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 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-06-27 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2410.00853 | Kent Vashaw | Merrick Cai, Kent B. Vashaw | Formal extension of noncommutative tensor-triangular support varieties | Example added | null | null | null | math.CT math.QA math.RA math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a support variety theory defined on the compact part of a monoidal triangulated category, we define an extension to the non-compact part following the blueprint of Benson--Carlson--Rickard, Benson--Iyengar--Krause, Balmer--Favi, and Stevenson. We generalize important aspects of the theory of extended support varieties to the noncommutative case, and give characterizations of when an extended support theory detects the zero object, under certain assumptions. In particular, we show that when the original support variety theory is based on a Noetherian topological space, detects the zero object, satisfies a generalized tensor product property, and comes equipped with a comparison map, then the extended support variety also detects the zero object. In the case of stable categories of finite tensor categories, this gives conditions under which the central cohomological support admits an extension that detects the zero object, confirming part of a recent conjecture made by the second author together with Nakano and Yakimov. | [
{
"version": "v1",
"created": "Tue, 01 Oct 2024 16:39:49 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Oct 2024 17:55:32 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 17:59:10 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Cai",
"Merrick",
""
],
[
"Vashaw",
"Kent B.",
""
]
] |
2410.05090 | Simin Fan | Xinyu Zhou, Simin Fan, Martin Jaggi | HyperINF: Unleashing the HyperPower of the Schulz's Method for Data Influence Estimation | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Influence functions provide a principled method to assess the contribution of individual training samples to a specific target. Yet, their high computational costs limit their applications on large-scale models and datasets. Existing methods proposed for influence function approximation have significantly reduced the computational overheads. However, they mostly suffer from inaccurate estimation due to the lack of strong convergence guarantees from the algorithm. The family of hyperpower methods are well-known for their rigorous convergence guarantees on matrix inverse approximation, while the matrix multiplication operation can involve intractable memory and computation costs on large-scale models. We propose HyperINF, an efficient and accurate influence function approximation method which leverages the hyperpower method, specifically Schulz's iterative algorithm. To deal with the computation-intensive matrix multiplication, we incorporate the generalized fisher information (GFIM) as a low-rank approximation of the Hessian matrix, which reduces the memory and computation overheads to constant costs independent of ranks on LoRA-tuned models. We first demonstrate the superior accuracy and stability of HyperINF compared to other baselines through a synthetic convergence simulation for matrix inversion. We further validate the efficacy of HyperINF through extensive real-world data attribution tasks, including mislabeled data detection and data selection for LLM and VLM fine-tuning. On LoRA-tuned models, HyperINF achieves superior downstream performance with minimal memory and computational overhead, while other baselines suffer from significant degradation. Our codebase is available at https://github.com/Blackzxy/HyperINF. | [
{
"version": "v1",
"created": "Mon, 07 Oct 2024 14:42:45 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 23:23:23 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhou",
"Xinyu",
""
],
[
"Fan",
"Simin",
""
],
[
"Jaggi",
"Martin",
""
]
] |
2410.05138 | Nadezhda Orlova Nikolaevna | N.N. Orlova, V.D. Esin, A.V. Timonina, N.N. Kolesnikov, E.V. Deviatov | Magnetocaloric effect for the topological semimetal Co$_3$Sn$_2$S$_2$ due to the antiferromagnetic coupling of the bulk and surface spin-polarized phases | null | Journal of Magnetism and Magnetic Materials, 624, 172998 (2025) | 10.1016/j.jmmm.2025.172998 | null | cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We experimentally investigate magnetocaloric effect for the topological magnetic Weyl semimetal Co$_3$Sn$_2$S$_2$ in a wide temperature range. The isothermal magnetic entropy change $\Delta S$ is calculated from the experimental magnetization curves by using Maxwell relation. In addition to the expected $\Delta S$ peak at the Curie temperature $T_C$, we obtain another one at the temperature $T_{inv}$ of the hysteresis inversion, which is the main experimental result. The inverted hysteresis usually originates from the antiferromagnetic coupling between two magnetic phases. For Co$_3$Sn$_2$S$_2$ topological magnetic Weyl semimetal these phases are the ferromagnetic bulk and the spin-polarized topological surface states. Thus, the pronounced magnetocaloric effect at $T_{inv}$ is determined by the bulk magnetization switching by the exchange bias field of the surface spin-polarizad phase, in contrast to the ferromagnetic-paramagnetic transition at the Curie temperature $T_C$. For possible applications of magnetocaloric effect, Weyl semimetals open a new way to shift from ferromagnetic to the antiferromagnetic systems without loss of efficiency, but with higher reversibility and with smaller energy costs. | [
