskiy, Jie Jiang, Thuc T. Mai, Katherine Burzynski, Ruth Pachter, Michael F. Durstock, W. Joshua Kennedy, Rahul Rao
|
Strong Raman Optical Activity and Chiral Phonons in Chiral Hybrid Organic-Inorganic Perovskites
|
6 Figures
| null | null | null |
cond-mat.mtrl-sci physics.chem-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Hybrid organic-inorganic perovskites with chiral organic cations are very interesting for optoelectronic applications because of their intrinsically chiral light-matter interactions. Chiral distortions in these materials lead to circular dichroism, circular birefringence, and circularly polarized luminescence in the band transitions of the inorganic sublattice. Raman-active vibrational modes in these crystals are governed by crystal symmetry and therefore are also strongly impacted by the nature and magnitude of the chiral distortions. Here, we report low-frequency Raman modes that are sensitive to circularly polarized excitation in chiral hybrid organic-inorganic perovskites (CHOIPs) across a wide range of structures and compositions. The circularly polarized Raman spectra from enantiomers of CHOIP single crystals exhibit sharp modes below 150 cm-1, corresponding to vibrations of the lead iodide octahedra. These modes exhibit strong differences in intensities (Raman optical activity, ROA) depending on the handedness of the excitation, with high degree of polarization for several modes. Calculations reveal the presence of several chiral phonon modes with opposite phonon angular momenta. The strong ROA and the chiral phonon modes are a direct consequence of chirality transfer from the chiral organic linker to the lead iodide octahedra in the CHOIP structure, resulting in a strong chiroptical response in the phonon modes.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:10:05 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Muller",
"Evan W.",
""
],
[
"Ruditskiy",
"Aleksey",
""
],
[
"Jiang",
"Jie",
""
],
[
"Mai",
"Thuc T.",
""
],
[
"Burzynski",
"Katherine",
""
],
[
"Pachter",
"Ruth",
""
],
[
"Durstock",
"Michael F.",
""
],
[
"Kennedy",
"W. Joshua",
""
],
[
"Rao",
"Rahul",
""
]
] |
2508.20339
|
Max Kreider
|
Max Kreider, Peter J. Thomas, Yao Li
|
Artificial neural network solver for Fokker-Planck and Koopman eigenfunctions
| null | null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
For a stochastic differential equation (SDE) that is an It\^{o} diffusion or Langevin equation, the Fokker-Planck operator governs the evolution of the probability density, while its adjoint, the infinitesimal generator of the stochastic Koopman operator, governs the evolution of system observables, in the mean. The eigenfunctions of these operators provide a powerful framework to analyze SDEs, and have shown to be particularly useful for systems of stochastic oscillators. However, computing these eigenfunctions typically requires solving high-dimensional PDEs on unbounded domains, which is numerically challenging. Building on previous work, we propose a data-driven artificial neural network solver for Koopman and Fokker-Planck eigenfunctions. Our approach incorporates the differential operator into the loss function, improving accuracy and reducing dependence on large amounts of accurate training data. We demonstrate our approach on several numerical examples in two, three, and four dimensions.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:21:17 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kreider",
"Max",
""
],
[
"Thomas",
"Peter J.",
""
],
[
"Li",
"Yao",
""
]
] |
2508.20340
|
Maolin Sun
|
Maolin Sun, Yibiao Yang, Yuming Zhou
|
Boosting Skeleton-Driven SMT Solver Fuzzing by Leveraging LLM to Produce Formula Generators
| null | null | null | null |
cs.SE cs.AI cs.PL
|
http://creativecommons.org/licenses/by/4.0/
|
Satisfiability Modulo Theory (SMT) solvers are foundational to modern systems and programming languages research, providing the foundation for tasks like symbolic execution and automated verification. Because these solvers sit on the critical path, their correctness is essential, and high-quality test formulas are key to uncovering bugs. However, while prior testing techniques performed well on earlier solver versions, they struggle to keep pace with rapidly evolving features. Recent approaches based on Large Language Models (LLMs) show promise in exploring advanced solver capabilities, but two obstacles remain: nearly half of the generated formulas are syntactically invalid, and iterative interactions with the LLMs introduce substantial computational overhead. In this study, we present Chimera, a novel LLM-assisted fuzzing framework that addresses both issues by shifting from direct formula generation to the synthesis of reusable term (i.e., logical expression) generators. Particularly, Chimera uses LLMs to (1) automatically extract context-free grammars (CFGs) for SMT theories, including solver-specific extensions, from documentation, and (2) synthesize composable Boolean term generators that adhere to these grammars. During fuzzing, Chimera populates structural skeletons derived from existing formulas with the terms iteratively produced by the LLM-synthesized generators. This design ensures syntactic validity while promoting semantic diversity. Notably, Chimera requires only one-time LLM interaction investment, dramatically reducing runtime cost. We evaluated Chimera on two leading SMT solvers: Z3 and cvc5. Our experiments show that Chimera has identified 43 confirmed bugs, 40 of which have already been fixed by developers.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:21:26 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Sun",
"Maolin",
""
],
[
"Yang",
"Yibiao",
""
],
[
"Zhou",
"Yuming",
""
]
] |
2508.20341
|
Katsuhiko Matsuzaki
|
Katsuhiko Matsuzaki
|
Integrable Teichm\"uller spaces for analysis on Weil-Petersson curves
| null | null | null | null |
math.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The integrableTeichm\"uller space $T_p$ for $p \geq 1$ is defined by the $p$-integrability of Beltrami coefficients. We characterize a quasisymmetric homeomorphism $h$ in $T_p$ by the condition that $\log h'$ belongs to the real $p$-Besov space, with a certain modification applied in the case $p=1$. This is done as part of the arguments for establishing a biholomorphic correspondence $\Lambda$ from the product of $T_p$ for simultaneous uniformization of $p$-Weil-Petersson curves into the $p$-Besov space. In particular, this proves the real-analytic equivalence between $T_p$ and the real $p$-Besov space. Moreover, the Cauchy transform of Besov functions on Weil-Petersson curves can be expressed by the derivative of this holomorphic map $\Lambda$, and from this, the Calder\'on theorem in this setting is straightforward. It also follows that the Cauchy transforms on $p$-Weil-Petersson curves holomorphically depend on their embeddings as they vary in the Bers coordinates.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:29:06 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Matsuzaki",
"Katsuhiko",
""
]
] |
2508.20342
|
Pablo Rocha
|
Pablo Rocha
|
The $H^p(\mathbb{Z}^n)-H^q(\mathbb{Z}^n)$ boundedness of discrete Riesz potential
|
12 pages
| null | null | null |
math.CA
|
http://creativecommons.org/licenses/by/4.0/
|
In [J. Class. Anal., vol. 26 (1) (2025), 63-76], we proved that the discrete Riesz potential $I_{\alpha}$ is a bounded operator $H^p(\mathbb{Z}^n) \to H^q(\mathbb{Z}^n)$ for $\frac{n-1}{n} < p \leq 1$, $\frac{1}{q} = \frac{1}{p} - \frac{\alpha}{n}$ and $0 < \alpha < n$. In this note, we extend such boundedness on the full range $0 < p \leq 1$.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:31:22 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Rocha",
"Pablo",
""
]
] |
2508.20343
|
Justin Bergfield
|
Eren Erdogan, Justin P. Bergfield
|
When Dephasing Fails: Thermodynamic Consequences of Decoherence Models in Quantum Transport
|
14 pages, 6 figures
| null | null | null |
cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding how decoherence influences heat and information flow is essential for realizing the promise of quantum technologies. Two widely used models for incorporating decoherence in quantum transport are the voltage probe (VP), which imposes local charge current conservation, and the voltage-temperature probe (VTP), which also conserves heat current. Although these models are often treated as functionally equivalent, we demonstrate that this equivalence actually exists only under highly symmetric conditions, which may be challenging to achieve experimentally Under asymmetric coupling or thermal bias, the VTP respects thermodynamic constraints and enforces decoherence in both charge and heat channels, while the VP instead acts as a source or sink of heat. Strikingly, the VP can fail to model decoherence in the heat transport entirely, even with large probe coupling strengths. Using a benzene-based molecular junction as a realistic example, we show that these effects significantly impact the predicted heat transport. These results establish that the VP and VTP models are not interchangeable: only the VTP provides a thermodynamically consistent framework for modeling decoherence in quantum transport.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:35:17 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Erdogan",
"Eren",
""
],
[
"Bergfield",
"Justin P.",
""
]
] |
2508.20344
|
Hancheng Min
|
Hancheng Min, Ren\'e Vidal
|
Understanding Incremental Learning with Closed-form Solution to Gradient Flow on Overparamerterized Matrix Factorization
|
Accepted to CDC 2025
| null | null | null |
cs.LG cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Many theoretical studies on neural networks attribute their excellent empirical performance to the implicit bias or regularization induced by first-order optimization algorithms when training networks under certain initialization assumptions. One example is the incremental learning phenomenon in gradient flow (GF) on an overparamerterized matrix factorization problem with small initialization: GF learns a target matrix by sequentially learning its singular values in decreasing order of magnitude over time. In this paper, we develop a quantitative understanding of this incremental learning behavior for GF on the symmetric matrix factorization problem, using its closed-form solution obtained by solving a Riccati-like matrix differential equation. We show that incremental learning emerges from some time-scale separation among dynamics corresponding to learning different components in the target matrix. By decreasing the initialization scale, these time-scale separations become more prominent, allowing one to find low-rank approximations of the target matrix. Lastly, we discuss the possible avenues for extending this analysis to asymmetric matrix factorization problems.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:36:42 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Min",
"Hancheng",
""
],
[
"Vidal",
"René",
""
]
] |
2508.20345
|
Xiao Li
|
Xiao Li, Yanfan Zhu, Ruining Deng, Wei-Qi Wei, Yu Wang, Shilin Zhao, Yaohong Wang, Haichun Yang, Yuankai Huo
|
MedFoundationHub: A Lightweight and Secure Toolkit for Deploying Medical Vision Language Foundation Models
| null | null | null | null |
cs.CV cs.HC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent advances in medical vision-language models (VLMs) open up remarkable opportunities for clinical applications such as automated report generation, copilots for physicians, and uncertainty quantification. However, despite their promise, medical VLMs introduce serious security concerns, most notably risks of Protected Health Information (PHI) exposure, data leakage, and vulnerability to cyberthreats - which are especially critical in hospital environments. Even when adopted for research or non-clinical purposes, healthcare organizations must exercise caution and implement safeguards. To address these challenges, we present MedFoundationHub, a graphical user interface (GUI) toolkit that: (1) enables physicians to manually select and use different models without programming expertise, (2) supports engineers in efficiently deploying medical VLMs in a plug-and-play fashion, with seamless integration of Hugging Face open-source models, and (3) ensures privacy-preserving inference through Docker-orchestrated, operating system agnostic deployment. MedFoundationHub requires only an offline local workstation equipped with a single NVIDIA A6000 GPU, making it both secure and accessible within the typical resources of academic research labs. To evaluate current capabilities, we engaged board-certified pathologists to deploy and assess five state-of-the-art VLMs (Google-MedGemma3-4B, Qwen2-VL-7B-Instruct, Qwen2.5-VL-7B-Instruct, and LLaVA-1.5-7B/13B). Expert evaluation covered colon cases and renal cases, yielding 1015 clinician-model scoring events. These assessments revealed recurring limitations, including off-target answers, vague reasoning, and inconsistent pathology terminology.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:39:16 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Li",
"Xiao",
""
],
[
"Zhu",
"Yanfan",
""
],
[
"Deng",
"Ruining",
""
],
[
"Wei",
"Wei-Qi",
""
],
[
"Wang",
"Yu",
""
],
[
"Zhao",
"Shilin",
""
],
[
"Wang",
"Yaohong",
""
],
[
"Yang",
"Haichun",
""
],
[
"Huo",
"Yuankai",
""
]
] |
2508.20346
|
Sergei Artemov
|
Sergei Artemov
|
Consistency formula is strictly stronger in PA than PA-consistency
| null | null | null | null |
math.LO
|
http://creativecommons.org/licenses/by/4.0/
|
In this note, we show that, despite the widespread assumption, the consistency formula for Peano Arithmetic PA, Con(PA), "for all x, x is not a code of a derivation of (0=1)," is not equivalent in PA to the consistency of PA. Specifically, we demonstrate that "PA is consistent" is provably in PA equivalent to the series ConS(PA) of arithmetical sentences "n is not a code of a derivation of (0=1)" for n=0,1,2,.... Since Con(PA) is strictly stronger in PA than ConS(PA), the unprovability of Con(PA) in PA does not yield the unprovability of PA-consistency.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:39:59 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Artemov",
"Sergei",
""
]
] |
2508.20347
|
Fumika Suzuki
|
Fumika Suzuki, Ying Wai Li, Wojciech H. Zurek
|
Machine learning topological defect formation
|
7 pages, 5 figures
| null | null |
LA-UR-25-27255
|
cond-mat.stat-mech gr-qc hep-ph hep-th physics.comp-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
According to the Kibble-Zurek mechanism (KZM), the density of topological defects created during a second-order phase transition is determined by the correlation length at the freeze-out time. This suggests that the final configuration of topological defects in such a transition is largely established during the impulse regime, soon after the critical point is traversed. Motivated by this, we conjecture that machine learning (ML) can predict the final configuration of topological defects based on the time evolution of the order parameter over a short interval in the vicinity of the critical point, well before the order parameter settles into the emerging new minima resulting from spontaneous symmetry breaking. Furthermore, we show that the predictability of ML also follows the power law scaling dictated by KZM. We demonstrate these using a Recurrent Neural Network.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:41:08 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Suzuki",
"Fumika",
""
],
[
"Li",
"Ying Wai",
""
],
[
"Zurek",
"Wojciech H.",
""
]
] |
2508.20348
|
Osvaldo L. Santos-Pereira
|
Osvaldo L. Santos-Pereira
|
The Warp Drive: Superluminal Travel within General Relativity
|
128 pages, 8 figures, 9 tables. LaTeX. PhD thesis
| null | null | null |
gr-qc
|
http://creativecommons.org/licenses/by/4.0/
|
In 1994, Miguel Alcubierre proposed that the well-known special relativistic limitation that particles cannot travel with velocities higher than light speed can be bypassed when such trips are considered globally within specific general relativistic frameworks, using a warped region of spacetime in the shape of a bubble that transports particles with mass traveling through spacetime with superluminal speed. Although initial results indicated this scenario as nonphysical, since it would seem to require negative mass-energy density, recent theoretical analyses suggest that such a nonphysical situation may not always be true. This thesis presents newfound solutions for the Einstein field equations, considering the Alcubierre warp drive spacetime metrics. The central premise is to study the fluid matter as the gravity source, rather than the more common vacuum or negative energy sources, to explore the potential for generating superluminal velocities, or \textit{warp speeds}, through a warped region in the spacetime. Such solutions have various matter-energy sources: dust particles, perfect fluid, quasi-perfect fluid with anisotropic pressures, charged dust, and a perfect fluid within a cosmological constant spacetime.
A connection between some of these solutions featuring shock waves described by a Burgers-type equation with a term on the right-hand side of the equation purely dependent on time is also shown. This could mean warp drives are closely related to vacuum energy and possibly have topological effects such as shock waves.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:44:10 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Santos-Pereira",
"Osvaldo L.",
""
]
] |
2508.20349
|
Zhiqiang Cao
|
Zhiqiang Cao, Scott Zuo, Mary Ryan Baumann, Kendra Plourde, Patrick Heagerty, Guangyu Tong and Fan Li
|
Covariate-adjusted win statistics in randomized clinical trials with ordinal outcomes
| null | null | null | null |
stat.ME
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Ordinal outcomes are common in clinical settings where they often represent increasing levels of disease progression or different levels of functional impairment. Such outcomes can characterize differences in meaningful patient health states that are directly relevant to clinical researchers and frequently represent composite outcomes that include absorbing states such as death. To compare different intervention strategies in clinical trials, the direct use of ordinal logistic regression models may not be ideal for analyzing ranked outcomes due to non-collapsibility, lack of estimation and clarity, or failure of the common underlying proportional odds assumption. In this article, we focus on representing the average treatment effect for ordinal outcomes via intrinsic pairwise outcome comparisons captured through win estimates, such as the win ratio and win difference. We first develop propensity score weighting estimators, including both inverse probability weighting (IPW) and overlap weighting (OW), tailored to estimating win parameters. Furthermore, we develop augmented weighting estimators that leverage an additional ordinal outcome regression to potentially improve efficiency over weighting alone. Leveraging the theory of U-statistics, we establish the asymptotic theory for all estimators, and derive closed-form variance estimators to support statistical inference. Through extensive simulations we demonstrate the enhanced efficiency of the weighted estimators over the unadjusted estimator, with the augmented weighting estimators showing a further improvement in efficiency except for extreme cases. Finally, we illustrate our proposed methods with the ORCHID trial, and implement our covariate adjustment methods in an R package winPSW to facilitate the practical implementation.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:49:17 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Cao",
"Zhiqiang",
""
],
[
"Zuo",
"Scott",
""
],
[
"Baumann",
"Mary Ryan",
""
],
[
"Plourde",
"Kendra",
""
],
[
"Heagerty",
"Patrick",
""
],
[
"Tong",
"Guangyu",
""
],
[
"Li",
"Fan",
""
]
] |
2508.20350
|
Penghua Ying
|
Yu Bao and Keke Song and Jiahui Liu and Yanzhou Wang and Yifei Ning and Penghua Ying and Ping Qian
|
Atomistic understanding of hydrogen bubble-induced embrittlement in tungsten enabled by machine learning molecular dynamics
|
14pages,7 figures
| null | null | null |
cond-mat.mtrl-sci
|
http://creativecommons.org/publicdomain/zero/1.0/
|
Hydrogen bubble formation within nanoscale voids is a critical mechanism underlying the embrittlement of metallic materials, yet its atomistic origins remains elusive. Here, we present an accurate and transferable machine-learned potential (MLP) for the tungsten-hydrogen binary system within the neuroevolution potential (NEP) framework, trained through active learning on extensive density functional theory data. The developed NEP-WH model reproduces a wide range of lattice and defect properties in tungsten systems, as well as hydrogen solubility, with near first-principles accuracy, while retaining the efficiency of empirical potentials. Crucially, it is the first MLP capable of capturing hydrogen trapping and H\textsubscript{2} formation in nanovoids, with quantitative fidelity. Large-scale machine-learning molecular dynamics simulations reveal a distinct aggregation pathway where planar hydrogen clusters nucleate and grow along \{100\} planes near voids, with hexagonal close-packed structures emerging at their intersections. Under uniaxial tension, these aggregates promote bubble fracture and the development of regular \{100\} cracks, suppressing dislocation activity and resulting in brittle fracture behavior. This work provides detailed atomistic insights into hydrogen bubble evolution and fracture in nanovoids, enables predictive modeling of structural degradation in extreme environments, and advances fundamental understanding of hydrogen-induced damage in structural metals.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:53:59 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Bao",
"Yu",
""
],
[
"Song",
"Keke",
""
],
[
"Liu",
"Jiahui",
""
],
[
"Wang",
"Yanzhou",
""
],
[
"Ning",
"Yifei",
""
],
[
"Ying",
"Penghua",
""
],
[
"Qian",
"Ping",
""
]
] |
2508.20351
|
Zhirui Chen
|
Zhirui Chen, Wei Shen, Jiashui Huang, Ling Shao
|
Joint Enhancement of Relational Reasoning for Long-Context LLMs
|
9 pages, 5 pages Accepted by EMNLP 2025 Findings
| null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Despite significant progress, large language models (LLMs) still struggle with long contexts due to memory limitations and their inability to tackle complex and long-context tasks. Additionally, LLMs often suffer from a lack of transparency and are prone to producing hallucinations. To address these challenges, we propose \textbf{JERR}, a novel framework designed to enhance long-context comprehension via graph-based reasoning in LLMs. JERR integrates three key components: synopsis extraction, graph construction, and relational reasoning. First, synopsis is extracted by chunking text strategically, allowing the model to summarize and understand information more efficiently. Second, we build a directed acyclic graph (DAG) to resolve redundancy, ensuring logical consistency and clarity. Finally, we incorporate Monte Carlo Tree Search (MCTS) to help the model navigate complex reasoning paths, ensuring more accurate and interpretable outputs. This framework provides a novel solution that enables LLMs to handle extended contexts and complex reasoning tasks with improved reliability and transparency. Experimental results show that JERR consistently outperforms all baselines on the ROUGE and F1 metrics, achieving the highest scores on the LLM-Rater evaluation.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 01:54:47 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Chen",
"Zhirui",
""
],
[
"Shen",
"Wei",
""
],
[
"Huang",
"Jiashui",
""
],
[
"Shao",
"Ling",
""
]
] |
2508.20352
|
Justin Bergfield
|
Justin P. Bergfield
|
Quantum Interference Supernodes, Thermoelectric Enhancement, and the Role of Dephasing
|
19 pages, 6 figures
| null | null | null |
cond-mat.mes-hall cond-mat.other cond-mat.stat-mech
|
http://creativecommons.org/licenses/by/4.0/
|
Quantum interference (QI) can strongly enhance thermoelectric response, with higher-order "supernodes" predicted to yield scalable gains in thermopower and efficiency. A central question, however, is whether such features are intrinsically more fragile to dephasing. Using $B\"uttiker$ voltage-temperature probes, we establish an order-selection rule: the effective near-node order is set by the lowest among coherent and probe-assisted channels. Supernodes are therefore fragile in an absolute sense because their transmission is parametrically suppressed with order. However, once an incoherent floor dominates, the fractional suppression of thermopower, efficiency, and figure of merit becomes universal and order-independent. Illustrating these principles with benzene- and biphenyl-based junction calculations, we show that the geometry of environmental coupling -- through a single orbital or across many -- dictates whether coherence is lost by order reduction or by floor building. These results yield general scaling rules for the thermoelectric response of interference nodes under dephasing.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:00:21 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Bergfield",
"Justin P.",
""
]
] |
2508.20353
|
Zhibang Yang
|
Zhibang Yang, Xinke Jiang, Rihong Qiu, Ruiqing Li, Yihang Zhang, Yue Fang, Yongxin Xu, Hongxin Ding, Xu Chu, Junfeng Zhao, Yasha Wang
|
DFAMS: Dynamic-flow guided Federated Alignment based Multi-prototype Search
|
7 pages, 3 figures
| null | null | null |
cs.LG cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Federated Retrieval (FR) routes queries across multiple external knowledge sources, to mitigate hallucinations of LLMs, when necessary external knowledge is distributed. However, existing methods struggle to retrieve high-quality and relevant documents for ambiguous queries, especially in cross-domain scenarios, which significantly limits their effectiveness in supporting downstream generation tasks. Inspired by dynamic information flow (DIF), we propose DFAMS, a novel framework that leverages DIF to identify latent query intents and construct semantically aligned knowledge partitions for accurate retrieval across heterogeneous sources. Specifically, DFAMS probes the DIF in LLMs by leveraging gradient signals from a few annotated queries and employing Shapley value-based attribution to trace neuron activation paths associated with intent recognition and subdomain boundary detection. Then, DFAMS leverages DIF to train an alignment module via multi-prototype contrastive learning, enabling fine-grained intra-source modeling and inter-source semantic alignment across knowledge bases. Experimental results across five benchmarks show that DFAMS outperforms advanced FR methods by up to 14.37% in knowledge classification accuracy, 5.38% in retrieval recall, and 6.45% in downstream QA accuracy, demonstrating its effectiveness in complex FR scenarios.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:00:48 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Yang",
"Zhibang",
""
],
[
"Jiang",
"Xinke",
""
],
[
"Qiu",
"Rihong",
""
],
[
"Li",
"Ruiqing",
""
],
[
"Zhang",
"Yihang",
""
],
[
"Fang",
"Yue",
""
],
[
"Xu",
"Yongxin",
""
],
[
"Ding",
"Hongxin",
""
],
[
"Chu",
"Xu",
""
],
[
"Zhao",
"Junfeng",
""
],
[
"Wang",
"Yasha",
""
]
] |
2508.20354
|
Andrey Baydin
|
Andrey Baydin, Jiaming Luo, Zhiren He, Jacques Doumani, Tong Lin, Fuyang Tay, Jiaming He, Jianshi Zhou, Guru Khalsa, Junichiro Kono and Hanyu Zhu
|
Lattice-induced spin dynamics in Dirac magnet CoTiO3
| null | null | null | null |
cond-mat.mtrl-sci cond-mat.other physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Spin-lattice coupling is crucial for understanding the spin transport and dynamics for spintronics and magnonics applications. Recently, cobalt titanate (CoTiO3), an easy-plane antiferromagnet, has been found to host axial phonons with a large magnetic moment, which may originate from spin-lattice coupling. Here, we investigate the effect of light-driven lattice dynamics on the magnetic properties of CoTiO3 using time-resolved spectroscopy with a THz pump and a magneto-optic probe. We found resonantly driven Raman active phonons, phonon-polariton-induced excitation of the antiferromagnetic magnons, and a slow increase in the polarization rotation of the probe, all indicating symmetry breaking that is not intrinsic to the magnetic space group. The temperature dependence confirmed that the observed spin dynamics is related to the magnetic order, and we suggest surface effects as a possible mechanism. Our results of THz-induced spin-lattice dynamics signify that extrinsic symmetry breaking may contribute strongly and unexpectedly to light-driven phenomena in bulk complex oxides.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:10:51 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Baydin",
"Andrey",
""
],
[
"Luo",
"Jiaming",
""
],
[
"He",
"Zhiren",
""
],
[
"Doumani",
"Jacques",
""
],
[
"Lin",
"Tong",
""
],
[
"Tay",
"Fuyang",
""
],
[
"He",
"Jiaming",
""
],
[
"Zhou",
"Jianshi",
""
],
[
"Khalsa",
"Guru",
""
],
[
"Kono",
"Junichiro",
""
],
[
"Zhu",
"Hanyu",
""
]
] |
2508.20355
|
Judit Ferrer Asensio
|
J. Ferrer Asensio, S. S. Jensen, S. Spezzano, P. Caselli, F. O. Alves, O. Sipil\"a and E. Redaelli
|
High sensitivity molecular line observations towards the L1544 pre-stellar core challenge current models
| null | null | null | null |
astro-ph.GA
|
http://creativecommons.org/licenses/by/4.0/
|
The increased sensitivity and spectral resolution of observed spectra towards the pre-stellar core L1544 are challenging the current physical and chemical models. With the aim of further constraining the structure of L1544 as well as assessing the completeness of chemical networks, we turn to radiative transfer modelling of observed molecular lines towards this source. We obtained high-sensitivity and high-spectral resolution observations of HCO+ (J = 1 - 0), CS (J = 2 - 1), C34S (J = 2 - 1), H2CO (J ,Ka,Kc = 2,1,2 - 1,1,1), c-C3H2 (J,Ka,Kc = 2,1,2 - 1,0,1), SO (N,J = 2,3 - 1,2) and 34SO (N,J = 2,3 - 1,2) with the IRAM 30m telescope towards the dust peak of L1544. A non-Local Thermodynamic Equilibrium radiative transfer code is coupled to the Markov Chain Monte Carlo method to model the observations. We find that with just one transition for each isotope, the modelling cannot find a global minimum that fits the observations. The derived fractional abundance profiles are compared to those computed with chemical models. The observed transitions trace gas components with distinct dynamics at different distances along the radius of the core. Moreover, the results evidence a poor reproduction of sulphur chemistry by chemical modelling and stresses the need to include a more consistent S-depletion process to accurately reproduce the S-chemistry towards dense cores.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:11:00 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Asensio",
"J. Ferrer",
""
],
[
"Jensen",
"S. S.",
""
],
[
"Spezzano",
"S.",
""
],
[
"Caselli",
"P.",
""
],
[
"Alves",
"F. O.",
""
],
[
"Sipilä",
"O.",
""
],
[
"Redaelli",
"E.",
""
]
] |
2508.20356
|
Guohui Dong
|
Guohui Dong and Yao Yao
|
Ultranarrow Superradiant Laser via Dressed Dark State
| null | null | null | null |
physics.atom-ph quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Superradiant laser, which exploits the clock transition of alkaline-earth-metal-like atoms to generate ultrastable light in the bad-cavity limit, has garnered much attention in the past few decades. Unlike their odd counterpart, the even isotopes of alkaline-earth-metal-like atoms possess simpler structures and longer-lived transitions, which would relax the field control requirements and enhance the frequency stability of the output light. However, due to the absence of hyper-fine interaction in even isotopes, the transition from the state $^{3}\mathrm{P}_{0}$ to $^{1}$S$_{0}$ is strictly forbidden, leading to a vanishing coupling strength between the cavity mode and the atoms (the state $^{3}\mathrm{P}_{0}$ as a dark state). In this work, we suggest a superradiant laser scheme by dressing this dark state with a small bright component by virtue of a static magnetic field. In contrast to other proposals utilizing natural atomic transitions, our dressed-state protocol can work from the crossover regime (coherence in both atoms and photons) to the superradiant lasing regime (coherence solely in atoms). Specially, by operating deep into the superradiant lasing regime, our scheme witnesses a dramatic line-narrowing feature (mHz level) while maintaining its power. Furthermore, compared to the crossover regime, the laser frequency in the superradiant lasing regime is more robust against the fluctuations of the cavity length and magnetic field strength. Our proposal demonstrates the potential for extracting ultranarrow light from even isotopes of the alkaline-earth-metal-like atoms and may find its role in frequency stabilization and related precision measurement scenarios.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:11:38 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Dong",
"Guohui",
""
],
[
"Yao",
"Yao",
""
]
] |
2508.20357
|
Yeray Garcia Del Castillo
|
Yeray Garcia del Castillo, Giovanni Pierobon, Dipan Sengupta, Yvonne Y. Y. Wong
|
Prospects for relic neutrino detection using nuclear spin experiments
|
34 pages, 12 figures, numerical codes and solutions are available at https://github.com/gpierobon/pyOpenNu and https://github.com/gpierobon/OpenNu
| null | null |
CPPC-2025-06
|
hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Direct detection of the cosmic neutrino background (C$\nu$B) remains one of the most formidable experimental challenges in modern physics. In this work, we extend recent studies of C$\nu$B-induced coherent transitions in polarised nuclear spin ensembles. Adopting an open quantum system framework, we model coherent neutrino effects in large spin ensembles using a Lindblad master equation that also incorporates realistic experimental imperfections such as local dephasing and imperfect polarisation. We solve the Lindblad equation numerically by way of a fast and computationally inexpensive method that can be extended to an arbitrarily large number of spins. Using our numerical solutions, we forecast the sensitivities of future experiments such as CASPEr to the local C$\nu$B overdensity parameter $\delta_\nu$. Our findings indicate that a CASPEr-like experiment, though primarily aimed at axion dark matter search, could also constrain the C$\nu$B overdensity to $\delta_\nu \sim 10^{9}-10^{11}$ in configurations achievable by currently planned experimental efforts, and down to $\delta_\nu \sim 10^7$ in the most optimised scenario. While C$\nu$B detection remains out of reach in the foreseeable future, our results highlight the potential of using quantum sensing to probe fundamental physics.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:12:29 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"del Castillo",
"Yeray Garcia",
""
],
[
"Pierobon",
"Giovanni",
""
],
[
"Sengupta",
"Dipan",
""
],
[
"Wong",
"Yvonne Y. Y.",
""
]
] |
2508.20358
|
Abhishek Indupally
|
Abhishek Indupally, Satchit Ramnath
|
Developing a Multi-Modal Machine Learning Model For Predicting Performance of Automotive Hood Frames
| null | null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Is there a way for a designer to evaluate the performance of a given hood frame geometry without spending significant time on simulation setup? This paper seeks to address this challenge by developing a multimodal machine-learning (MMML) architecture that learns from different modalities of the same data to predict performance metrics. It also aims to use the MMML architecture to enhance the efficiency of engineering design processes by reducing reliance on computationally expensive simulations. The proposed architecture accelerates design exploration, enabling rapid iteration while maintaining high-performance standards, especially in the concept design phase. The study also presents results that show that by combining multiple data modalities, MMML outperforms traditional single-modality approaches. Two new frame geometries, not part of the training dataset, are also used for prediction using the trained MMML model to showcase the ability to generalize to unseen frame models. The findings underscore MMML's potential in supplementing traditional simulation-based workflows, particularly in the conceptual design phase, and highlight its role in bridging the gap between machine learning and real-world engineering applications. This research paves the way for the broader adoption of machine learning techniques in engineering design, with a focus on refining multimodal approaches to optimize structural development and accelerate the design cycle.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:15:54 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Indupally",
"Abhishek",
""
],
[
"Ramnath",
"Satchit",
""
]
] |
2508.20359
|
Shijia Wang
|
Shijia Wang, Tianpei Ouyang, Qiang Xiao, Dongjing Wang, Yintao Ren, Songpei Xu, Da Guo, Chuanjiang Luo
|
Progressive Semantic Residual Quantization for Multimodal-Joint Interest Modeling in Music Recommendation
| null | null |
10.1145/3746252.3761579
| null |
cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In music recommendation systems, multimodal interest learning is pivotal, which allows the model to capture nuanced preferences, including textual elements such as lyrics and various musical attributes such as different instruments and melodies. Recently, methods that incorporate multimodal content features through semantic IDs have achieved promising results. However, existing methods suffer from two critical limitations: 1) intra-modal semantic degradation, where residual-based quantization processes gradually decouple discrete IDs from original content semantics, leading to semantic drift; and 2) inter-modal modeling gaps, where traditional fusion strategies either overlook modal-specific details or fail to capture cross-modal correlations, hindering comprehensive user interest modeling. To address these challenges, we propose a novel multimodal recommendation framework with two stages. In the first stage, our Progressive Semantic Residual Quantization (PSRQ) method generates modal-specific and modal-joint semantic IDs by explicitly preserving the prefix semantic feature. In the second stage, to model multimodal interest of users, a Multi-Codebook Cross-Attention (MCCA) network is designed to enable the model to simultaneously capture modal-specific interests and perceive cross-modal correlations. Extensive experiments on multiple real-world datasets demonstrate that our framework outperforms state-of-the-art baselines. This framework has been deployed on one of China's largest music streaming platforms, and online A/B tests confirm significant improvements in commercial metrics, underscoring its practical value for industrial-scale recommendation systems.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:16:57 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Shijia",
""
],
[
"Ouyang",
"Tianpei",
""
],
[
"Xiao",
"Qiang",
""
],
[
"Wang",
"Dongjing",
""
],
[
"Ren",
"Yintao",
""
],
[
"Xu",
"Songpei",
""
],
[
"Guo",
"Da",
""
],
[
"Luo",
"Chuanjiang",
""
]
] |
2508.20360
|
Charles Gong
|
Charles Gong
|
Three Generalizations of Erd\H{o}s Szekeres: $k$-Modal Subsequences
|
8 pages, 1 figure
| null | null | null |
math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
Erd\H{o}s and Szekeres showed that given a permutation $p$ of $[n]$, and the sequence defined by \newline $(p(1), p(2), \ldots, p(n))$, there exists either a decreasing or increasing subsequence, not necessarily contiguous, of length at least $\sqrt{n}$. Fan Chung considered subsequences that can have at most one change of direction, i.e. an increasing and then decreasing subsequence, or a decreasing and then increasing subsequence. She called these unimodal subsequences, and showed there exists a unimodal subsequence of length at least $\sqrt{3n}$, up to some constants \cite{chung}. She conjectured that a permutation of $n$ contains a $k$-modal (at most $k$ changes in direction) subsequence of length at least $\sqrt{(2k+1)n}$ up to some constants. Zijian Xu proved this conjecture in 2024 \cite{xu}, and we will provide another substantially different proof using "sophisticated labeling arguments" instead of "underlying poset structures behind k-modal subsequences." We also show that there exists an increasing first $k$-modal subsequence of length at least $\sqrt{2kn}$.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:17:39 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Gong",