{
"version": "v1",
"created": "Mon, 07 Oct 2024 15:45:18 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Orlova",
"N. N.",
""
],
[
"Esin",
"V. D.",
""
],
[
"Timonina",
"A. V.",
""
],
[
"Kolesnikov",
"N. N.",
""
],
[
"Deviatov",
"E. V.",
""
]
] |
2410.07254 | Juntao Huang | Zhiting Ma, Juntao Huang | Uniform accuracy of implicit-explicit Runge-Kutta methods for linear hyperbolic relaxation systems | to be published in Journal of Scientific Computing. arXiv admin note: text overlap with arXiv:2306.08742 by other authors | null | null | null | math.NA cs.NA physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | In this paper, we study the uniform accuracy of implicit-explicit (IMEX) Runge-Kutta (RK) schemes for general linear hyperbolic relaxation systems satisfying the structural stability condition proposed in \cite{yong_singular_1999}. We establish the uniform stability and accuracy of a class of IMEX-RK schemes with spatial discretization using a Fourier spectral method. Our results demonstrate that the accuracy of the fully discretized schemes is independent of the relaxation time across all regimes. Numerical experiments on applications in traffic flows and kinetic theory verify our theoretical analysis. | [
{
"version": "v1",
"created": "Tue, 08 Oct 2024 14:26:09 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 16:41:03 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Ma",
"Zhiting",
""
],
[
"Huang",
"Juntao",
""
]
] |
2410.08082 | Yifan Zhan | Yifan Zhan, Qingtian Zhu, Muyao Niu, Mingze Ma, Jiancheng Zhao, Zhihang Zhong, Xiao Sun, Yu Qiao, Yinqiang Zheng | ToMiE: Towards Explicit Exoskeleton for the Reconstruction of Complicated 3D Human Avatars | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we highlight a critical yet often overlooked factor in most 3D human tasks, namely modeling complicated 3D human with with hand-held objects or loose-fitting clothing. It is known that the parameterized formulation of SMPL is able to fit human skin; while hand-held objects and loose-fitting clothing, are difficult to get modeled within the unified framework, since their movements are usually decoupled with the human body. To enhance the capability of SMPL skeleton in response to this situation, we propose a growth strategy that enables the joint tree of the skeleton to expand adaptively. Specifically, our method, called ToMiE, consists of parent joints localization and external joints optimization. For parent joints localization, we employ a gradient-based approach guided by both LBS blending weights and motion kernels. Once the external joints are obtained, we proceed to optimize their transformations in SE(3) across different frames, enabling rendering and explicit animation. ToMiE manages to outperform other methods across various cases with hand-held objects and loose-fitting clothing, not only in rendering quality but also by offering free animation of grown joints, thereby enhancing the expressive ability of SMPL skeleton for a broader range of applications. | [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 16:25:52 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 06:47:51 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhan",
"Yifan",
""
],
[
"Zhu",
"Qingtian",
""
],
[
"Niu",
"Muyao",
""
],
[
"Ma",
"Mingze",
""
],
[
"Zhao",
"Jiancheng",
""
],
[
"Zhong",
"Zhihang",
""
],
[
"Sun",
"Xiao",
""
],
[
"Qiao",
"Yu",
""
],
[
"Zheng",
"Yinqiang",
""
]
] |
2410.08165 | Aryo Lotfi | Aryo Lotfi, Enrico Fini, Samy Bengio, Moin Nabi, Emmanuel Abbe | Chain-of-Sketch: Enabling Global Visual Reasoning | additional experiments added, title changed from "Visual Scratchpads: Enabling Global Reasoning in Vision" to "Chain-of-Sketch: Enabling Global Visual Reasoning" | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern vision models have achieved remarkable success in benchmarks where local features provide critical information about the target. There is now a growing interest in tackling tasks requiring more global reasoning, where local features do not provide significant information. Minsky and Papert put forward such tasks in 1969 with their connectivity study, exposing the limitations of the perceptron model. In this paper, we introduce an expanded set of global visual datasets involving graphs, strings, mazes, and image grids. We show that large vision models still struggle to learn these tasks efficiently. Similarly, state-of-the-art multi-modal LLMs perform poorly on these datasets. We explain this learning inefficiency by means of the 'globality degree' measure. To mitigate this, we propose a method called chain-of-sketch (CoS). Similar to the chain-of-thought and scratchpad techniques used in language models, CoS breaks the original task into intermediate visual steps to help learn a complex task. In addition, we show that not all CoS strategies perform equally well. Our key insight is to impose a Markovian structure on the CoS frames. This leads to the introduction of 'inductive CoS' which achieves better out-of-distribution generalization and performs well even with smaller models compared to non-inductive variants. | [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 17:44:13 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:48:33 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Lotfi",
"Aryo",
""
],
[
"Fini",
"Enrico",
""
],
[
"Bengio",
"Samy",
""
],
[
"Nabi",
"Moin",
""
],
[
"Abbe",
"Emmanuel",
""
]
] |
2410.08297 | Dirk Lorenz | Jonas Bresch, Dirk A. Lorenz, Felix Schneppe, Maximilian Winkler | Matrix-free stochastic calculation of operator norms without using adjoints | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | This paper considers the problem of computing the operator norm of a linear map between finite dimensional Hilbert spaces when only evaluations of the linear map are available and under restrictive storage assumptions. We propose a stochastic method of random search type to maximize the Rayleigh quotient and employ an exact line search in the random search directions. Moreover, we show that the proposed algorithm converges to the global maximum (the operator norm) almost surely, show a sublinear convergence behavior for the corresponding eigenvector and eigenvalue equation, and illustrate the performance of the method with numerical experiments. | [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 18:35:49 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 13:20:37 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Bresch",
"Jonas",
""
],
[
"Lorenz",
"Dirk A.",
""
],
[
"Schneppe",
"Felix",
""
],
[
"Winkler",
"Maximilian",
""
]
] |
2410.09942 | Alireza Salemi | Alireza Salemi, Hamed Zamani | Learning to Rank for Multiple Retrieval-Augmented Models through Iterative Utility Maximization | null | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | This paper investigates the design of a unified search engine to serve multiple retrieval-augmented generation (RAG) agents, each with a distinct task, backbone large language model (LLM), and RAG strategy. We introduce an iterative approach where the search engine generates retrieval results for the RAG agents and gathers feedback on the quality of the retrieved documents during an offline phase. This feedback is then used to iteratively optimize the search engine using an expectation-maximization algorithm, with the goal of maximizing each agent's utility function. Additionally, we adapt this to an online setting, allowing the search engine to refine its behavior based on real-time individual agents feedback to better serve the results for each of them. Experiments on datasets from the Knowledge-Intensive Language Tasks (KILT) benchmark demonstrates that our approach significantly on average outperforms baselines across 18 RAG models. We demonstrate that our method effectively ``personalizes'' the retrieval for each RAG agent based on the collected feedback. Finally, we provide a comprehensive ablation study to explore various aspects of our method. | [
{
"version": "v1",
"created": "Sun, 13 Oct 2024 17:53:50 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:06:17 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Salemi",
"Alireza",
""
],
[
"Zamani",
"Hamed",
""
]
] |
2410.10298 | Jiwei Chen | Jiwei Chen, Yubao Sun, Laiyan Ding and Rui Huang | ROA-BEV: 2D Region-Oriented Attention for BEV-based 3D Object Detection | accepted by IROS 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-based Bird's-Eye-View (BEV) 3D object detection has recently become popular in autonomous driving. However, objects with a high similarity to the background from a camera perspective cannot be detected well by existing methods. In this paper, we propose a BEV-based 3D Object Detection Network with 2D Region-Oriented Attention (ROA-BEV), which enables the backbone to focus more on feature learning of the regions where objects exist. Moreover, our method further enhances the information feature learning ability of ROA through multi-scale structures. Each block of ROA utilizes a large kernel to ensure that the receptive field is large enough to catch information about large objects. Experiments on nuScenes show that ROA-BEV improves the performance based on BEVDepth. The source codes of this work will be available at https://github.com/DFLyan/ROA-BEV. | [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 08:51:56 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 11:02:33 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Chen",
"Jiwei",
""
],
[
"Sun",
"Yubao",
""
],
[
"Ding",
"Laiyan",
""
],
[
"Huang",
"Rui",
""
]
] |