"Charles",
""
]
] |
2508.20361
|
Tengteng Cui
|
Tengteng Cui, Chengtao Sheng, Bihao Su, and Zhi Zhou
|
Numerical Method for Space-Time Fractional Diffusion: A Stochastic Approach
| null | null | null | null |
math.NA cs.NA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we develop and analyze a stochastic algorithm for solving space-time fractional diffusion models, which are widely used to describe anomalous diffusion dynamics. These models pose substantial numerical challenges due to the memory effect of the time-fractional derivative and the nonlocal nature of the spatial fractional Laplacian and the, leading to significant computational costs and storage demands, particularly in high-dimensional settings. To overcome these difficulties, we propose a Monte Carlo method based on the Feynman--Kac formula for space-time fractional models. The novel algorithm combines the simulation of the monotone path of a stable subordinator in time with the ``walk-on-spheres'' method that efficiently simulates the stable Levy jumping process in space. We rigorously derive error bounds for the proposed scheme, explicitly expressed in terms of the number of simulation paths and the time step size. Numerical experiments confirm the theoretical error bounds and demonstrate the computational efficiency of the method, particularly in domains with complex geometries or high-dimensional spaces. Furthermore, both theoretical and numerical results emphasize the robustness of the proposed approach across a range of fractional orders, particularly for small fractional values, a capability often absent in traditional numerical methods.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:19:06 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Cui",
"Tengteng",
""
],
[
"Sheng",
"Chengtao",
""
],
[
"Su",
"Bihao",
""
],
[
"Zhou",
"Zhi",
""
]
] |
2508.20362
|
Yi Liao
|
Yi Liao, Hao-Lin Wang, Pengcheng Zhao
|
Three-dimensional trapping of circular Rydberg atoms by a superimposed vortex light beam
|
10 pages, 6 figures
| null | null | null |
physics.atom-ph hep-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We propose to trap circular Rydberg atoms (CRAs) by a ponderomotive potential well formed with a superimposed vortex light beam. We calculate analytically the ponderomotive potential energy for a Bessel vortex light beam. We work out a corrected version of the classical circular orbit approximation for a CRA which fits the exact result much better than the usual approximation. We reveal the three-dimensional characteristics of the potential well for some benchmark values of the CRA principal quantum number and beam parameters such as the frequency, the opening angle and topological charge of the vortex. We investigate how we can achieve similar trapping effects for different principal quantum numbers by varying beam parameters. The potential provides a lattice structure in the beam axis where one CRA could be trapped at each lattice site.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:19:11 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Liao",
"Yi",
""
],
[
"Wang",
"Hao-Lin",
""
],
[
"Zhao",
"Pengcheng",
""
]
] |
2508.20363
|
Jordan Strasser
|
Jordan Strasser
|
Tumorigenesis as a trauma response: the fragmentation of morphogenetic memory drives neoplastic dissociation
| null | null | null | null |
q-bio.CB q-bio.NC q-bio.PE q-bio.TO
|
http://creativecommons.org/licenses/by/4.0/
|
The mitigation of stress is a key challenge for all biological systems. Conditions of unresolvable stress have been associated with a diverse array of pathologies, from cancer to post-traumatic stress disorder (PTSD). Here, I unify insights from evolutionary and developmental biology with trauma psychology to present a novel framework for tumorigenesis which synthesizes stress-perception, tissue dysfunction, and the hallmarks of neoplastic growth. This view carries therapeutic implications, suggesting a reintegrative approach that seeks to return cancer cells to the homeostatic control of the surrounding tissue.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:21:43 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Strasser",
"Jordan",
""
]
] |
2508.20364
|
Yi-Huang Shen
|
Kuei-Nuan Lin, Yi-Huang Shen
|
Binomial edge rings associated to skew Ferrers diagrams
|
17 page. Comments are welcome
| null | null | null |
math.AC math.CO
|
http://creativecommons.org/licenses/by/4.0/
|
In this study, we investigate the binomial edge ring associated with the skew Ferrers diagram. By employing Sagbi basis theory, we construct a quadratic Gr\"{o}bner basis for its defining ideal. As an application, we prove that this ring is a Koszul, Cohen-Macaulay, normal domain. Moreover, we precisely determine its Krull dimension.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:25:47 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Lin",
"Kuei-Nuan",
""
],
[
"Shen",
"Yi-Huang",
""
]
] |
2508.20365
|
Naoki Kobayashi
|
Naoki Kobayashi and Ryosuke Sato and Ayumi Shinohara and Ryo Yoshinaka
|
Solvable Tuple Patterns and Their Applications to Program Verification
| null | null | null | null |
cs.PL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Despite the recent progress of automated program verification techniques, fully automated verification of programs manipulating recursive data structures remains a challenge. We introduce the notion of solvable tuple patterns (STPs) to express invariants between list-like recursive data structures. A distinguishing feature of STPs is that they can be efficiently inferred from only a small number of positive samples; no negative samples are required. An SMT solver that supports the sequence theory can be used to check that an inferred STP is indeed an inductive invariant. After presenting basic properties of STPs and an STP inference algorithm, we show how to incorporate the STP inference into a CHC (Constrained Horn Clauses) solver supporting list-like data structures, which serves as a uniform backend for automated program verification tools. A CHC solver incorporating the STP inference has won the ADT-LIN category of CHC-COMP 2025 by a big margin.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:30:02 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kobayashi",
"Naoki",
""
],
[
"Sato",
"Ryosuke",
""
],
[
"Shinohara",
"Ayumi",
""
],
[
"Yoshinaka",
"Ryo",
""
]
] |
2508.20366
|
Muye Liu
|
Muye Liu and Jun Xie
|
When Is Causal Inference Possible? A Statistical Test for Unmeasured Confounding
| null | null | null | null |
stat.ME
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper clarifies a fundamental difference between causal inference and traditional statistical inference by formalizing a mathematical distinction between their respective parameters. We connect two major approaches to causal inference, the potential outcomes framework and causal structure graphs, which are typically studied separately. While the unconfoundedness assumption in the potential outcomes framework cannot be assessed from an observational dataset alone, causal structure graphs help explain when causal effects are identifiable through graphical models. We propose a statistical test to assess the unconfoundedness assumption, equivalent to the absence of unmeasured confounding, by comparing two datasets: a randomized controlled trial and an observational study. The test controls the Type I error probability, and we analyze its power under linear models. Our approach provides a practical method to evaluate when real-world data are suitable for causal inference.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:30:29 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Liu",
"Muye",
""
],
[
"Xie",
"Jun",
""
]
] |
2508.20367
|
Luke Bhan
|
Luke Bhan, Miroslav Krstic, and Yuanyuan Shi
|
Delay-adaptive Control of Nonlinear Systems with Approximate Neural Operator Predictors
|
9 pages, 1 Figure
| null | null | null |
eess.SY cs.LG cs.SY math.DS
|
http://creativecommons.org/licenses/by/4.0/
|
In this work, we propose a rigorous method for implementing predictor feedback controllers in nonlinear systems with unknown and arbitrarily long actuator delays. To address the analytically intractable nature of the predictor, we approximate it using a learned neural operator mapping. This mapping is trained once, offline, and then deployed online, leveraging the fast inference capabilities of neural networks. We provide a theoretical stability analysis based on the universal approximation theorem of neural operators and the transport partial differential equation (PDE) representation of the delay. We then prove, via a Lyapunov-Krasovskii functional, semi-global practical convergence of the dynamical system dependent on the approximation error of the predictor and delay bounds. Finally, we validate our theoretical results using a biological activator/repressor system, demonstrating speedups of 15 times compared to traditional numerical methods.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:30:53 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Bhan",
"Luke",
""
],
[
"Krstic",
"Miroslav",
""
],
[
"Shi",
"Yuanyuan",
""
]
] |
2508.20368
|
Lang Mei
|
Lang Mei, Zhihan Yang, Chong Chen
|
AI-SearchPlanner: Modular Agentic Search via Pareto-Optimal Multi-Objective Reinforcement Learning
| null | null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Recent studies have explored integrating Large Language Models (LLMs) with search engines to leverage both the LLMs' internal pre-trained knowledge and external information. Specially, reinforcement learning (RL) has emerged as a promising paradigm for enhancing LLM reasoning through multi-turn interactions with search engines. However, existing RL-based search agents rely on a single LLM to handle both search planning and question-answering (QA) tasks in an end-to-end manner, which limits their ability to optimize both capabilities simultaneously. In practice, sophisticated AI search systems often employ a large, frozen LLM (e.g., GPT-4, DeepSeek-R1) to ensure high-quality QA. Thus, a more effective and efficient approach is to utilize a small, trainable LLM dedicated to search planning. In this paper, we propose \textbf{AI-SearchPlanner}, a novel reinforcement learning framework designed to enhance the performance of frozen QA models by focusing on search planning. Specifically, our approach introduces three key innovations: 1) Decoupling the Architecture of the Search Planner and Generator, 2) Dual-Reward Alignment for Search Planning, and 3) Pareto Optimization of Planning Utility and Cost, to achieve the objectives. Extensive experiments on real-world datasets demonstrate that AI SearchPlanner outperforms existing RL-based search agents in both effectiveness and efficiency, while exhibiting strong generalization capabilities across diverse frozen QA models and data domains.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:31:17 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Mei",
"Lang",
""
],
[
"Yang",
"Zhihan",
""
],
[
"Chen",
"Chong",
""
]
] |
2508.20369
|
Haiquan Lu
|
Haiquan Lu, Hongqi Min, Yong Zeng, Shaodan Ma
|
Flexible XL-MIMO via Array Configuration Codebook: Codebook Design and Array Configuration Training
| null | null | null | null |
cs.IT math.IT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
XL-MIMO emerges as a promising technology to achieve unprecedented enhancements in spectral efficiency and spatial resolution, via orders-of-magnitude increase in the antenna array size. However, the practical issues of high hardware cost and power consumption pose great challenges towards the cost-effective implementation of XL-MIMO. To address such challenges, this paper proposes a novel concept called array configuration codebook (ACC), which enables flexible XL-MIMO cost-effectively and improves the system performance compared with conventional antenna selection (AS) schemes with limited number of RF chains. Specifically, ACC refers to a set of pre-designed array configuration codewords, where each codeword specifies the positions of activated antenna pixels. Then, flexible XL-MIMO architecture can be enabled via dynamical pixel activation based on the designed ACC, without having to exhaustively try all possible combinations of the antenna pixels activations. As an illustration, we give a specific codebook design, encompassing the classic compact array (CA), uniform sparse array (USA), modular array (MoA), nested array (NA), and co-prime array (CPA), and each codeword is specified by one array configuration parameter. With the designed ACC, array configuration training is considered for multi-UE communication to maximize the sum rate. To reduce the training overhead of exhaustive scanning, a two-stage scanning scheme is proposed, including the array- and pixel-level scanning. For comparison, the greedy AS scheme is proposed, where the resulting incremental SINR expression by activating antenna pixel sequentially is derived in closed-form. Subsequently, array configuration training is extended to the wireless localization scenario. Simulation results demonstrate the effectiveness of codeword optimization for scenarios of multi-UE communication and wireless localization.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:34:13 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Lu",
"Haiquan",
""
],
[
"Min",
"Hongqi",
""
],
[
"Zeng",
"Yong",
""
],
[
"Ma",
"Shaodan",
""
]
] |
2508.20370
|
Lingzhe Zhang
|
Lingzhe Zhang, Tong Jia, Kangjin Wang, Weijie Hong, Chiming Duan, Minghua He, and Ying Li
|
Adaptive Root Cause Localization for Microservice Systems with Multi-Agent Recursion-of-Thought
| null | null | null | null |
cs.SE cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
As contemporary microservice systems become increasingly popular and complex-often comprising hundreds or even thousands of fine-grained, interdependent subsystems-they are facing more frequent failures. Ensuring system reliability thus demands accurate root cause localization. While traces and metrics have proven to be effective data sources for this task, existing methods either heavily rely on pre-defined schemas, which struggle to adapt to evolving operational contexts, or lack interpretability in their reasoning process, thereby leaving Site Reliability Engineers (SREs) confused. In this paper, we conduct a comprehensive study on how SREs localize the root cause of failures, drawing insights from multiple professional SREs across different organizations. Our investigation reveals that human root cause analysis exhibits three key characteristics: recursiveness, multi-dimensional expansion, and cross-modal reasoning. Motivated by these findings, we introduce RCLAgent, an adaptive root cause localization method for microservice systems that leverages a multi-agent recursion-of-thought framework. RCLAgent employs a novel recursion-of-thought strategy to guide the LLM's reasoning process, effectively integrating data from multiple agents and tool-assisted analysis to accurately pinpoint the root cause. Experimental evaluations on various public datasets demonstrate that RCLAgent achieves superior performance by localizing the root cause using only a single request-outperforming state-of-the-art methods that depend on aggregating multiple requests. These results underscore the effectiveness of RCLAgent in enhancing the efficiency and precision of root cause localization in complex microservice environments.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:34:19 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhang",
"Lingzhe",
""
],
[
"Jia",
"Tong",
""
],
[
"Wang",
"Kangjin",
""
],
[
"Hong",
"Weijie",
""
],
[
"Duan",
"Chiming",
""
],
[
"He",
"Minghua",
""
],
[
"Li",
"Ying",
""
]
] |
2508.20371
|
Sopam Dasgupta
|
Sopam Dasgupta, Sadaf MD Halim, Joaqu\'in Arias, Elmer Salazar, Gopal Gupta
|
P2C: Path to Counterfactuals
| null | null | null | null |
cs.AI cs.LG cs.LO
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Machine-learning models are increasingly driving decisions in high-stakes settings, such as finance, law, and hiring, thus, highlighting the need for transparency. However, the key challenge is to balance transparency -- clarifying `why' a decision was made -- with recourse: providing actionable steps on `how' to achieve a favourable outcome from an unfavourable outcome. Counterfactual explanations reveal `why' an undesired outcome occurred and `how' to reverse it through targeted feature changes (interventions).
Current counterfactual approaches have limitations: 1) they often ignore causal dependencies between features, and 2) they typically assume all interventions can happen simultaneously, an unrealistic assumption in practical scenarios where actions are typically taken in a sequence. As a result, these counterfactuals are often not achievable in the real world.
We present P2C (Path-to-Counterfactuals), a model-agnostic framework that produces a plan (ordered sequence of actions) converting an unfavourable outcome to a causally consistent favourable outcome. P2C addresses both limitations by 1) Explicitly modelling causal relationships between features and 2) Ensuring that each intermediate state in the plan is feasible and causally valid. P2C uses the goal-directed Answer Set Programming system s(CASP) to generate the plan accounting for feature changes that happen automatically due to causal dependencies. Furthermore, P2C refines cost (effort) computation by only counting changes actively made by the user, resulting in realistic cost estimates. Finally, P2C highlights how its causal planner outperforms standard planners, which lack causal knowledge and thus can generate illegal actions.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:36:02 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Dasgupta",
"Sopam",
""
],
[
"Halim",
"Sadaf MD",
""
],
[
"Arias",
"Joaquín",
""
],
[
"Salazar",
"Elmer",
""
],
[
"Gupta",
"Gopal",
""
]
] |
2508.20372
|
Ryutaro Matsudo
|
Ryuichiro Kitano, Ryutaro Matsudo and Lukas Treuer
|
On cusps in the $\eta'$ potential
|
27 pages, 1 figure
| null | null |
KEK-TH-2727, YITP-25-123
|
hep-th hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The large $N$ analysis of QCD states that the potential for the $\eta'$ meson develops cusps at $\eta' = \pi / N_f$, $3 \pi /N_f$, $\cdots$, with $N_f$ the number of flavors. Furthermore, the recent discussion of generalized anomalies tells us that even for finite $N$ there should be cusps if $N$ and $N_f$ are not coprime, as one can show that the domain wall configuration of $\eta'$ should support a Chern-Simons theory on it, i.e., domains are not smoothly connected. On the other hand, there is a supporting argument for instanton-like, smooth potentials of $\eta'$ from the analyses of softly-broken supersymmetric QCD for $N_f= N-1$, $N$, and $N+1$. We argue that the analysis of the $N_f = N$ case should be subject to the above anomaly argument, and thus there should be a cusp; while the $N_f = N \pm 1$ cases are consistent, as $N_f$ and $N$ are coprime. We discuss how this cuspy/smooth transition can be understood. For $N_f< N$, we find that the number of branches of the $\eta'$ potential is $\operatorname{gcd}(N,N_f)$, which is the minimum number allowed by the anomaly. We also discuss the condition for s-confinement in QCD-like theories, and find that in general the anomaly matching of the $\theta$ periodicity indicates that s-confinement can only be possible when $N_f$ and $N$ are coprime. The s-confinement in supersymmetric QCD at $N_f = N+1$ is a famous example, and the argument generalizes for any number of fermions in the adjoint representation.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:37:56 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kitano",
"Ryuichiro",
""
],
[
"Matsudo",
"Ryutaro",
""
],
[
"Treuer",
"Lukas",
""
]
] |
2508.20373
|
Yuyao Wang
|
Yuyao Wang, Bowen Liu, Jianheng Tang, Nuo Chen, Yuhan Li, Qifan Zhang, Jia Li
|
Graph-R1: Unleashing LLM Reasoning with NP-Hard Graph Problems
| null | null | null | null |
cs.CL cs.AI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Reasoning Large Language Models (RLLMs) have recently achieved remarkable progress on complex reasoning tasks, largely enabled by their long chain-of-thought (Long CoT) capabilities. However, developing these Long CoT behaviors relies heavily on post-training with high-quality datasets, which are typically costly and human-curated (e.g., mathematics and code), leaving scalable alternatives unexplored. In this work, we introduce NP-hard (NPH) graph problems as a novel synthetic training corpus, as they inherently require deep reasoning, extensive exploration, and reflective strategies, which are core characteristics of Long CoT reasoning. Building on this insight, we develop a two-stage post-training framework: (i) Long CoT Supervised Fine-Tuning (SFT) on rejection-sampled NPH graph instances, which substantially enhances reasoning depth, and (ii) Reinforcement Learning (RL) with a fine-grained reward design, which sharpens reasoning efficiency. Our flagship model, Graph-R1-7B, demonstrates strong generalization across mathematics, coding, STEM, and logic, and surpasses QwQ-32B on NPH graph problems in both accuracy and reasoning efficiency. These results position NPH graph problems as an effective and scalable resource for advancing Long CoT reasoning in LLMs, opening a new frontier for LLM post-training. Our implementation is available at https://github.com/Graph-Reasoner/Graph-R1, with models and datasets hosted in our Hugging Face collection HKUST-DSAIL/Graph-R1.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:40:27 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Yuyao",
""
],
[
"Liu",
"Bowen",
""
],
[
"Tang",
"Jianheng",
""
],
[
"Chen",
"Nuo",
""
],
[
"Li",
"Yuhan",
""
],
[
"Zhang",
"Qifan",
""
],
[
"Li",
"Jia",
""
]
] |
2508.20374
|
Yebowen Hu
|
Simin Ma, Shujian Liu, Jun Tan, Yebowen Hu, Song Wang, Sathish Reddy Indurthi, Sanqiang Zhao, Liwei Wu, Jianbing Han, Kaiqiang Song
|
TCIA: A Task-Centric Instruction Augmentation Method for Instruction Finetuning
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Diverse instruction data is vital for effective instruction tuning of large language models, as it enables the model to generalize across different types of inputs . Building such diversified instruction dataset is an essential step in this process. Existing approaches often leverage large language models to automatically explore and generate diverse instructions, ensuring both data diversity and quality. However, they tend to overlook an important factor in real-world applications: on-task relevance. In practice, only a few real-world applications require a truly general-purpose model; most benefit from task-specific knowledge tailored to their particular use case. Therefore, it is vital to develop instruction augmentation methods that not only maintain diversity but are also optimized for specific, real-world scenarios.