2410.12529 | Christian Ferko | Christian Ferko, Eashan Iyer, Kasra Mossayebi, Gregor Sanfey | A Hodge Theory for Entanglement Cohomology | 53 pages; v3: final published version | Phys.Rev.A 111 (2025) 3, 032422 | null | null | hep-th math-ph math.AT math.MP quant-ph | http://creativecommons.org/licenses/by/4.0/ | We explore and extend the application of homological algebra to describe quantum entanglement, initiated in arXiv:1901.02011, focusing on the Hodge-theoretic structure of entanglement cohomology in finite-dimensional quantum systems. We construct analogues of the Hodge star operator, inner product, codifferential, and Laplacian for entanglement $k$-forms. We also prove that such $k$-forms obey versions of the Hodge isomorphism theorem and Hodge decomposition, and that they exhibit Hodge duality. As a corollary, we conclude that the dimensions of the $k$-th and $(n-k)$-th cohomologies coincide for entanglement in $n$-partite pure states, which explains a symmetry property ("Poincare duality") of the associated Poincare polynomials. | [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 13:09:56 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Dec 2024 03:14:20 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 15:18:46 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Ferko",
"Christian",
""
],
[
"Iyer",
"Eashan",
""
],
[
"Mossayebi",
"Kasra",
""
],
[
"Sanfey",
"Gregor",
""
]
] |
2410.12590 | Keith Barnatchez | Keith Barnatchez, Rachel Nethery, Bryan E. Shepherd, Giovanni Parmigiani, Kevin P. Josey | Flexible and Efficient Estimation of Causal Effects with Error-Prone Exposures: A Control Variates Approach for Measurement Error | 42 pages, 6 figures | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Exposure measurement error is a ubiquitous but often overlooked challenge in causal inference with observational data. Existing methods accounting for exposure measurement error largely rely on restrictive parametric assumptions, while emerging data-adaptive estimation approaches allow for less restrictive assumptions but at the cost of flexibility, as they are typically tailored towards rigidly-defined statistical quantities. There remains a critical need for assumption-lean estimation methods that are both flexible and possess desirable theoretical properties across a variety of study designs. In this paper, we introduce a general framework for estimation of causal quantities in the presence of exposure measurement error, adapted from the control variates approach of Yang and Ding (2019). Our method can be implemented in various two-phase sampling study designs, where one obtains gold-standard exposure measurements for a small subset of the full study sample, called the validation data. The control variates framework leverages both the error-prone and error-free exposure measurements by augmenting an initial consistent estimator from the validation data with a variance reduction term formed from the full data. We show that our method inherits double-robustness properties under standard causal assumptions. Simulation studies show that our approach performs favorably compared to leading methods under various two-phase sampling schemes. We illustrate our method with observational electronic health record data on HIV outcomes from the Vanderbilt Comprehensive Care Clinic. | [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 14:10:29 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:03:03 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Barnatchez",
"Keith",
""
],
[
"Nethery",
"Rachel",
""
],
[
"Shepherd",
"Bryan E.",
""
],
[
"Parmigiani",
"Giovanni",
""
],
[
"Josey",
"Kevin P.",
""
]
] |
2410.15953 | Gunnar Wilken | Gunnar Wilken | Fundamental sequences based on localization | Extended version of the paper published under the same title in the LNCS proceedings of the 20th Conference on Computability in Europe (CiE) 2024, containing complete proofs and generalization to simultaneously defined collapsing functions First revision: Overall improvement of readability. Minor corrections | null | null | null | math.LO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Building on Buchholz' assignment for ordinals below Bachmann-Howard ordinal, see Buchholz 2003, we introduce systems of fundamental sequences for two kinds of relativized $\vartheta$-function-based notation systems of strength $\Pi^1_1{\operatorname{-CA}_0}$ and prove Bachmann property for these systems, which is essential for monotonicity properties of subrecursive hierarchies defined on the basis of fundamental sequences. The central notion of our construction is the notion of localization, which was introduced in Wilken 2007.
The first kind of stepwise defined $\vartheta$-functions over ordinal addition as basic function fits the framework of the ordinal arithmetical toolkit developed in Wilken 2007, whereas the second kind of $\vartheta$-functions is defined simultaneously and will allow for further generalization to larger proof-theoretic ordinals, see Weiermann and Wilken 2011.
The systems of fundamental sequences given here enable the investigation of fundamental sequences and independence phenomena also in the context of patterns of resemblance, an approach to ordinal notations that is both semantic and combinatorial and was first introduced by Carlson 2001 and further analyzed in Wilken 2006, 2007, and Carlson and Wilken 2012.
Our exposition is put into the context of the abstract approach to fundamental sequences developed by Buchholz, Cichon, and Weiermann 1994. The results of this paper will be applied in the theory of Goodstein sequences, extending results of Fern\`andez-Duque and Weiermann 2024. | [
{
"version": "v1",
"created": "Mon, 21 Oct 2024 12:33:50 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 04:08:47 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wilken",
"Gunnar",
""
]
] |
2410.16155 | Tianyi Men | Tianyi Men, Pengfei Cao, Zhuoran Jin, Yubo Chen, Kang Liu, Jun Zhao | A Troublemaker with Contagious Jailbreak Makes Chaos in Honest Towns | ACL 2025 Main | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the development of large language models, they are widely used as agents in various fields. A key component of agents is memory, which stores vital information but is susceptible to jailbreak attacks. Existing research mainly focuses on single-agent attacks and shared memory attacks. However, real-world scenarios often involve independent memory. In this paper, we propose the Troublemaker Makes Chaos in Honest Town (TMCHT) task, a large-scale, multi-agent, multi-topology text-based attack evaluation framework. TMCHT involves one attacker agent attempting to mislead an entire society of agents. We identify two major challenges in multi-agent attacks: (1) Non-complete graph structure, (2) Large-scale systems. We attribute these challenges to a phenomenon we term toxicity disappearing. To address these issues, we propose an Adversarial Replication Contagious Jailbreak (ARCJ) method, which optimizes the retrieval suffix to make poisoned samples more easily retrieved and optimizes the replication suffix to make poisoned samples have contagious ability. We demonstrate the superiority of our approach in TMCHT, with 23.51%, 18.95%, and 52.93% improvements in line topology, star topology, and 100-agent settings. Encourage community attention to the security of multi-agent systems. | [
{
"version": "v1",
"created": "Mon, 21 Oct 2024 16:21:24 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 13:45:10 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Men",
"Tianyi",
""
],
[
"Cao",
"Pengfei",
""
],
[
"Jin",
"Zhuoran",
""
],
[
"Chen",
"Yubo",
""
],
[
"Liu",
"Kang",
""
],
[
"Zhao",
"Jun",
""
]
] |
2410.19232 | Kiwan Park | Kiwan Park | $\alpha$ Effect and Magnetic Diffusivity $\beta$ in Helical Plasma under Turbulence Growth | Published in Universe | null | 10.3390/universe11070203 | null | physics.plasm-ph astro-ph.SR physics.space-ph | http://creativecommons.org/licenses/by-sa/4.0/ | We investigate the transport coefficients $\alpha$ and $\beta$ in plasma systems with varying Reynolds numbers while maintaining a unit magnetic Prandtl number. {The $\alpha$ and $\beta$ tensors parameterize the turbulent electromotive force (EMF) in terms of the large-scale magnetic field ${\bf \overline{B}}$ and current density ${\bf \overline{}}$ as follows : $\langle {\bf u}\times {\bf b} \rangle = \alpha {\bf \overline{B}}-\beta {\nabla\times \bf \overline{B}}$.} In astrophysical plasmas, high fluid Reynolds numbers ($Re$) and magnetic Reynolds numbers ($Re_\mathrm{M}$) drive turbulence, where $Re$ governs flow dynamics and $Re_\mathrm{M}$ controls magnetic field evolution. The coefficients $\alpha_{\text{semi}}$ and $\beta_{\text{semi}}$ are obtained from large-scale magnetic field data as estimates of the $\alpha$ and $\beta$ tensors, while $\beta_{\text{theo}}$ is derived from turbulent kinetic energy data. The reconstructed large-scale field $\overline{B}$ agrees with simulations, confirming consistency among $\alpha$, $\beta$, and $\overline{B}$ in weakly nonlinear regimes. This highlights the need to incorporate magnetic effects under strong nonlinearity. To clarify $\alpha$ and $\beta$, we introduce a field structure model, identifying $\alpha$ as the electrodynamic induction effect and $\beta$ as the fluid-like diffusion effect. The agreement between our method and direct simulations suggests that plasma turbulence and magnetic interactions can be analyzed using fundamental physical quantities. Moreover, $\alpha_{\text{semi}}$ and $\beta_{\text{semi}}$, which successfully reproduce the numerically obtained magnetic field, provide a benchmark for future theoretical studies. | [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 00:41:53 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 22:44:41 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Park",
"Kiwan",
""
]
] |