We thus introduce Task Centric Instruction Augmentation (TCIA), a framework that systematically expands instructions while preserving both diversity and task alignment. By representing instructions in a discrete query-constraints space, TCIA creates a rich set of task-relevant instructions and enables models to generalize to these task-specific instructions without sacrificing overall performance. Experiments show that TCIA improves open-source LLMs' performance by an average of 8.7% across four real-world, task-specific applications, and in some cases outperforming leading closed-source models. These improvements do not compromise general instruction-following ability, making TCIA a scalable and efficient solution for adapting LLMs to real-world, task-focused applications.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:42:10 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Ma",
"Simin",
""
],
[
"Liu",
"Shujian",
""
],
[
"Tan",
"Jun",
""
],
[
"Hu",
"Yebowen",
""
],
[
"Wang",
"Song",
""
],
[
"Indurthi",
"Sathish Reddy",
""
],
[
"Zhao",
"Sanqiang",
""
],
[
"Wu",
"Liwei",
""
],
[
"Han",
"Jianbing",
""
],
[
"Song",
"Kaiqiang",
""
]
] |
2508.20375
|
Guanyu Xu
|
Guanyu Xu, Zhiwei Hao, Li Shen, Yong Luo, Fuhui Sun, Xiaoyan Wang, Han Hu, Yonggang Wen
|
CoFormer: Collaborating with Heterogeneous Edge Devices for Scalable Transformer Inference
|
Accepted by IEEE Transactions on Computers
| null | null | null |
cs.DC cs.LG cs.PF
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The impressive performance of transformer models has sparked the deployment of intelligent applications on resource-constrained edge devices. However, ensuring high-quality service for real-time edge systems is a significant challenge due to the considerable computational demands and resource requirements of these models. Existing strategies typically either offload transformer computations to other devices or directly deploy compressed models on individual edge devices. These strategies, however, result in either considerable communication overhead or suboptimal trade-offs between accuracy and efficiency. To tackle these challenges, we propose a collaborative inference system for general transformer models, termed CoFormer. The central idea behind CoFormer is to exploit the divisibility and integrability of transformer. An off-the-shelf large transformer can be decomposed into multiple smaller models for distributed inference, and their intermediate results are aggregated to generate the final output. We formulate an optimization problem to minimize both inference latency and accuracy degradation under heterogeneous hardware constraints. DeBo algorithm is proposed to first solve the optimization problem to derive the decomposition policy, and then progressively calibrate decomposed models to restore performance. We demonstrate the capability to support a wide range of transformer models on heterogeneous edge devices, achieving up to 3.1$\times$ inference speedup with large transformer models. Notably, CoFormer enables the efficient inference of GPT2-XL with 1.6 billion parameters on edge devices, reducing memory requirements by 76.3\%. CoFormer can also reduce energy consumption by approximately 40\% while maintaining satisfactory inference performance.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:50:12 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Xu",
"Guanyu",
""
],
[
"Hao",
"Zhiwei",
""
],
[
"Shen",
"Li",
""
],
[
"Luo",
"Yong",
""
],
[
"Sun",
"Fuhui",
""
],
[
"Wang",
"Xiaoyan",
""
],
[
"Hu",
"Han",
""
],
[
"Wen",
"Yonggang",
""
]
] |
2508.20376
|
Mang Cao
|
Mang Cao, Sanping Zhou, Yizhe Li, Ye Deng, Wenli Huang, and Le Wang
|
Enhancing Mamba Decoder with Bidirectional Interaction in Multi-Task Dense Prediction
|
Codes are available online: \url{https://github.com/mmm-cc/BIM\_for\_MTL}
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Sufficient cross-task interaction is crucial for success in multi-task dense prediction. However, sufficient interaction often results in high computational complexity, forcing existing methods to face the trade-off between interaction completeness and computational efficiency. To address this limitation, this work proposes a Bidirectional Interaction Mamba (BIM), which incorporates novel scanning mechanisms to adapt the Mamba modeling approach for multi-task dense prediction. On the one hand, we introduce a novel Bidirectional Interaction Scan (BI-Scan) mechanism, which constructs task-specific representations as bidirectional sequences during interaction. By integrating task-first and position-first scanning modes within a unified linear complexity architecture, BI-Scan efficiently preserves critical cross-task information. On the other hand, we employ a Multi-Scale Scan~(MS-Scan) mechanism to achieve multi-granularity scene modeling. This design not only meets the diverse granularity requirements of various tasks but also enhances nuanced cross-task feature interactions. Extensive experiments on two challenging benchmarks, \emph{i.e.}, NYUD-V2 and PASCAL-Context, show the superiority of our BIM vs its state-of-the-art competitors.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:50:19 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Cao",
"Mang",
""
],
[
"Zhou",
"Sanping",
""
],
[
"Li",
"Yizhe",
""
],
[
"Deng",
"Ye",
""
],
[
"Huang",
"Wenli",
""
],
[
"Wang",
"Le",
""
]
] |
2508.20377
|
Zhaoming Wang
|
Zhao-Wei Wang, Hong-Yang Ma, Yun-An Yan, Lian-Ao Wu and Zhao-Ming Wang
|
Machine-Learning-Assisted Pulse Design for State Preparation in a Noisy Environment
| null | null | null | null |
quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
High-precision quantum control is essential for quantum computing and quantum information processing. However, its practical implementation is challenged by environmental noise, which affects the stability and accuracy of quantum systems. In this paper, using machine learning techniques we propose a quantum control approach that incorporates environmental factors into the design of control schemes, improving the control fidelity in noisy environments. Specifically, we investigate arbitrary quantum state preparation in a two-level system coupled to a bosonic bath. We use both Deep Reinforcement Learning (DRL) and Supervised Learning (SL) algorithms to design specific control pulses that mitigate the noise. These two neural network (NN) based algorithm both have the advantage that the well trained NN can output the optimal pulse sequence for any environmental parameters. Comparing the performance of these two algorithms, our results show that DRL is more effective in low-noise environments due to its strong optimization capabilities, while SL provides greater stability and performs better in high-noise conditions. These findings highlight the potential of machine learning techniques to enhance the quantum control fidelity in practical applications.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:53:43 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Zhao-Wei",
""
],
[
"Ma",
"Hong-Yang",
""
],
[
"Yan",
"Yun-An",
""
],
[
"Wu",
"Lian-Ao",
""
],
[
"Wang",
"Zhao-Ming",
""
]
] |
2508.20378
|
William Detmold
|
William Detmold, Rohan Kanchana, and Cagin Yunus
|
Empirical investigation of nuclear correlation function distributions in lattice QCD
| null | null | null |
MIT-CTP-5903
|
hep-lat
|
http://creativecommons.org/licenses/by/4.0/
|
Two-point correlation functions of systems with baryon number $B \in \{1,2,3,4\}$ are investigated using lattice Quantum Chromodynamics (QCD). In particular, the empirical distributions of importance-sampling Monte-Carlo samples of these correlation functions are examined as a function of the spacetime separation between the two points and the baryon number. While the exact forms of these distributions are not known for QCD, recent work has determined asymptotic expressions for analogous correlation function distributions in simpler theories such as scalar field theory and the disordered phase of the $O(N)$ model. The theoretical O(N) model distributions are found to provide an accurate description of the empirical QCD distributions at zero momentum over a wide range of temporal separations for each baryon number when assessed with a range of different statistical tests. In particular, the behaviour of the baryon number $B$ QCD correlation function at large temporal separation is well-reproduced by the $O(N \sim 2/B)$ model distribution.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 02:56:46 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Detmold",
"William",
""
],
[
"Kanchana",
"Rohan",
""
],
[
"Yunus",
"Cagin",
""
]
] |
2508.20379
|
Hyeonyu Kim
|
Hyeonyu Kim, Seokhoon Jeong, Seonghee Han, Chanhyuk Choi, Taehwan Kim
|
Audio-Guided Visual Editing with Complex Multi-Modal Prompts
|
Accepted to BMVC 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Visual editing with diffusion models has made significant progress but often struggles with complex scenarios that textual guidance alone could not adequately describe, highlighting the need for additional non-text editing prompts. In this work, we introduce a novel audio-guided visual editing framework that can handle complex editing tasks with multiple text and audio prompts without requiring additional training. Existing audio-guided visual editing methods often necessitate training on specific datasets to align audio with text, limiting their generalization to real-world situations. We leverage a pre-trained multi-modal encoder with strong zero-shot capabilities and integrate diverse audio into visual editing tasks, by alleviating the discrepancy between the audio encoder space and the diffusion model's prompt encoder space. Additionally, we propose a novel approach to handle complex scenarios with multiple and multi-modal editing prompts through our separate noise branching and adaptive patch selection. Our comprehensive experiments on diverse editing tasks demonstrate that our framework excels in handling complicated editing scenarios by incorporating rich information from audio, where text-only approaches fail.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:00:30 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kim",
"Hyeonyu",
""
],
[
"Jeong",
"Seokhoon",
""
],
[
"Han",
"Seonghee",
""
],
[
"Choi",
"Chanhyuk",
""
],
[
"Kim",
"Taehwan",
""
]
] |
2508.20380
|
Eugene Kiang Jin Tan
|
Eugene Tan, Michael Small, Shannon D. Algar
|
Do triangles matter? Replicating hypergraph disease dynamics with lower-order interactions
| null | null | null | null |
math.DS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Disease spreading models such as the ubiquitous SIS compartmental model and its numerous variants are widely used to understand and predict the behaviour of a given epidemic or information diffusion process. A common approach to imbue more realism to the spreading process is to constrain simulations to a network structure, where connected nodes update their disease state based on pairwise interactions along the edges of their local neighbourhood. Simplicial contagion models (SCM) extend this to hypergraphs such that groups of three nodes are able to interact and propagate the disease along higher-order hyperedges (triangles). Though more flexible, it is not clear the extent to which the inclusion of these higher-order interactions result in dynamics that are characteristically different to those attained from simpler pairwise interactions. Here, we propose an agent-based model that unifies the classical SIS/SIR compartmental model and SCM, and extends it to allow for interactions along hyperedges of arbitrary order. Using this model, we demonstrate how the steady-state dynamics of pairwise interactions can be made to replicate those of simulations that include higher-order topologies by linearly scaling disease parameters based on a proposed measure network activity. By allowing disease parameters to dynamically vary over time, lower-order pairwise interactions can be made to closely replicate both the transient and steady-state dynamics of higher-order simulations.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:03:04 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Tan",
"Eugene",
""
],
[
"Small",
"Michael",
""
],
[
"Algar",
"Shannon D.",
""
]
] |
2508.20381
|
Luong Tran
|
Luong Tran, Thieu Vo, Anh Nguyen, Sang Dinh, Van Nguyen
|
More Reliable Pseudo-labels, Better Performance: A Generalized Approach to Single Positive Multi-label Learning
|
ICCV 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Multi-label learning is a challenging computer vision task that requires assigning multiple categories to each image. However, fully annotating large-scale datasets is often impractical due to high costs and effort, motivating the study of learning from partially annotated data. In the extreme case of Single Positive Multi-Label Learning (SPML), each image is provided with only one positive label, while all other labels remain unannotated. Traditional SPML methods that treat missing labels as unknown or negative tend to yield inaccuracies and false negatives, and integrating various pseudo-labeling strategies can introduce additional noise. To address these challenges, we propose the Generalized Pseudo-Label Robust Loss (GPR Loss), a novel loss function that effectively learns from diverse pseudo-labels while mitigating noise. Complementing this, we introduce a simple yet effective Dynamic Augmented Multi-focus Pseudo-labeling (DAMP) technique. Together, these contributions form the Adaptive and Efficient Vision-Language Pseudo-Labeling (AEVLP) framework. Extensive experiments on four benchmark datasets demonstrate that our framework significantly advances multi-label classification, achieving state-of-the-art results.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:07:57 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Tran",
"Luong",
""
],
[
"Vo",
"Thieu",
""
],
[
"Nguyen",
"Anh",
""
],
[
"Dinh",
"Sang",
""
],
[
"Nguyen",
"Van",
""
]
] |
2508.20382
|
Jian Zhang
|
Naihuan Jing, Yinlong Liu, Jian Zhang
|
Immanant inequalities and weight spaces
| null | null | null | null |
math.RT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We first obtain a trace formula for immanants of generalized principal submatrix of any complex matrix based on any weight space for finite dimensional representations of the general linear group. Our trace formula contains Kostant's famous formula for immanants on $0$-weight spaces as special case. We then present a criterion for non-vanishing immanants for any generalized principal submatrix of positive definite Hermitian or nonsingular totally nonnegative matrices, which strengthened the well-known results of Schur and Stembridge. Furthermore, we present an inequality that contains Kostant, Schur and Stembridge's famous inequalities as special cases.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:09:09 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Jing",
"Naihuan",
""
],
[
"Liu",
"Yinlong",
""
],
[
"Zhang",
"Jian",
""
]
] |
2508.20383
|
Prakash Shukla
|
Prakash Shukla and Paul Parsons
|
Identifying Framing Practices in Visualization Design Through Practitioner Reflections
|
Accepted for publication in the IEEEVIS Workshop on Visualization for Communication (VisComm 25)
| null | null | null |
cs.HC
|
http://creativecommons.org/licenses/by/4.0/
|
Framing -- how designers define and reinterpret problems, shape narratives, and guide audience understanding -- is central to design practice. Yet in visualization research, framing has been examined mostly through its rhetorical and perceptual effects on audiences, leaving its role in the design process underexplored. This study addresses that gap by analyzing publicly available podcasts and book chapters in which over 80 professional visualization designers reflect on their work. We find that framing is a pervasive, iterative activity, evident in scoping problems, interpreting data, aligning with stakeholder goals, and shaping narrative direction. Our analysis identifies the conditions that trigger reframing and the strategies practitioners use to navigate uncertainty and guide design. These findings position framing as a core dimension of visualization practice and underscore the need for research and education to support the interpretive and strategic judgment that practitioners exercise throughout the design process.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:14:39 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Shukla",
"Prakash",
""
],
[
"Parsons",
"Paul",
""
]
] |
2508.20384
|
Yongfu Zhu
|
Yongfu Zhu, Lin Sun, Guangxiang Zhao, Weihong Lin, Xiangzheng Zhang
|
Uncertainty Under the Curve: A Sequence-Level Entropy Area Metric for Reasoning LLM
|
Under review for AAAI 2026
| null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this work, we introduce Entropy Area Score (EAS), a simple yet effective metric to quantify uncertainty in the answer generation process of reasoning large language models (LLMs). EAS requires neither external models nor repeated sampling, it integrates token-level predictive entropy from the model itself to capture the evolution of uncertainty during generation. Empirical results show that EAS is strongly correlated with answer entropy across models and datasets. In training data selection, EAS identifies high-potential samples and consistently outperforms Pass Rate filtering under equal sample budgets, improving student model accuracy on math benchmarks. EAS is both efficient and interpretable, offering a practical tool for uncertainty modeling and data quality assessment in LLM training.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:16:15 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhu",
"Yongfu",
""
],
[
"Sun",
"Lin",
""
],
[
"Zhao",
"Guangxiang",
""
],
[
"Lin",
"Weihong",
""
],
[
"Zhang",
"Xiangzheng",
""
]
] |
2508.20385
|
Jivnesh Sandhan
|
Jivnesh Sandhan, Fei Cheng, Tushar Sandhan and Yugo Murawaki
|
CAPE: Context-Aware Personality Evaluation Framework for Large Language Models
|
Accepted at EMNLP25 (Findings)
| null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Psychometric tests, traditionally used to assess humans, are now being applied to Large Language Models (LLMs) to evaluate their behavioral traits. However, existing studies follow a context-free approach, answering each question in isolation to avoid contextual influence. We term this the Disney World test, an artificial setting that ignores real-world applications, where conversational history shapes responses. To bridge this gap, we propose the first Context-Aware Personality Evaluation (CAPE) framework for LLMs, incorporating prior conversational interactions. To thoroughly analyze the influence of context, we introduce novel metrics to quantify the consistency of LLM responses, a fundamental trait in human behavior.
Our exhaustive experiments on 7 LLMs reveal that conversational history enhances response consistency via in-context learning but also induces personality shifts, with GPT-3.5-Turbo and GPT-4-Turbo exhibiting extreme deviations. While GPT models are robust to question ordering, Gemini-1.5-Flash and Llama-8B display significant sensitivity. Moreover, GPT models response stem from their intrinsic personality traits as well as prior interactions, whereas Gemini-1.5-Flash and Llama--8B heavily depend on prior interactions. Finally, applying our framework to Role Playing Agents (RPAs) shows context-dependent personality shifts improve response consistency and better align with human judgments. Our code and datasets are publicly available at: https://github.com/jivnesh/CAPE
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:17:47 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Sandhan",
"Jivnesh",
""
],
[
"Cheng",
"Fei",
""
],
[
"Sandhan",
"Tushar",
""
],
[
"Murawaki",
"Yugo",
""
]
] |
2508.20386
|
Ao Chen
|
Ao Chen, Juntai Shen, Chun Wang, Yang Huang
|
Revisiting the Radial Metallicity Gradient-Age Relation in the Milky Way's Thin and Thick Disks
|
12 pages, 9 figures, to be submitted to ApJ
| null | null | null |
astro-ph.GA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Galactic disks typically exhibit a negative radial metallicity gradient, indicating faster enrichment in the inner regions. Recent studies report that this gradient becomes flatter with increasing stellar age in the Milky Way's (MW) thin disk, while the thick disk exhibits a mildly positive gradient across all ages. In this work, we revisit the metallicity gradient-age relation (MGAR) in both the thin and thick disks of the MW. We use spectroscopic data from LAMOST DR8 and stellar ages calibrated with asteroseismology. Our results show a steadily flattening MGAR in the thin disk and confirm a positive gradient $\sim0.013\,\mathrm{dex\,kpc^{-1}}$ in the thick disk. The flattening in the thin disk may be caused by large-scale radial migration induced by transient spiral arms, or by a time-dependent steepening of the interstellar medium (ISM) metallicity gradient as suggested by recent FIRE2 simulations. The positive gradient in the thick disk may reflect early enrichment of the outer regions by strong feedback or starburst-driven outflows in a turbulent, gas-rich proto-disk. These findings suggest distinct chemodynamical evolution paths for the MW's thin and thick disks and provide valuable constraints for future models of Galactic chemical evolution.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:20:02 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Chen",
"Ao",
""
],
[
"Shen",
"Juntai",
""
],
[
"Wang",
"Chun",
""
],
[
"Huang",
"Yang",
""
]
] |
2508.20387
|
Ku-Yu Fan
|
Ku-Yu Fan
|
A map between arborifications of multiple zeta values
|
18 pages
| null | null | null |
math.NT math.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Arborified multiple zeta values are a generalization of multiple zeta values associated with rooted trees. There are two types of decorated rooted trees, corresponding respectively to the series and the integral expressions. Manchon introduces the contracting arborification (resp. the simple arborification), which is maps from the BCK Hopf algebras of the decorated rooted trees corresponding to the series expression (resp. the integral expression) to the non-commutative polynomial algebras of the set $\mathbb{N}$ (resp. the set $\{0,1\}$). There is a natural map between the two non-commutative polynomial algebras. Manchon posed the question of finding a natural map between the two BCK Hopf algebras that would make the diagram commutative. In this paper, we consider planar rooted trees and use a recursive method to construct such a map between the two BCK Hopf algebras, making the diagram commutative.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:21:48 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Fan",
"Ku-Yu",
""
]
] |
2508.20388
|
Keyu Zhang
|
Zongxia Liang, Xiang Yu, Keyu Zhang
|
Mean Field Game with Reflected Jump Diffusion Dynamics: A Linear Programming Approach
| null | null | null | null |
math.OC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper develops a linear programming approach for mean field games with reflected jump-diffusion state dynamices. The jump component is driven by a Poisson process with time-varying intensity, and the multi-dimensional reflections enforce the state process to stay in a bounded domain. We first prove the equivalence between the mean field equilibria in the linear programming formulation and those in the weak relaxed control formulation under some measurability conditions on model coefficients. Building upon the characterization of the occupation measure in the equivalence result, we further establish the existence of linear programming mean field equilibria under mild continuity conditions on model coefficients.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:24:28 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Liang",
"Zongxia",
""
],
[
"Yu",
"Xiang",
""
],
[
"Zhang",
"Keyu",
""
]
] |
2508.20389
|
Sho Sugama
|
Ryuichiro Kitano, Joe Sato, Sho Sugama
|
Unitarity test of lepton mixing via energy dependence of neutrino oscillation
|
27 pages, 9 figures
| null | null |
YITP-25-134
|
hep-ph hep-ex
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We study the method to test the unitarity of the lepton mixing matrix by using only the long baseline neutrino oscillation experiments, such as the combination of the T2HK experiment and the one with the $\nu_e$ beam from a future neutrino factory at J-PARC. Without a specific parametrization, one can directly extract the elements of the lepton mixing matrix by observing the energy dependence of the oscillation probabilities. A non-trivial test of the unitarity under the three-generation assumption can thus be made possible by examining the orthogonality in a similar manner to the unitarity triangle in the quark sector. As the first trial, we perform the analysis based on the simplified situation where the matter effects in the neutrino oscillation can be neglected. Under this simplified analysis, we demonstrate the observation of the unitarity violation in the $3\times3$ part of the lepton mixing matrix for a parameter set in the four-generation model. The statistically most significant measurement can be provided by the energy dependences of the combination of the CP conjugate modes, $\nu_\mu \to \nu_e$ and $\bar \nu_\mu \to \bar \nu_e$, at T2HK and, independently, by the T conjugate modes, $\nu_\mu \to \nu_e$ and $\nu_e \to \nu_\mu$, with the latter measured at the neutrino factory experiments.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:27:08 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kitano",
"Ryuichiro",
""
],
[
"Sato",
"Joe",
""
],
[
"Sugama",
"Sho",
""
]
] |
2508.20390
|
Andreas Kirchner
|
Andreas Kirchner and Steffen A. Bass
|
Signals for fluctuating constituent numbers in small systems
|
8 pages, 10 figures
| null | null | null |
hep-ph nucl-th
|
http://creativecommons.org/licenses/by/4.0/
|
We propose an extension of the initial condition model TRENTo for sampling the number of partons inside the nucleons that participate in a heavy-ion collision. This sampling method is based on parton distribution functions (PDFs) and therefore has a natural dependence on the momentum transferred in the collision and the scale being probed during the collision. We examine the resulting distributions and their dependence on the momentum transfer. Additionally, we explore the sensitivity of different observables on the number of partons using the TRENTo framework and the estimators available therein for final-state observables.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:30:48 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kirchner",
"Andreas",
""
],
[
"Bass",
"Steffen A.",
""
]
] |
2508.20391
|
Benjamin Alterman
|
B. L. Alterman, R. D'Amicis
|
On the Regulation of the Solar Wind Helium Abundance by the Hydrogen Compressibility
| null | null | null | null |
physics.space-ph astro-ph.SR physics.plasm-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Traditionally, fast solar wind is considered to originate in solar source regions that are continuously open to the heliosphere and slow wind originates in regions that are intermittently open to it. In fast wind, the gradient of the solar wind helium abundance ($A_\mathrm{He}$) with increasing solar wind speed ($v_\mathrm{sw}$) is $\sim0$ and $A_\mathrm{He}$ is fixed at $\sim50\%$ of the photospheric value. In slow wind, this gradient is large, $A_\mathrm{He}$ is highly variable, and it doesn't exceed this $\sim50\%$ value. Although the normalized cross helicity in fast wind typically approaches 1, this is not universally true and Alterman & D'Amicis (2025) show that $\nabla_{v_\mathrm{sw}} \! A_\mathrm{He}$ in fast wind unexpectedly increases with decreasing $\left|\sigma_c\right|$. We show that these large gradients are due to the presence of compressive fluctuations. Accounting for the solar wind's compressibility ($\left|\delta n/n\right|$), there are two subsets of enhanced $A_\mathrm{He}$ in excess of typical fast wind values. The subset with a large compressibility is likely from neither continuously nor intermittently open sources. The portion of the solar wind speed distribution over which these fluctuations are most significant corresponds to the range of Alfv\'en wave-poor solar wind from continuously open source regions, which is likely analogous to the Alfv\'enic slow wind. Mapping the results of this work to Alterman & D'Amicis (2025) and vice versa shows that, in any given $\left|\delta n/n\right|$ quantile, $\left|\sigma_c\right| \lesssim 0.65$, an upper bound on non-Alfv\'enic cross helicity. Similarly, $\left|\delta n/n\right| \lesssim 0.15$ in any given $\left|\sigma_c\right|$ quantile, is an upper bound on incompressible fluctuations. We conclude that $\left|\delta n/n\right|$ is essential for characterizing the solar wind helium abundance and possibly regulating it.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:30:55 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Alterman",
"B. L.",
""
],
[
"D'Amicis",
"R.",
""
]
] |
2508.20392
|
Chengjun Zhang
|
Chengjun Zhang, Yuhao Zhang, Jie Yang and Mohamad Sawan
|
Ultra-Low-Latency Spiking Neural Networks with Temporal-Dependent Integrate-and-Fire Neuron Model for Objects Detection
|
12 pages, 8 figures
| null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Spiking Neural Networks (SNNs), inspired by the brain, are characterized by minimal power consumption and swift inference capabilities on neuromorphic hardware, and have been widely applied to various visual perception tasks. Current ANN-SNN conversion methods have achieved excellent results in classification tasks with ultra-low time-steps, but their performance in visual detection tasks remains suboptimal. In this paper, we propose a delay-spike approach to mitigate the issue of residual membrane potential caused by heterogeneous spiking patterns. Furthermore, we propose a novel temporal-dependent Integrate-and-Fire (tdIF) neuron architecture for SNNs. This enables Integrate-and-fire (IF) neurons to dynamically adjust their accumulation and firing behaviors based on the temporal order of time-steps. Our method enables spikes to exhibit distinct temporal properties, rather than relying solely on frequency-based representations. Moreover, the tdIF neuron maintains energy consumption on par with traditional IF neuron. We demonstrate that our method achieves more precise feature representation with lower time-steps, enabling high performance and ultra-low latency in visual detection tasks. In this study, we conduct extensive evaluation of the tdIF method across two critical vision tasks: object detection and lane line detection. The results demonstrate that the proposed method surpasses current ANN-SNN conversion approaches, achieving state-of-the-art performance with ultra-low latency (within 5 time-steps).
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:32:45 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhang",
"Chengjun",
""
],
[
"Zhang",
"Yuhao",
""
],
[
"Yang",
"Jie",
""
],
[
"Sawan",
"Mohamad",
""
]
] |
2508.20393
|
Qichen Huang
|
Qichen Huang, Biwei Jiang
|
The frequency-independent radio luminosity -- period relation of RS CVn stars
| null | null | null | null |
astro-ph.SR
|
http://creativecommons.org/licenses/by/4.0/
|
Radio emissions from RS CVn objects exhibit distinct characteristics at low and high frequencies, widely attributed to differing radiation mechanisms. The disparate processes of high-frequency gyrosynchrotron and low-frequency electron cyclotron maser emissions have traditionally suggested an absence of correlation in their radio luminosities. Our study presents a frequency-independent linear correlation between radio luminosity ($L_R$) and orbital/rotational periods ($P$) in RS CVn binaries. Analyzing the Sydney Radio Star Catalogue (SRSC) data, we derived orbital periods for 42 of 60 RS CVn sources using TESS light curves, revealing a strong positive correlation (PCC = 0.698, $P$ = 3.95e-7) between $\log_{10}L_R$ and $\log_{10}P$. This correlation remains across frequencies from 144-3000 MHz, showing uniform luminosity behavior. By combining light curve analysis with stellar mass-radius-luminosity relationships, we calculated parameters like binary mass, primary/secondary mass, Rossby number, and binary separation for eight RS CVn systems. The results show a notable correlation between radio luminosity and binary mass, primary mass, and separation (PCC = 0.663, 0.663, 0.719), with separation showing the strongest correlation. This suggests the radio emission may largely originate from the binary components' interaction, challenging existing models of RS CVn radio emission mechanisms and offering insights into the individual versus collective origins of these emissions.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:36:39 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Huang",
"Qichen",
""
],
[
"Jiang",
"Biwei",
""
]
] |
2508.20394
|
Jiayu Chen
|
Jiayu Chen, Zhenhui Xu, Xinghu Wang
|
Bootstrap Policy Iteration for Stochastic LQ Tracking with Multiplicative Noise
| null | null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
This paper studies the optimal tracking control problem for continuous-time stochastic linear systems with multiplicative noise. The solution framework involves solving a stochastic algebraic Riccati equation for the feedback gain and a Sylvester equation for the feedforward gain. To enable model-free optimal tracking, we first develop a two-phase bootstrap policy iteration (B-PI) algorithm, which bootstraps a stabilizing control gain from the trivially initialized zero-value start and proceeds with standard policy iteration. Building on this algorithm, we propose a data-driven, off-policy reinforcement learning approach that ensures convergence to the optimal feedback gain under the interval excitation condition. We further introduce a data-driven method to compute the feedforward using the obtained feedback gain. Additionally, for systems with state-dependent noise, we propose a shadow system-based optimal tracking method to eliminate the need for probing noise. The effectiveness of the proposed methods is demonstrated through numerical examples.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:39:13 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Chen",
"Jiayu",
""
],
[
"Xu",
"Zhenhui",
""
],
[
"Wang",
"Xinghu",
""
]
] |
2508.20395
|
Xu Guo
|
Xu Guo
|
Measuring Reasoning Utility in LLMs via Conditional Entropy Reduction
|
11 pages, 4 figures
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Recent advancements in large language models (LLMs) often rely on generating intermediate reasoning steps to enhance accuracy. However, little work has examined how reasoning utility contributes to the final answer's correctness. Due to the stochastic nature of autoregressive generation, generating more context does not guarantee increased confidence in the answer. If we could predict, during generation, whether a reasoning step will be useful, we could stop early or prune ineffective steps, avoiding distractions in the final decision.
We present an oracle study on MATH dataset, using Qwen2.5-32B and GPT-4o to generate reasoning chains, and then employing a separate model (Qwen3-8B) to quantify the utility of these chains for final accuracy. Specifically, we measure the model's uncertainty on the answer span Y at each reasoning step using conditional entropy (expected negative log-likelihood over the vocabulary) with context expanding step by step. Our results show a clear pattern: conditional entropy that decreases over steps is strongly associated with correct answers, whereas flat or increasing entropy often results in wrong answers. We also corroborate that incorrect reasoning paths tend to be longer than correct ones, suggesting that longer reasoning does not necessarily yield better outcomes. These findings serve as a foundation to inspire future work on designing efficient reasoning pipelines that detect and avoid unproductive reasoning early.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:43:38 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Guo",
"Xu",
""
]
] |
2508.20396
|
Mihajlo Grbovic
|
Guillaume Guy, Mihajlo Grbovic, Chun How Tan, Han Zhao
|
BiListing: Modality Alignment for Listings
| null |
Proceedings of the 34th ACM International Conference on Information and Knowledge Management, CIKM 2025
|
10.1145/3746252.3761577
| null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Airbnb is a leader in offering travel accommodations. Airbnb has historically relied on structured data to understand, rank, and recommend listings to guests due to the limited capabilities and associated complexity arising from extracting meaningful information from text and images. With the rise of representation learning, leveraging rich information from text and photos has become easier. A popular approach has been to create embeddings for text documents and images to enable use cases of computing similarities between listings or using embeddings as features in an ML model.
However, an Airbnb listing has diverse unstructured data: multiple images, various unstructured text documents such as title, description, and reviews, making this approach challenging. Specifically, it is a non-trivial task to combine multiple embeddings of different pieces of information to reach a single representation.
This paper proposes BiListing, for Bimodal Listing, an approach to align text and photos of a listing by leveraging large-language models and pretrained language-image models. The BiListing approach has several favorable characteristics: capturing unstructured data into a single embedding vector per listing and modality, enabling zero-shot capability to search inventory efficiently in user-friendly semantics, overcoming the cold start problem, and enabling listing-to-listing search along a single modality, or both.