2410.19252 | Himanshu Pandey | Sadhana Matth, S. Pandey, Himanshu Pandey | High-Performance Thermoelectric Properties of Half-Heusler CoHfSi: A First-Principles Study with Temperature-Dependent Relaxation Time | 18 pages, 8 figures | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In the ongoing search for innovative thermoelectric (TE) materials with superior TE performance globally, we aim to investigate the possible use of half-Heusler alloy CoHfSi in TE applications. We analyzed the structure stability, thermodynamic inertia and electrical and thermal transport properties using density functional formalism and semi-classical Boltzmann transport theory. Positive phonon frequencies confirm this alloy's dynamical stability, and the Born-Huang stability criterion is also satisfied, confirming the robust mechanical stability. A large Seebeck voltage of more than 150 {\mu}V/K is estimated, an essential and typical requirement for improved heat-to-electricity conversion efficiency. This Seebeck voltage can be further increased by an order of magnitude with suitable doping. The PHONO3PY algorithm and Slack's model are used to compare the lattice thermal conductivity. The latter method gives more values than the former algorithm. Despite the commonly used constant relaxation time approximation to estimate the TE performance, we adopt the temperature-dependent relaxation time and found a clear drop in figure-of merit (zT) from those estimated without considering the lattice thermal conductivity and relaxation time both, still, the zT values are remarkably more than two, for the temperatures above 500 K, which is a striking numeral in the field of TE materials. | [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 02:08:44 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:49:35 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Matth",
"Sadhana",
""
],
[
"Pandey",
"S.",
""
],
[
"Pandey",
"Himanshu",
""
]
] |
2410.19596 | Riccardo Passeggeri | Davy Paindaveine, Riccardo Passeggeri | On the robustness of semi-discrete optimal transport | null | null | null | null | math.PR math.ST stat.ME stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We derive the breakdown point for solutions of semi-discrete optimal transport problems, which characterizes the robustness of the multivariate quantiles based on optimal transport proposed in \cite{GS}. We do so under very mild assumptions: the absolutely continuous reference measure is only assumed to have a support that is \textcolor{mygreen}{convex}, whereas the target measure is a general discrete measure on a finite number, $n$ say, of atoms. The breakdown point depends on the target measure only through its probability weights (hence not on the location of the atoms) and involves the geometry of the reference measure through the \cite{Tuk1975} concept of halfspace depth. Remarkably, depending on this geometry, the breakdown point of the optimal transport median can be strictly smaller than the breakdown point of the univariate median or the breakdown point of the spatial median, namely~$\lceil n/2\rceil /2$. In the context of robust location estimation, our results provide a subtle insight on how to perform multivariate trimming when constructing trimmed means based on optimal transport. | [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 14:43:58 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:17:55 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Paindaveine",
"Davy",
""
],
[
"Passeggeri",
"Riccardo",
""
]
] |
2410.19644 | El Mahdi Chayti | El Mahdi Chayti and Nikita Doikov and Martin Jaggi | Improving Stochastic Cubic Newton with Momentum | null | null | null | null | math.OC cs.LG | http://creativecommons.org/licenses/by/4.0/ | We study stochastic second-order methods for solving general non-convex optimization problems. We propose using a special version of momentum to stabilize the stochastic gradient and Hessian estimates in Newton's method. We show that momentum provably improves the variance of stochastic estimates and allows the method to converge for any noise level. Using the cubic regularization technique, we prove a global convergence rate for our method on general non-convex problems to a second-order stationary point, even when using only a single stochastic data sample per iteration. This starkly contrasts with all existing stochastic second-order methods for non-convex problems, which typically require large batches. Therefore, we are the first to demonstrate global convergence for batches of arbitrary size in the non-convex case for the Stochastic Cubic Newton. Additionally, we show improved speed on convex stochastic problems for our regularized Newton methods with momentum. | [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 15:49:16 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 16:07:20 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Chayti",
"El Mahdi",
""
],
[
"Doikov",
"Nikita",
""
],
[
"Jaggi",
"Martin",
""
]
] |