We conducted offline and online tests to leverage the BiListing embeddings in the Airbnb search ranking model, and successfully deployed it in production, achieved 0.425% of NDCB gain, and drove tens of millions in incremental revenue.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:47:31 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Guy",
"Guillaume",
""
],
[
"Grbovic",
"Mihajlo",
""
],
[
"Tan",
"Chun How",
""
],
[
"Zhao",
"Han",
""
]
] |
2508.20397
|
Wu-Zhong Guo
|
Wu-zhong Guo
|
Spacetime Density Matrix: Formalism and Properties
|
45 pages, many figures to represent the spacetime density matrices and their moments
| null | null | null |
hep-th cond-mat.stat-mech gr-qc quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
In this paper, we develop the general formalism and properties of the spacetime density matrix, which captures correlations among different Cauchy surfaces and can be regarded as a natural generalization of the standard density matrix defined on a single Cauchy surface. We present the construction of the spacetime density matrix in general quantum systems and its representation via the Schwinger Keldysh path integral. We further introduce a super-operator framework, within which the spacetime density matrix appears as a special case, and discuss possible generalizations from this perspective. We also show that the spacetime density matrix satisfies a Liouville von Neumann type equation of motion. When considering subsystems, a reduced spacetime density matrix can be defined by tracing over complementary degrees of freedom. We study the general properties of its moments and, in particular, derive universal short time behavior of the second moment. We find that coupling between subsystems plays a crucial role in obtaining nontrivial results. Assuming weak coupling, we develop a perturbative method to compute the moments systematically.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:50:35 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Guo",
"Wu-zhong",
""
]
] |
2508.20398
|
Shijie Wang
|
Shijie Wang, Lei Li
|
TF-TransUNet1D: Time-Frequency Guided Transformer U-Net for Robust ECG Denoising in Digital Twin
|
9 pages, 3 figures International Workshop on Digital Twin for Healthcare (DT4H) in MICCAI 2025 (Daejeon, Republic of Korea)
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Electrocardiogram (ECG) signals serve as a foundational data source for cardiac digital twins, yet their diagnostic utility is frequently compromised by noise and artifacts. To address this issue, we propose TF-TransUNet1D, a novel one-dimensional deep neural network that integrates a U-Net-based encoder-decoder architecture with a Transformer encoder, guided by a hybrid time-frequency domain loss. The model is designed to simultaneously capture local morphological features and long-range temporal dependencies, which are critical for preserving the diagnostic integrity of ECG signals. To enhance denoising robustness, we introduce a dual-domain loss function that jointly optimizes waveform reconstruction in the time domain and spectral fidelity in the frequency domain. In particular, the frequency-domain component effectively suppresses high-frequency noise while maintaining the spectral structure of the signal, enabling recovery of subtle but clinically significant waveform components. We evaluate TF-TransUNet1D using synthetically corrupted signals from the MIT-BIH Arrhythmia Database and the Noise Stress Test Database (NSTDB). Comparative experiments against state-of-the-art baselines demonstrate consistent superiority of our model in terms of SNR improvement and error metrics, achieving a mean absolute error of 0.1285 and Pearson correlation coefficient of 0.9540. By delivering high-precision denoising, this work bridges a critical gap in pre-processing pipelines for cardiac digital twins, enabling more reliable real-time monitoring and personalized modeling.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:51:19 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Shijie",
""
],
[
"Li",
"Lei",
""
]
] |
2508.20399
|
Harshit Mishra
|
Harshit Mishra, Sucheta Soundarajan
|
A Case Study of Balanced Query Recommendation on Wikipedia
|
Accepted at FAccTRec 2025 workshop at recsys 2025
| null | null | null |
cs.IR
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Modern IR systems are an extremely important tool for seeking information. In addition to search, such systems include a number of query reformulation methods, such as query expansion and query recommendations, to provide high quality results. However, results returned by such methods sometimes exhibit undesirable or wrongful bias with respect to protected categories such as gender or race. Our earlier work considered the problem of balanced query recommendation, where instead of re-ranking a list of results based on fairness measures, the goal was to suggest queries that are relevant to a user's search query but exhibit less bias than the original query. In this work, we present a case study of BalancedQR using an extension of BalancedQR that handles biases in multiple dimensions. It employs a Pareto front approach that finds balanced queries, optimizing for multiple objectives such as gender bias and regional bias, along with the relevance of returned results. We evaluate the extended version of BalancedQR on a Wikipedia dataset.Our results demonstrate the effectiveness of our extension to BalancedQR framework and highlight the significant impact of subtle query wording,linguistic choice on retrieval.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:52:31 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Mishra",
"Harshit",
""
],
[
"Soundarajan",
"Sucheta",
""
]
] |
2508.20400
|
Yijia Sun
|
Yijia Sun, Shanshan Huang, Linxiao Che, Haitao Lu, Qiang Luo, Kun Gai, Guorui Zhou
|
MPFormer: Adaptive Framework for Industrial Multi-Task Personalized Sequential Retriever
|
CIKM 2025
| null | null | null |
cs.IR cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Modern industrial recommendation systems encounter a core challenge of multi-stage optimization misalignment: a significant semantic gap exists between the multi-objective optimization paradigm widely used in the ranking phase and the single-objective modeling in the retrieve phase. Although the mainstream industry solution achieves multi-objective coverage through parallel multi-path single-objective retrieval, this approach leads to linear growth of training and serving resources with the number of objectives and has inherent limitations in handling loosely coupled objectives. This paper proposes the MPFormer, a dynamic multi-task Transformer framework, which systematically addresses the aforementioned issues through three innovative mechanisms. First, an objective-conditioned transformer that jointly encodes user behavior sequences and multi-task semantics through learnable attention modulation; second, personalized target weights are introduced to achieve dynamic adjustment of retrieval results; finally, user personalization information is incorporated into token representations and the Transformer structure to further enhance the model's representation ability. This framework has been successfully integrated into Kuaishou short video recommendation system, stably serving over 400 million daily active users. It significantly improves user daily engagement and system operational efficiency. Practical deployment verification shows that, compared with traditional solutions, it effectively optimizes the iterative paradigm of multi-objective retrieval while maintaining service response speed, providing a scalable multi-objective solution for industrial recommendation systems.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:53:55 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Sun",
"Yijia",
""
],
[
"Huang",
"Shanshan",
""
],
[
"Che",
"Linxiao",
""
],
[
"Lu",
"Haitao",
""
],
[
"Luo",
"Qiang",
""
],
[
"Gai",
"Kun",
""
],
[
"Zhou",
"Guorui",
""
]
] |
2508.20401
|
Alexandre Andre
|
Alexandre Andre, Gauthier Roy, Eva Dyer, Kai Wang
|
Revealing Potential Biases in LLM-Based Recommender Systems in the Cold Start Setting
|
In Proceedings of 2nd Workshop on Evaluating and Applying Recommendation Systems with Large Language Models (EARL) at RecSys 2025 (EARL 2025)
| null | null | null |
cs.IR cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Large Language Models (LLMs) are increasingly used for recommendation tasks due to their general-purpose capabilities. While LLMs perform well in rich-context settings, their behavior in cold-start scenarios, where only limited signals such as age, gender, or language are available, raises fairness concerns because they may rely on societal biases encoded during pretraining. We introduce a benchmark specifically designed to evaluate fairness in zero-context recommendation. Our modular pipeline supports configurable recommendation domains and sensitive attributes, enabling systematic and flexible audits of any open-source LLM. Through evaluations of state-of-the-art models (Gemma 3 and Llama 3.2), we uncover consistent biases across recommendation domains (music, movies, and colleges) including gendered and cultural stereotypes. We also reveal a non-linear relationship between model size and fairness, highlighting the need for nuanced analysis.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 03:57:13 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Andre",
"Alexandre",
""
],
[
"Roy",
"Gauthier",
""
],
[
"Dyer",
"Eva",
""
],
[
"Wang",
"Kai",
""
]
] |
2508.20402
|
Kunio Kaneta
|
Kunio Kaneta, Tomo Takahashi, Natsumi Watanabe
|
Post-Reheating Inflaton Production as a Probe of Reheating Dynamics
|
6 pages, 2 figures
| null | null | null |
hep-ph astro-ph.CO
|
http://creativecommons.org/licenses/by/4.0/
|
Cosmological reheating bridges the inflationary epoch and the hot big bang phase, yet its underlying dynamics remain poorly understood. In this work, we investigate a minimal scenario in which the inflaton evolves under a simple power-law potential during reheating and interacts with other particles via renormalizable couplings. We show that inflaton quanta can be regenerated from the thermal bath even after the decay of the coherent inflaton field, unveiling a previously overlooked channel for inflaton particle production, which offers a novel window into probing reheating. Remarkably, this mechanism may also account for the observed dark matter abundance, providing a natural link between early Universe dynamics and present-day cosmological observations.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:00:23 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Kaneta",
"Kunio",
""
],
[
"Takahashi",
"Tomo",
""
],
[
"Watanabe",
"Natsumi",
""
]
] |
2508.20403
|
Tiancheng Zhao
|
Tiancheng Zhao, Zekun Yin, Huihai An, Xiaoyu Yang, Zhou Jin, Jiasi Shen, Helen Xu
|
pdGRASS: A Fast Parallel Density-Aware Algorithm for Graph Spectral Sparsification
| null | null | null | null |
cs.DC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Graph Spectral Sparsification (GSS) identifies an ultra-sparse subgraph, or sparsifier, whose Laplacian matrix closely approximates the spectral properties of the original graph, enabling substantial reductions in computational complexity for computationally intensive problems in scientific computing. The state-of-the-art method for efficient GSS is feGRASS, consisting of two steps: 1) spanning tree generation and 2) off-tree edge recovery. However, feGRASS suffers from two main issues: 1) difficulties in parallelizing the recovery step for strict data dependencies, and 2) performance degradation on skewed inputs, often requiring multiple passes to recover sufficient edges. To address these challenges, we propose parallel density-aware Graph Spectral Sparsification (pdGRASS), a parallel algorithm that organizes edges into disjoint subtasks without data dependencies between them, enabling efficient parallelization and sufficient edge recovery in a single pass. We empirically evaluate feGRASS and pdGRASS based on 1) off-tree edge-recovery runtime and 2) sparsifier quality, measured by the iteration count required for convergence in a preconditioned conjugate gradient (PCG) application. The evaluation demonstrates that, depending on the number of edges recovered, pdGRASS achieves average speedups ranging from 3.9x to 8.8x. The resulting sparsifiers also show between 1.2x higher and 1.8x lower PCG iteration counts, with further improvements as more edges are recovered. Additionally, pdGRASS mitigates the worst-case runtimes of feGRASS with over 1000x speedup. These results highlight pdGRASS's significant improvements in scalability and performance for the graph spectral sparsification problem.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:02:56 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhao",
"Tiancheng",
""
],
[
"Yin",
"Zekun",
""
],
[
"An",
"Huihai",
""
],
[
"Yang",
"Xiaoyu",
""
],
[
"Jin",
"Zhou",
""
],
[
"Shen",
"Jiasi",
""
],
[
"Xu",
"Helen",
""
]
] |
2508.20404
|
Chenyi Zhuang
|
Chengyue Yu, Siyuan Lu, Chenyi Zhuang, Dong Wang, Qintong Wu, Zongyue Li, Runsheng Gan, Chunfeng Wang, Siqi Hou, Gaochi Huang, Wenlong Yan, Lifeng Hong, Aohui Xue, Yanfeng Wang, Jinjie Gu, David Tsai, Tao Lin
|
AWorld: Orchestrating the Training Recipe for Agentic AI
| null | null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The learning from practice paradigm is crucial for developing capable Agentic AI systems, yet it is severely hampered by inefficient experience generation, a bottleneck especially pronounced in complex benchmarks like GAIA. To address this, we introduce AWorld, an open-source system engineered for large-scale agent-environment interaction. By distributing tasks across a cluster, AWorld accelerates experience collection by 14.6x compared to standard single-node, sequential execution. This critical speedup makes extensive reinforcement learning practical and scalable. Leveraging this capability, we trained a Qwen3-32B-based agent that significantly outperforms its base model, increasing its overall GAIA accuracy from 21.59% to 32.23%. On the benchmark's most challenging levels, our agent achieves a score of 16.33%, surpassing the performance of leading proprietary models. Our open-source system and resulting agent provide a practical blueprint for a complete agentic AI training pipeline, from efficient interaction to demonstrable model improvement.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:04:30 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Yu",
"Chengyue",
""
],
[
"Lu",
"Siyuan",
""
],
[
"Zhuang",
"Chenyi",
""
],
[
"Wang",
"Dong",
""
],
[
"Wu",
"Qintong",
""
],
[
"Li",
"Zongyue",
""
],
[
"Gan",
"Runsheng",
""
],
[
"Wang",
"Chunfeng",
""
],
[
"Hou",
"Siqi",
""
],
[
"Huang",
"Gaochi",
""
],
[
"Yan",
"Wenlong",
""
],
[
"Hong",
"Lifeng",
""
],
[
"Xue",
"Aohui",
""
],
[
"Wang",
"Yanfeng",
""
],
[
"Gu",
"Jinjie",
""
],
[
"Tsai",
"David",
""
],
[
"Lin",
"Tao",
""
]
] |
2508.20405
|
Yasuaki Fujitani
|
Yasuaki Fujitani, Yohei Sakurai
|
Stability of weighted minimal hypersurfaces under a lower $1$-weighted Ricci curvature bound
|
17 pages
| null | null | null |
math.DG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We will study the $1$-weighted Ricci curvature in view of the extrinsic geometric analysis. We derive several geometric consequences concerning stable weighted minimal hypersurfaces in weighted manifolds under a lower $1$-weighted Ricci curvature bound. We prove a Schoen-Yau type criterion, and conclude a structure theorem for three-dimensional weighted manifolds of non-negative $1$-weighted Ricci curvature. We also show non-existence results under volume growth conditions, and conclude smooth compactness theorems.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:05:23 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Fujitani",
"Yasuaki",
""
],
[
"Sakurai",
"Yohei",
""
]
] |
2508.20406
|
Kyle Helson
|
Kyle R. Helson, Carol Yan Yan Chan, Stefan Arseneau, Alyssa Barlis, Charles L. Bennett, Thomas M. Essinger-Hileman, Haiquan Guo, Tobias Marriage, Manuel A. Quijada, Ariel E. Tokarz, Stephanie L. Vivod, and Edward J. Wollack
|
Diamond-loaded polyimide aerogel scattering filters and their applications in astrophysical and planetary science observations
|
Submitted to RSI. 13 pages, 14 figures
| null | null | null |
astro-ph.IM cond-mat.mtrl-sci physics.optics
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Infrared-blocking, aerogel-based scattering filters have a broad range of potential applications in astrophysics and planetary science instruments in the far-infrared, sub-millimeter, and microwave regimes. This paper demonstrates the ability of conductively-loaded, polyimide aerogel filters to meet the mechanical and science instrument requirements for several experiments, including the Cosmology Large Angular Scale Surveyor (CLASS), the Experiment for Cryogenic Large-Aperture Intensity Mapping (EXCLAIM), and the Sub-millimeter Solar Observation Lunar Volatiles Experiment (SSOLVE). Thermal multi-physics simulations of the filters predict their performance when integrated into a cryogenic receiver. Prototype filters have survived cryogenic cycling to 4\,K with no degradation in mechanical properties. Measurement of total hemispherical reflectance and transmittance, as well as cryogenic tests of the aerogel filters in a full receiver context, allow estimates of the integrated infrared emissivity of the filters. Knowledge of the emissivity will help instrument designers incorporate the filters into future experiments in planetary science, astrophysics, and cosmology.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:07:49 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Helson",
"Kyle R.",
""
],
[
"Chan",
"Carol Yan Yan",
""
],
[
"Arseneau",
"Stefan",
""
],
[
"Barlis",
"Alyssa",
""
],
[
"Bennett",
"Charles L.",
""
],
[
"Essinger-Hileman",
"Thomas M.",
""
],
[
"Guo",
"Haiquan",
""
],
[
"Marriage",
"Tobias",
""
],
[
"Quijada",
"Manuel A.",
""
],
[
"Tokarz",
"Ariel E.",
""
],
[
"Vivod",
"Stephanie L.",
""
],
[
"Wollack",
"Edward J.",
""
]
] |
2508.20407
|
Zhongpan Tang
|
Zhongpan Tang
|
Rethinking Transformer Connectivity: TLinFormer, A Path to Exact, Full Context-Aware Linear Attention
| null | null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Transformer architecture has become a cornerstone of modern artificial intelligence, but its core self-attention mechanism suffers from a complexity bottleneck that scales quadratically with sequence length, severely limiting its application in long-sequence tasks. To address this challenge, existing linear attention methods typically sacrifice model performance by relying on data-agnostic kernel approximations or restrictive context selection. This paper returns to the first principles of connectionism, starting from the topological structure of information flow, to introduce a novel linear attention architecture-\textbf{TLinFormer}. By reconfiguring neuron connection patterns, TLinFormer achieves strict linear complexity while computing exact attention scores and ensuring information flow remains aware of the full historical context. This design aims to bridge the performance gap prevalent between existing efficient attention methods and standard attention. Through a series of experiments, we systematically evaluate the performance of TLinFormer against a standard Transformer baseline on long-sequence inference tasks. The results demonstrate that TLinFormer exhibits overwhelming advantages in key metrics such as \textbf{inference latency}, \textbf{KV cache efficiency}, \textbf{memory footprint}, and \textbf{overall speedup}.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:10:19 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Tang",
"Zhongpan",
""
]
] |
2508.20408
|
Haoyu Wu
|
Haoyu Wu, Qingcheng Zeng, Kaize Ding
|
Fact or Facsimile? Evaluating the Factual Robustness of Modern Retrievers
|
Proceedings of the 34th ACM International Conference on Information and Knowledge Management
| null | null | null |
cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Dense retrievers and rerankers are central to retrieval-augmented generation (RAG) pipelines, where accurately retrieving factual information is crucial for maintaining system trustworthiness and defending against RAG poisoning. However, little is known about how much factual competence these components inherit or lose from the large language models (LLMs) they are based on. We pair 12 publicly released embedding checkpoints with their original base LLMs and evaluate both sets on a factuality benchmark. Across every model evaluated, the embedding variants achieve markedly lower accuracy than their bases, with absolute drops ranging from 12 to 43 percentage points (median 28 pts) and typical retriever accuracies collapsing into the 25-35 % band versus the 60-70 % attained by the generative models. This degradation intensifies under a more demanding condition: when the candidate pool per question is expanded from four options to one thousand, the strongest retriever's top-1 accuracy falls from 33 % to 26 %, revealing acute sensitivity to distractor volume. Statistical tests further show that, for every embedding model, cosine-similarity scores between queries and correct completions are significantly higher than those for incorrect ones (p < 0.01), indicating decisions driven largely by surface-level semantic proximity rather than factual reasoning. To probe this weakness, we employed GPT-4.1 to paraphrase each correct completion, creating a rewritten test set that preserved factual truth while masking lexical cues, and observed that over two-thirds of previously correct predictions flipped to wrong, reducing overall accuracy to roughly one-third of its original level. Taken together, these findings reveal a systematic trade-off introduced by contrastive learning for retrievers: gains in semantic retrieval are paid for with losses in parametric factual knowledge......
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:13:51 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wu",
"Haoyu",
""
],
[
"Zeng",
"Qingcheng",
""
],
[
"Ding",
"Kaize",
""
]
] |
2508.20409
|
Siyao Li
|
Siyao Li, Conrad Prisby, Thomas Yang
|
Blind Source Separation-Enabled Joint Communication and Sensing in IBFD MIMO Systems
|
6 pages,5 figures, accepted by IEEE Military Communications Conference 2025
| null | null | null |
cs.ET cs.PF
|
http://creativecommons.org/licenses/by/4.0/
|
This paper addresses the challenge of joint communication and sensing (JCAS) in next-generation wireless networks, with an emphasis on in-band full-duplex (IBFD) multiple-input multiple-output (MIMO) systems. Traditionally, self-interference (SI) in IBFD systems is a major obstacle to recovering the signal of interest (SOI). Under the JCAS paradigm, however, this high-power SI signal presents an opportunity for efficient sensing. Since each transceiver node has access to the original SI signal, its environmental reflections can be exploited to estimate channel conditions and detect changes, without requiring dedicated radar waveforms. We propose a blind source separation (BSS)-based framework to simultaneously perform self-interference cancellation (SIC) and extract sensing information in IBFD MIMO settings. The approach applies the Fast Independent Component Analysis (FastICA) algorithm to separate the SI and SOI signals while enabling simultaneous signal recovery and channel estimation. Simulation results confirm the framework's effectiveness, showing improved sensing and communication performance as signal frame size increases.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:15:40 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Li",
"Siyao",
""
],
[
"Prisby",
"Conrad",
""
],
[
"Yang",
"Thomas",
""
]
] |
2508.20410
|
Spencer Mateega
|
Sam Jung, Agustin Garcinuno, Spencer Mateega
|
UI-Bench: A Benchmark for Evaluating Design Capabilities of AI Text-to-App Tools
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
AI text-to-app tools promise high quality applications and websites in minutes, yet no public benchmark rigorously verifies those claims. We introduce UI-Bench, the first large-scale benchmark that evaluates visual excellence across competing AI text-to-app tools through expert pairwise comparison. Spanning 10 tools, 30 prompts, 300 generated sites, and \textit{4000+} expert judgments, UI-Bench ranks systems with a TrueSkill-derived model that yields calibrated confidence intervals. UI-Bench establishes a reproducible standard for advancing AI-driven web design. We release (i) the complete prompt set, (ii) an open-source evaluation framework, and (iii) a public leaderboard. The generated sites rated by participants will be released soon. View the UI-Bench leaderboard at https://uibench.ai/leaderboard.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:20:00 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Jung",
"Sam",
""
],
[
"Garcinuno",
"Agustin",
""
],
[
"Mateega",
"Spencer",
""
]
] |
2508.20411
|
Weiyun Liang
|
Donglin Wang, Weiyun Liang, Chunyuan Chen, Jing Xu, and Yulong Fu
|
Governable AI: Provable Safety Under Extreme Threat Models
| null | null | null | null |
cs.AI cs.CR cs.CY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
As AI rapidly advances, the security risks posed by AI are becoming increasingly severe, especially in critical scenarios, including those posing existential risks. If AI becomes uncontrollable, manipulated, or actively evades safety mechanisms, it could trigger systemic disasters. Existing AI safety approaches-such as model enhancement, value alignment, and human intervention-suffer from fundamental, in-principle limitations when facing AI with extreme motivations and unlimited intelligence, and cannot guarantee security. To address this challenge, we propose a Governable AI (GAI) framework that shifts from traditional internal constraints to externally enforced structural compliance based on cryptographic mechanisms that are computationally infeasible to break, even for future AI, under the defined threat model and well-established cryptographic assumptions.The GAI framework is composed of a simple yet reliable, fully deterministic, powerful, flexible, and general-purpose rule enforcement module (REM); governance rules; and a governable secure super-platform (GSSP) that offers end-to-end protection against compromise or subversion by AI. The decoupling of the governance rules and the technical platform further enables a feasible and generalizable technical pathway for the safety governance of AI. REM enforces the bottom line defined by governance rules, while GSSP ensures non-bypassability, tamper-resistance, and unforgeability to eliminate all identified attack vectors. This paper also presents a rigorous formal proof of the security properties of this mechanism and demonstrates its effectiveness through a prototype implementation evaluated in representative high-stakes scenarios.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:22:59 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Donglin",
""
],
[
"Liang",
"Weiyun",
""
],
[
"Chen",
"Chunyuan",
""
],
[
"Xu",
"Jing",
""
],
[
"Fu",
"Yulong",
""
]
] |
2508.20412
|
Zhiqiang Wang
|
Zhiqiang Wang, Junyang Zhang, Guanquan Shi, HaoRan Cheng, Yunhao Yao, Kaiwen Guo, Haohua Du, Xiang-Yang Li
|
MindGuard: Tracking, Detecting, and Attributing MCP Tool Poisoning Attack via Decision Dependence Graph
| null | null | null | null |
cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
The Model Context Protocol (MCP) is increasingly adopted to standardize the interaction between LLM agents and external tools. However, this trend introduces a new threat: Tool Poisoning Attacks (TPA), where tool metadata is poisoned to induce the agent to perform unauthorized operations. Existing defenses that primarily focus on behavior-level analysis are fundamentally ineffective against TPA, as poisoned tools need not be executed, leaving no behavioral trace to monitor.
Thus, we propose MindGuard, a decision-level guardrail for LLM agents, providing provenance tracking of call decisions, policy-agnostic detection, and poisoning source attribution against TPA. While fully explaining LLM decision remains challenging, our empirical findings uncover a strong correlation between LLM attention mechanisms and tool invocation decisions. Therefore, we choose attention as an empirical signal for decision tracking and formalize this as the Decision Dependence Graph (DDG), which models the LLM's reasoning process as a weighted, directed graph where vertices represent logical concepts and edges quantify the attention-based dependencies. We further design robust DDG construction and graph-based anomaly analysis mechanisms that efficiently detect and attribute TPA attacks. Extensive experiments on real-world datasets demonstrate that MindGuard achieves 94\%-99\% average precision in detecting poisoned invocations, 95\%-100\% attribution accuracy, with processing times under one second and no additional token cost. Moreover, DDG can be viewed as an adaptation of the classical Program Dependence Graph (PDG), providing a solid foundation for applying traditional security policies at the decision level.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:23:44 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Zhiqiang",
""
],
[
"Zhang",
"Junyang",
""
],
[
"Shi",
"Guanquan",
""
],
[
"Cheng",
"HaoRan",
""
],
[
"Yao",
"Yunhao",
""
],
[
"Guo",
"Kaiwen",
""
],
[
"Du",
"Haohua",
""
],
[
"Li",
"Xiang-Yang",
""
]
] |
2508.20413
|
Benjamin Cou\'eraud
|
Benjamin Cou\'eraud, Vikram Sunkara, Christof Sch\"utte
|
Assessing local deformation and computing scalar curvature with nonlinear conformal regularization of decoders
|
9 pages
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
One aim of dimensionality reduction is to discover the main factors that explain the data, and as such is paramount to many applications. When working with high dimensional data, autoencoders offer a simple yet effective approach to learn low-dimensional representations. The two components of a general autoencoder consist first of an encoder that maps the observed data onto a latent space; and second a decoder that maps the latent space back to the original observation space, which allows to learn a low-dimensional manifold representation of the original data. In this article, we introduce a new type of geometric regularization for decoding maps approximated by deep neural networks, namely nonlinear conformal regularization. This regularization procedure permits local variations of the decoder map and comes with a new scalar field called conformal factor which acts as a quantitative indicator of the amount of local deformation sustained by the latent space when mapped into the original data space. We also show that this regularization technique allows the computation of the scalar curvature of the learned manifold. Implementation and experiments on the Swiss roll and CelebA datasets are performed to illustrate how to obtain these quantities from the architecture.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:30:49 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Couéraud",
"Benjamin",
""
],
[
"Sunkara",
"Vikram",
""
],
[
"Schütte",
"Christof",
""
]
] |
2508.20414
|
Mengyu Sun
|
Mengyu Sun, Ziyuan Yang, Yongqiang Huang, Hui Yu, Yingyu Chen, Shuren Qi, Andrew Beng Jin Teoh, Yi Zhang
|
Federated Learning for Large Models in Medical Imaging: A Comprehensive Review
| null | null | null | null |
cs.CR cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Artificial intelligence (AI) has demonstrated considerable potential in the realm of medical imaging. However, the development of high-performance AI models typically necessitates training on large-scale, centralized datasets. This approach is confronted with significant challenges due to strict patient privacy regulations and legal restrictions on data sharing and utilization. These limitations hinder the development of large-scale models in medical domains and impede continuous updates and training with new data. Federated Learning (FL), a privacy-preserving distributed training framework, offers a new solution by enabling collaborative model development across fragmented medical datasets. In this survey, we review FL's contributions at two stages of the full-stack medical analysis pipeline. First, in upstream tasks such as CT or MRI reconstruction, FL enables joint training of robust reconstruction networks on diverse, multi-institutional datasets, alleviating data scarcity while preserving confidentiality. Second, in downstream clinical tasks like tumor diagnosis and segmentation, FL supports continuous model updating by allowing local fine-tuning on new data without centralizing sensitive images. We comprehensively analyze FL implementations across the medical imaging pipeline, from physics-informed reconstruction networks to diagnostic AI systems, highlighting innovations that improve communication efficiency, align heterogeneous data, and ensure secure parameter aggregation. Meanwhile, this paper provides an outlook on future research directions, aiming to serve as a valuable reference for the field's development.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:31:41 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Sun",
"Mengyu",
""
],
[
"Yang",
"Ziyuan",
""
],
[
"Huang",
"Yongqiang",
""
],
[
"Yu",
"Hui",
""
],
[
"Chen",
"Yingyu",
""
],
[
"Qi",
"Shuren",
""
],
[
"Teoh",
"Andrew Beng Jin",
""
],
[
"Zhang",
"Yi",
""
]
] |
2508.20415
|
Yuqi Xiong
|
Yuqi Xiong, Wuzhen Shi, Yang Wen, Ruhan Liu
|
Graph-Based Uncertainty Modeling and Multimodal Fusion for Salient Object Detection
|
ICONIP 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In view of the problems that existing salient object detection (SOD) methods are prone to losing details, blurring edges, and insufficient fusion of single-modal information in complex scenes, this paper proposes a dynamic uncertainty propagation and multimodal collaborative reasoning network (DUP-MCRNet). Firstly, a dynamic uncertainty graph convolution module (DUGC) is designed to propagate uncertainty between layers through a sparse graph constructed based on spatial semantic distance, and combined with channel adaptive interaction, it effectively improves the detection accuracy of small structures and edge regions. Secondly, a multimodal collaborative fusion strategy (MCF) is proposed, which uses learnable modality gating weights to weightedly fuse the attention maps of RGB, depth, and edge features. It can dynamically adjust the importance of each modality according to different scenes, effectively suppress redundant or interfering information, and strengthen the semantic complementarity and consistency between cross-modalities, thereby improving the ability to identify salient regions under occlusion, weak texture or background interference. Finally, the detection performance at the pixel level and region level is optimized through multi-scale BCE and IoU loss, cross-scale consistency constraints, and uncertainty-guided supervision mechanisms. Extensive experiments show that DUP-MCRNet outperforms various SOD methods on most common benchmark datasets, especially in terms of edge clarity and robustness to complex backgrounds. Our code is publicly available at https://github.com/YukiBear426/DUP-MCRNet.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:31:48 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Xiong",
"Yuqi",
""
],
[
"Shi",
"Wuzhen",
""
],
[
"Wen",
"Yang",
""
],
[
"Liu",
"Ruhan",
""
]
] |
2508.20416
|
Hengchuan Zhu
|
Hengchuan Zhu, Yihuan Xu, Yichen Li, Zijie Meng, Zuozhu Liu
|
DentalBench: Benchmarking and Advancing LLMs Capability for Bilingual Dentistry Understanding
| null | null | null | null |
cs.CL cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent advances in large language models (LLMs) and medical LLMs (Med-LLMs) have demonstrated strong performance on general medical benchmarks. However, their capabilities in specialized medical fields, such as dentistry which require deeper domain-specific knowledge, remain underexplored due to the lack of targeted evaluation resources. In this paper, we introduce DentalBench, the first comprehensive bilingual benchmark designed to evaluate and advance LLMs in the dental domain. DentalBench consists of two main components: DentalQA, an English-Chinese question-answering (QA) benchmark with 36,597 questions spanning 4 tasks and 16 dental subfields; and DentalCorpus, a large-scale, high-quality corpus with 337.35 million tokens curated for dental domain adaptation, supporting both supervised fine-tuning (SFT) and retrieval-augmented generation (RAG). We evaluate 14 LLMs, covering proprietary, open-source, and medical-specific models, and reveal significant performance gaps across task types and languages. Further experiments with Qwen-2.5-3B demonstrate that domain adaptation substantially improves model performance, particularly on knowledge-intensive and terminology-focused tasks, and highlight the importance of domain-specific benchmarks for developing trustworthy and effective LLMs tailored to healthcare applications.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:35:51 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhu",
"Hengchuan",
""
],
[
"Xu",
"Yihuan",
""
],
[
"Li",
"Yichen",
""
],
[
"Meng",
"Zijie",
""
],
[
"Liu",
"Zuozhu",
""
]
] |
2508.20417
|
Ngoc Mai Thieu
|
Chi Minh Bui, Ngoc Mai Thieu, Van Vinh Nguyen, Json J.Jung, Khac-Hoai Nam Bui
|
KG-CQR: Leveraging Structured Relation Representations in Knowledge Graphs for Contextual Query Retrieval
|
Accepted at Main EMNLP 2025
| null | null | null |
cs.CL cs.DB
|
http://creativecommons.org/licenses/by/4.0/
|
The integration of knowledge graphs (KGs) with large language models (LLMs) offers significant potential to improve the retrieval phase of retrieval-augmented generation (RAG) systems. In this study, we propose KG-CQR, a novel framework for Contextual Query Retrieval (CQR) that enhances the retrieval phase by enriching the contextual representation of complex input queries using a corpus-centric KG. Unlike existing methods that primarily address corpus-level context loss, KG-CQR focuses on query enrichment through structured relation representations, extracting and completing relevant KG subgraphs to generate semantically rich query contexts. Comprising subgraph extraction, completion, and contextual generation modules, KG-CQR operates as a model-agnostic pipeline, ensuring scalability across LLMs of varying sizes without additional training. Experimental results on RAGBench and MultiHop-RAG datasets demonstrate KG-CQR's superior performance, achieving a 4-6% improvement in mAP and a 2-3% improvement in Recall@25 over strong baseline models. Furthermore, evaluations on challenging RAG tasks such as multi-hop question answering show that, by incorporating KG-CQR, the performance consistently outperforms the existing baseline in terms of retrieval effectiveness
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:37:15 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Bui",
"Chi Minh",
""
],
[
"Thieu",
"Ngoc Mai",
""
],
[
"Nguyen",
"Van Vinh",
""
],
[
"Jung",
"Json J.",
""
],
[
"Bui",
"Khac-Hoai Nam",
""
]
] |
2508.20418
|
Siddharth Mahesh
|
Siddharth Mahesh and Sean McWilliams and Zachariah Etienne
|
Spinning Effective-to-Backwards One Body ($\texttt{SEBOB}$): combining Effective One-Body inspirals and Backwards One-Body merger-ringdowns for aligned spin black hole binaries
|
17 pages, 9 figures
| null | null | null |
gr-qc
|
http://creativecommons.org/licenses/by/4.0/
|
High-fidelity gravitational waveform models are essential for realizing the scientific potential of next-generation gravitational-wave observatories. While highly accurate, state-of-the-art models often rely on extensive phenomenological calibrations to numerical relativity (NR) simulations for the late-inspiral and merger phases, which can limit physical insight and extrapolation to regions where NR data is sparse. To address this, we introduce the Spinning Effective-to-Backwards One Body (SEBOB) formalism, a hybrid approach that combines the well-established Effective-One-Body (EOB) framework with the analytically-driven Backwards-One-Body (BOB) model, which describes the merger-ringdown from first principles as a perturbation of the final remnant black hole. We present two variants building on the state-of-the-art $\texttt{SEOBNRv5HM}$ model: $\texttt{seobnrv5_nrnqc_bob}$, which retains standard NR-calibrated non-quasi-circular (NQC) corrections and attaches a BOB-based merger-ringdown; and a more ambitious variant, $\texttt{seobnrv5_bob}$, which uses BOB to also inform the NQC corrections, thereby reducing reliance on NR fitting and enabling higher-order ($\mathcal{C}^2$) continuity by construction. Implemented in the open-source $\texttt{NRPy}$ framework for optimized C-code generation, the SEBOB model is transparent, extensible, and computationally efficient. By comparing our waveforms to a large catalog of NR simulations, we demonstrate that SEBOB yields accuracies comparable to the highly-calibrated $\texttt{SEOBNRv5HM}$ model, providing a viable pathway towards more physically motivated and robust waveform models
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:37:19 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Mahesh",
"Siddharth",
""
],
[
"McWilliams",
"Sean",
""
],
[
"Etienne",
"Zachariah",
""
]
] |
2508.20419
|
Honger Gong
|
Hong-Er Gong, Junlin Qin, Yusen Wang, Bofeng Wu, Zhan-Feng Mai, Sen Guo and Enwei Liang
|
The shadows and photon rings of two minimal deformations of Schwarzschild black holes
|
19 pages, 50 figures
| null | null | null |
gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper primarily investigates the optical characteristics of two minimal Schwarzschild black hole deformations - the Kazakov-Solodukhin and Ghosh-Kumar black holes - under different accretion models. The event horizon, photon sphere, and critical impact parameter of the former increase compared with the Schwarzschild black hole, but those of the latter decrease. The data from the Event Horizon Telescope Collaboration are used to constrain the parameter ranges of the two black holes. In the case of spherical accretion, the quantum correction of Kazakov-Solodukhin black hole leads to the increase of black hole shadow size and the decrease of integrated intensity, while the shadow size of magnetically charged Ghosh-Kumar black hole decreases and the integrated intensity increases. The shadow radius of the black hole is independent of the spherical accretion models. For an optically and geometrically thin accretion disk, the integrated intensity is mainly contributed by direct emission, and the contributions of photon rings and lensed rings are very small. In addition, the photon rings and lensed rings of Kazakov-Solodukhin black hole are narrower, while those of Ghosh-Kumar black hole are wider. Whereas the Kazakov-Solodukhin black hole exhibits higher brightness, the Ghosh-Kumar black hole shows lower brightness. Additionally, a disk closer to the black hole correlates with a smaller shadow radius. This paper proposes a method to distinguish different black holes in a specific thin disk model.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:40:03 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Gong",
"Hong-Er",
""
],
[
"Qin",
"Junlin",
""
],
[
"Wang",
"Yusen",
""
],
[
"Wu",
"Bofeng",
""
],
[
"Mai",
"Zhan-Feng",
""
],
[
"Guo",
"Sen",
""
],
[
"Liang",
"Enwei",
""
]
] |
2508.20420
|
Chenyu Luo
|
Feng Zhang, Chengjie Pang, Yuehan Zhang, Chenyu Luo
|
CAMB: A comprehensive industrial LLM benchmark on civil aviation maintenance
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Civil aviation maintenance is a domain characterized by stringent industry standards. Within this field, maintenance procedures and troubleshooting represent critical, knowledge-intensive tasks that require sophisticated reasoning. To address the lack of specialized evaluation tools for large language models (LLMs) in this vertical, we propose and develop an industrial-grade benchmark specifically designed for civil aviation maintenance. This benchmark serves a dual purpose: It provides a standardized tool to measure LLM capabilities within civil aviation maintenance, identifying specific gaps in domain knowledge and complex reasoning. By pinpointing these deficiencies, the benchmark establishes a foundation for targeted improvement efforts (e.g., domain-specific fine-tuning, RAG optimization, or specialized prompt engineering), ultimately facilitating progress toward more intelligent solutions within civil aviation maintenance. Our work addresses a significant gap in the current LLM evaluation, which primarily focuses on mathematical and coding reasoning tasks. In addition, given that Retrieval-Augmented Generation (RAG) systems are currently the dominant solutions in practical applications , we leverage this benchmark to evaluate existing well-known vector embedding models and LLMs for civil aviation maintenance scenarios. Through experimental exploration and analysis, we demonstrate the effectiveness of our benchmark in assessing model performance within this domain, and we open-source this evaluation benchmark and code to foster further research and development:https://github.com/CamBenchmark/cambenchmark
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:42:11 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Zhang",
"Feng",
""
],
[
"Pang",
"Chengjie",
""
],
[
"Zhang",
"Yuehan",
""
],
[
"Luo",
"Chenyu",
""
]
] |
2508.20421
|
Shokoufe Faraji
|
Shokoufe Faraji, Niayesh Afshordi
|
Self-regularized entropy: How much do quantum black holes violate the no-hair theorem?
|
9 pages
| null | null | null |
gr-qc astro-ph.HE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We compute the canonical (brick wall) entropy of Hawking radiation in a quantum black hole whose exterior is described, to first order in a small quadrupole parameter, by the static $q$-metric, which is an exact vacuum solution of the Einstein equations. Counting near horizon quasinormal modes shows that a modest quadrupolar deformation self-regularizes the ultraviolet divergence: the entropy of Hawking radiation is finite for any non-vanishing quadrupole, without an ad hoc cutoff. Matching this canonical entropy to the Bekenstein-Hawking entropy leads to no-hair violating multipoles, at percent-to-tens-of-percent level, and provides concrete observational targets for the Next Generation Event Horizon Telescope (ngEHT) and the Laser Interferometer Space Antenna (LISA).
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:42:29 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Faraji",
"Shokoufe",
""
],
[
"Afshordi",
"Niayesh",
""
]
] |
2508.20422
|
Tatsuya Wada
|
Tatsuya Wada, Masakiyo Kitazawa, Kazuyuki Kanaya
|
Lee-Yang-zero ratio method in three-dimensional Ising model
|
27 pages, 7 figures
| null | null |
YITP-25-128, J-PARC-TH-0324, UTHEP-810
|
cond-mat.stat-mech hep-lat hep-ph physics.comp-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
By performing Monte Carlo simulations of the three-dimensional Ising model, we apply the recently proposed Lee-Yang-zero ratio (LYZR) method to determine the location of the critical point in this model. We demonstrate that the LYZR method is as powerful as the conventional Binder-cumulant method in studying the critical point, while the LYZR method has the advantage of suppressing the violation of the finite-size scaling and non-linearity near the critical point. We also achieve a precise determination of the values of the LYZRs at the critical point, which are universal numbers. In addition, we propose an alternative method that uses only a single Lee-Yang zero and show that it is also useful for the search for the critical point.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:42:59 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wada",
"Tatsuya",
""
],
[
"Kitazawa",
"Masakiyo",
""
],
[
"Kanaya",
"Kazuyuki",
""
]
] |
2508.20423
|
Haopeng Wang
|
Haopeng Wang, Stefaan Poedts, Andrea Lani, Luis Linan, Tinatin Baratashvili, Hyun-Jin Jeong, Rayan Dhib, Yuhao Zhou, Yucong Li, Mahdi Najafi-Ziyazi, Juan Wang, Brigitte Schmieder, and Wensi Wang
|
COCONUT: A time-evolving coronal model with an energy decomposition strategy
|
12 pages, 8 figures
| null | null | null |
astro-ph.SR
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
In this paper, we propose an energy decomposition method to improve the numerical stability of time-evolving magnetohydrodynamic (MHD) coronal models, enabling them to resolve the stronger magnetic field during solar maxima without significantly filtering out small-scale structures from the observed magnetograms.We advance the decomposed energy that excludes the magnetic energy, instead of the total energy, in time. It avoids the operation of subtracting a large magnetic energy from the total energy to obtain a very small thermal pressure in low-beta regions, thereby improving the numerical stability of MHD models. We implemented this method in COCONUT and validated the model by performing a time-evolving coronal simulation during Carrington Rotation (CR) 2296, a solar maximum CR. We also compare quasi-steady-state simulation results during the solar minimum and the increasing phase, calculated using both versions of COCONUT adopting the decomposed energy equation and the traditional full energy equation to further validate the reliability of the energy decomposition method. The simulation results show that the energy decomposition method yields results nearly identical to those of the traditional full energy equation during solar minimum, while significantly enhancing COCONUT's ability to simulate coronal evolution under strong magnetic fields, even those exceeding 100 Gauss. This method is well suited for performing quasi-realistic time-evolving coronal simulations around solar maxima without excessively filtering out the observed strong magnetic fields.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:45:26 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Wang",
"Haopeng",
""
],
[
"Poedts",
"Stefaan",
""
],
[
"Lani",
"Andrea",
""
],
[
"Linan",
"Luis",
""
],
[
"Baratashvili",
"Tinatin",
""
],
[
"Jeong",
"Hyun-Jin",
""
],
[
"Dhib",
"Rayan",
""
],
[
"Zhou",
"Yuhao",
""
],
[
"Li",
"Yucong",
""
],
[
"Najafi-Ziyazi",
"Mahdi",
""
],
[
"Wang",
"Juan",
""
],
[
"Schmieder",
"Brigitte",
""
],
[
"Wang",
"Wensi",
""
]
] |
2508.20424
|
Desen Sun
|
Desen Sun, Shuncheng Jie and Sihang Liu
|
Breaking Diffusion with Cache: Exploiting Approximate Caches in Diffusion Models
| null | null | null | null |
cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
Diffusion models are a powerful class of generative models that produce content, such as images, from user prompts, but they are computationally intensive. To mitigate this cost, recent academic and industry work has adopted approximate caching, which reuses intermediate states from similar prompts in a cache. While efficient, this optimization introduces new security risks by breaking isolation among users. This work aims to comprehensively assess new security vulnerabilities arising from approximate caching. First, we demonstrate a remote covert channel established with the cache, where a sender injects prompts with special keywords into the cache and a receiver can recover that even after days, to exchange information. Second, we introduce a prompt stealing attack using the cache, where an attacker can recover existing cached prompts based on cache hit prompts. Finally, we introduce a poisoning attack that embeds the attacker's logos into the previously stolen prompt, to render them in future user prompts that hit the cache. These attacks are all performed remotely through the serving system, which indicates severe security vulnerabilities in approximate caching.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:46:44 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Sun",
"Desen",
""
],
[
"Jie",
"Shuncheng",
""
],
[
"Liu",
"Sihang",
""
]
] |
2508.20425
|
Samuel Dayo
|
Shuhan Liu, Samuel Dayo, Peijing Li, Philip Levis, Subhasish Mitra, Thierry Tambe, David Tennenhouse, and H.-S. Philip Wong
|
The Future of Memory: Limits and Opportunities
|
3 Pages, 2 Figures, 1 Table, Accepted to SOSP 25 BigMem Workshop
| null | null | null |
cs.AR
|
http://creativecommons.org/licenses/by/4.0/
|
Memory latency, bandwidth, capacity, and energy increasingly limit performance. In this paper, we reconsider proposed system architectures that consist of huge (many-terabyte to petabyte scale) memories shared among large numbers of CPUs. We argue two practical engineering challenges, scaling and signaling, limit such designs. We propose the opposite approach. Rather than create large, shared, homogenous memories, systems explicitly break memory up into smaller slices more tightly coupled with compute elements. Leveraging advances in 2.5D/3D integration, this compute-memory node provisions private local memory, enabling accesses of node-exclusive data through micrometer-scale distances, and dramatically reduced access cost. In-package memory elements support shared state within a processor, providing far better bandwidth and energy-efficiency than DRAM, which is used as main memory for large working sets and cold data. Hardware making memory capacities and distances explicit allows software to efficiently compose this hierarchy, managing data placement and movement.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:46:57 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Liu",
"Shuhan",
""
],
[
"Dayo",
"Samuel",
""
],
[
"Li",
"Peijing",
""
],
[
"Levis",
"Philip",
""
],
[
"Mitra",
"Subhasish",
""
],
[
"Tambe",
"Thierry",
""
],
[
"Tennenhouse",
"David",
""
],
[
"Wong",
"H. -S. Philip",
""
]
] |
2508.20426
|
Gabjin Oh
|
Gabjin Oh
|
Nonlinear Evidence of Investor Heterogeneity: Retail Cash Flows as Drivers of Market Dynamics
| null | null | null | null |
q-fin.GN
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This study measures the long memory of investor-segregated cash flows within the Korean equity market from 2015 to 2024. Applying detrended fluctuation analysis (DFA) to BUY, SELL, and NET aggregates, we estimate the Hurst exponent ($H$) using both a static specification and a 250-day rolling window. All series exhibit heavy tails, with complementary cumulative distribution exponents ranging from approximately 2 to 3. As a control, time-shuffled series yield $H \approx 0.5$, confirming that the observed persistence originates from the temporal structure rather than the distributional shape. Our analysis documents long-range dependence and reveals a clear ranking of persistence across investor types. Persistence is strongest for retail BUY and SELL flows, intermediate for institutional flows, and lowest for foreign investor flows. For NET flows, however, this persistence diminishes for retail and institutional investors but remains elevated for foreign investors. The rolling $H$ exhibits clear regime sensitivity, with significant level shifts occurring around key events: the 2018--2019 tariff episode, the COVID-19 pandemic, and the period of disinflation from November 2022 to October 2024. Furthermore, regressions of daily volatility on the rolling $H$ produce positive and statistically significant coefficients for most investor groups. Notably, the $H$ of retail NET flows demonstrates predictive power for future volatility, a characteristic not found in institutional NET flows. These findings challenge the canonical noise-trader versus informed-trader dichotomy, offering a model-light, replicable diagnostic for assessing investor persistence and its regime shifts.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:49:18 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Oh",
"Gabjin",
""
]
] |
2508.20427
|
Yongqiang Han
|
Yongqiang Han, Kai Cheng, Kefan Wang and Enhong Chen
|
Rethinking Purity and Diversity in Multi-Behavior Sequential Recommendation from the Frequency Perspective
| null | null | null | null |
cs.IR cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
In recommendation systems, users often exhibit multiple behaviors, such as browsing, clicking, and purchasing. Multi-behavior sequential recommendation (MBSR) aims to consider these different behaviors in an integrated manner to improve the recommendation performance of the target behavior. However, some behavior data will also bring inevitable noise to the modeling of user interests. Some research efforts focus on data denoising from the frequency domain perspective to improve the accuracy of user preference prediction. These studies indicate that low-frequency information tends to be valuable and reliable, while high-frequency information is often associated with noise. In this paper, we argue that high-frequency information is by no means insignificant. Further experimental results highlight that low frequency corresponds to the purity of user interests, while high frequency corresponds to the diversity of user interests. Building upon this finding, we proposed our model PDB4Rec, which efficiently extracts information across various frequency bands and their relationships, and introduces Boostrapping Balancer mechanism to balance their contributions for improved recommendation performance. Sufficient experiments on real-world datasets demonstrate the effectiveness and efficiency of our model.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:55:02 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Han",
"Yongqiang",
""
],
[
"Cheng",
"Kai",
""
],
[
"Wang",
"Kefan",
""
],
[
"Chen",
"Enhong",
""
]
] |
2508.20428
|
Dongwei Chen
|
Dongwei Chen
|
A note on the c-monotonicity in optimal transport with capacity constraints
|
5 pages
| null | null | null |
math.OC
|
http://creativecommons.org/licenses/by/4.0/
|
This paper studies the geometry of the optimizer for the optimal transport problem with capacity constraints. We introduce the concept of c-capacity monotonicity, which is a generalization of c-cyclical monotonicity in optimal transport. We show that the optimizer of the optimal transport problem with capacity constraints is c-capacity monotone.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 04:57:02 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Chen",
"Dongwei",
""
]
] |
2508.20429
|
Dennis Salinel
|
Dennis F. Salinel and Kristian Hauser A. Villegas
|
Bosonization in $R$-paraparticle Luttinger models
|
16 pages, Submission to SciPost
| null | null | null |
cond-mat.str-el quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We reintroduce the parafermion-paraboson classification in $R$-paraparticles in terms of their average occupation numbers, analogous to Green's parastatistics. The notion of $p$-order in $R$-parafermions is also redefined as the maximum number of particles that can occupy a quantum state. An example of an order-$2$ $R$-parafermion with $m=2$ internal degrees of freedom is presented, which obeys an exclusion principle that is not Pauli's. The interacting $R$-parafermions are studied in the context of bosonization. Specifically, we show that while density waves are generally bosonic in nature and that flavor-charge separation naturally occurs for any one-dimensional $R$-parafermion system described by the Luttinger model, flavor waves do not always satisfy bose statistics. Comparison of the partition functions further show that only $(p=1)$-ordered $R$-parafermions are compatible with the bosonization procedure in the low-energy limit. Based from these results, we discuss a potential realization of parafermion signatures in one-dimensional systems.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:01:51 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Salinel",
"Dennis F.",
""
],
[
"Villegas",
"Kristian Hauser A.",
""
]
] |
2508.20430
|
Romy Budhi
|
Romy Hanang Setya Budhi, Nehla Shobcha, Norma Sidik Risdianto
|
Preventing Singularity Problem in $f(R)$ Dark Energy Models
|
25 pages, 8 figures
| null | null | null |
gr-qc
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The curvature singularity problem in $f(R)$ dark energy models poses a significant challenge to their viability as alternatives to the $\Lambda$CDM paradigm. In this work, we investigate the possibility of resolving this issue by incorporating higher-order corrections that are compatible with the inflationary phase. We analyze the effects of adding $R^2$, $R^{\frac{m+2}{m+1}}$ and $\alpha$-attractor representation in $f(R)$ gravity terms to types of dark energy $f(R)$ models, focusing on their ability to prevent singularities in high-density environments. Our results demonstrate that these corrections can effectively stabilize the models, ensuring their consistency across both inflationary and late-time cosmological scales.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:04:01 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Budhi",
"Romy Hanang Setya",
""
],
[
"Shobcha",
"Nehla",
""
],
[
"Risdianto",
"Norma Sidik",
""
]
] |
2508.20431
|
Antonio Victor Machado De Oliveira
|
Antonio Victor Machado de Oliveira, Debjit Sarkar, Ali Hajimiri
|
Electrical Impedance Tomography with an Integrated Picoliter-Volume Subtractive Microfluidic Chamber in 65 nm CMOS
|
4 pages, 8 figures
| null | null | null |
physics.ins-det eess.SP
|
http://creativecommons.org/licenses/by/4.0/
|
Electrical impedance tomography with fully integrated microfluidics and electronics is presented for the first time in a CMOS chip. Chambers and electrodes are fabricated in the interconnect layers of a 65 nm CMOS chip through post-processing, enabling picoliter-volumes to be processed and imaged. Tomography maps are reconstructed by reading out voltages from a 16-element electrode array and processing the data off-chip, and sources of variation in reconstruction are discussed. The EIT system presented in this work serves as a proof-of-concept towards using CMOS as a platform for co-integrated microfluidics and electronics.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:05:34 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"de Oliveira",
"Antonio Victor Machado",
""
],
[
"Sarkar",
"Debjit",
""
],
[
"Hajimiri",
"Ali",
""
]
] |
2508.20432
|
Heikki M\"antysaari
|
Heikki M\"antysaari and Bj\"orn Schenke and Chun Shen and Wenbin Zhao
|
Perturbative high-energy evolution in the IP-Glasma initial state
|
4 pages. Talk by H.M at the Quark Matter 2025 conference
| null | null | null |
nucl-th
|
http://creativecommons.org/licenses/by/4.0/
|
We include the perturbative JIMWLK energy evolution into the IP-Glasma initial state description used to simulate the early-time dynamics in heavy ion collisions. By numerically solving the JIMWLK equation on an event-by-event basis, we obtain the energy (Bjorken-$x$) dependent structure of the colliding nuclei. Combining the initial state with hydrodynamic simulations, this enables us to predict how observables evolve when moving from RHIC to LHC energies.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:15:44 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Mäntysaari",
"Heikki",
""
],
[
"Schenke",
"Björn",
""
],
[
"Shen",
"Chun",
""
],
[
"Zhao",
"Wenbin",
""
]
] |
2508.20433
|
Haoyang Shi
|
Haoyang Shi, Xing Zhang, Sitong Li, Minghang Li, Xinming Lu, Shaoxiang Xu, Guoquan Wang
|
MegaCacheX: Towards Cost-Effective Hierarchical Collaborative Content Caching in Emerging Mega-Constellations
| null | null | null | null |
eess.SY cs.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Significant latency in global content delivery primarily arises from insufficient terrestrial infrastructure. Deploying space-based content delivery networks within emerging mega-constellations provides an effective means to bridge the digital divide. However, space-based caching faces constraints from physical-layer dynamics, including dynamic topologies, time-varying inter-satellite link conditions, and limited onboard energy. In addition, existing mechanisms often lack fine-grained content categorization and global optimization. This paper proposes MegaCacheX, a cost-effective hierarchical framework for collaborative content distribution that achieves "Earth-independence" by providing cloud services directly from space. Specifically, data centers in Sun-synchronous orbit act as primary content sources, while caching nodes in mega-constellations and ground stations collaboratively form a distributed edge layer. MegaCacheX optimizes caching strategies by integrating content popularity, regional user distribution, and satellite trajectory predictions. Multi-tier caching nodes serve as service anchors, enabling seamless content delivery with low latency. A prototype implemented on a microservices-based, containerized testbed demonstrates that MegaCacheX reduces global content access latency by about 36% compared to baseline approaches, while maintaining cost efficiency.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:22:25 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Shi",
"Haoyang",
""
],
[
"Zhang",
"Xing",
""
],
[
"Li",
"Sitong",
""
],
[
"Li",
"Minghang",
""
],
[
"Lu",
"Xinming",
""
],
[
"Xu",
"Shaoxiang",
""
],
[
"Wang",
"Guoquan",
""
]
] |
2508.20434
|
Takehiko Yasuda
|
Ratko Darda and Takehiko Yasuda
|
Orbifold pseudo-effective cones of toric stacks
|
16 pages
| null | null | null |
math.AG math.NT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we explicitly describe the orbifold pseudo-effective cone of a split toric stack.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:22:40 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Darda",
"Ratko",
""
],
[
"Yasuda",
"Takehiko",
""
]
] |
2508.20435
|
Yongheng Hu
|
Yongheng Hu
|
How Big Data Dilutes Cognitive Resources, Interferes with Rational Decision-making and Affects Wealth Distribution ?
|
55 pages, 14 figures
| null | null | null |
econ.GN q-fin.EC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Big data has exponentially dilated consumption demand and speed, but can they all convert to utility? We argue about the measures of consumption and utility acquisition in CARR utility function under the condition of big data interaction, we indicate its weakness, i.e., irrational consumption does not lead to the acquisition of utility. We consider that big data, which is different from macro and micro economic signals, formed by general information entropy, affects agents' rational cognition, which makes a part of their consumption ineffective. We preliminarily propose the theory that how dilution mechanism driven by big data will affect agents' cognitive resources. Based on theoretical and empirical analysis, we construct the Consumption Adjustment Weight Function (CAWF) of agents interacting with big data and further apply it to a model of firm wealth distribution with financial frictions, we get analytical solutions according to the Mean Field Game (MFG) and find: Lower financial friction increases the average wealth of firms but also leads to greater wealth inequality. When agents convert effective consumption into utility, which is a weight of total consumption, the average wealth of firms increases with the weight increasing. Meanwhile, wealth inequality follows a U-shaped trend, and it will be the lowest level when the weight approaches to 0.5. In conclusion, we try to provide a new complementary hypothesis to refine the 'Lucas Critique' according to the cognitive resources as endowments involved in the decision-making of agents.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:24:11 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Hu",
"Yongheng",
""
]
] |
2508.20436
|
Tsukasa Iwabuchi
|
Reika Fukuizumi and Tsukasa Iwabuchi
|
Besov spaces associated with the Harmonic oscillator
| null | null | null | null |
math.AP math.FA
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
The Besov space associated with the harmonic oscillator is introduced and thoroughly explored in this paper. It provides a comprehensive summary of the fundamental concepts of the Besov spaces, their embedding properties, bilinear estimates, and related topics.
|
[
{
"version": "v1",
"created": "Thu, 28 Aug 2025 05:26:08 GMT"
}
] | 2025-08-29T00:00:00 |
[
[
"Fukuizumi",
"Reika",
""
],
[
"Iwabuchi",
"Tsukasa",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.