2410.21632 | Xiaoling Cui | Haitian Wang, Yu Chen, Xiaoling Cui | Boson-Anyon-Fermion Mapping and Anyon Construction in One Dimension | 6+4 pages, 2 figures; to appear as a Letter in Phys.Rev.Research | Physical Review Research 7, L022075 (2025) | 10.1103/np63-xnh8 | null | cond-mat.quant-gas | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We establish an exact mapping between identical particles in one dimension with arbitrary exchange statistics, including bosons, anyons and fermions, provided they share the same scattering length. This boson-anyon-fermion mapping facilitates the construction of anyons from a linear superposition of spatially symmetric and anti-symmetric states. This scheme is general and has been demonstrated in a spin-1/2 Fermi gas, where both s- and p-wave bound states can be supported by manipulating spin channels. With a suitable symmetry-breaking field, these bound states are hybridized to form a fractional-wave molecule. The condensation of these molecules in a many-body system leads to anyonic superfluidity, characterized by fractional statistics upon spin exchange within a Cooper pair. These anyonic states can be detected through asymmetric momentum distributions for each spin with a chiral $k^{-3}$ tail. Our results have demonstrated the inadequacy of contact interaction model for anyons in continuum and lattices, and meanwhile proposed a convenient route for engineering fractional phases in the platform of ultracold atoms. | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 00:31:15 GMT"
},
{
"version": "v2",
"created": "Sat, 09 Nov 2024 10:08:15 GMT"
},
{
"version": "v3",
"created": "Wed, 28 May 2025 10:20:24 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wang",
"Haitian",
""
],
[
"Chen",
"Yu",
""
],
[
"Cui",
"Xiaoling",
""
]
] |
2410.21635 | Andrew Zhao | Andrew Zhao | Learning the structure of any Hamiltonian from minimal assumptions | 45 pages | Proceedings of the 57th Symposium on Theory of Computing (STOC), pp. 1201-1211, 2025 | 10.1145/3717823.3718115 | null | quant-ph cs.DS cs.LG | http://creativecommons.org/licenses/by/4.0/ | We study the problem of learning an unknown quantum many-body Hamiltonian $H$ from black-box queries to its time evolution $e^{-\mathrm{i} H t}$. Prior proposals for solving this task either impose some assumptions on $H$, such as its interaction structure or locality, or otherwise use an exponential amount of computational postprocessing. In this paper, we present algorithms to learn any $n$-qubit Hamiltonian, which do not need to know the Hamiltonian terms in advance, nor are they restricted to local interactions. Our algorithms are efficient as long as the number of terms $m$ is polynomially bounded in the system size $n$. We consider two models of control over the time evolution:~the first has access to time reversal ($t < 0$), enabling an algorithm that outputs an $\epsilon$-accurate classical description of $H$ after querying its dynamics for a total of $\widetilde{\mathcal{O}}(m/\epsilon)$ evolution time. The second access model is more conventional, allowing only forward-time evolutions;~our algorithm requires $\widetilde{\mathcal{O}}(\|H\|^3/\epsilon^4)$ evolution time in this setting. Central to our results is the recently introduced concept of a pseudo-Choi state of $H$. We extend the utility of this learning resource by showing how to use it to learn the Fourier spectrum of $H$, how to achieve nearly Heisenberg-limited scaling with it, and how to prepare it even under our more restricted access models. | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 00:43:33 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 20:47:27 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhao",
"Andrew",
""
]
] |
2410.21909 | Xiao Xia | Xiao Xia, Dan Zhang, Zibo Liao, Zhenyu Hou, Tianrui Sun, Jing Li, Ling Fu, Yuxiao Dong | SceneGenAgent: Precise Industrial Scene Generation with Coding Agent | Accepted to ACL 2025 | null | null | null | cs.CL cs.LG cs.SE | http://creativecommons.org/licenses/by/4.0/ | The modeling of industrial scenes is essential for simulations in industrial manufacturing. While large language models (LLMs) have shown significant progress in generating general 3D scenes from textual descriptions, generating industrial scenes with LLMs poses a unique challenge due to their demand for precise measurements and positioning, requiring complex planning over spatial arrangement. To address this challenge, we introduce SceneGenAgent, an LLM-based agent for generating industrial scenes through C# code. SceneGenAgent ensures precise layout planning through a structured and calculable format, layout verification, and iterative refinement to meet the quantitative requirements of industrial scenarios. Experiment results demonstrate that LLMs powered by SceneGenAgent exceed their original performance, reaching up to 81.0% success rate in real-world industrial scene generation tasks and effectively meeting most scene generation requirements. To further enhance accessibility, we construct SceneInstruct, a dataset designed for fine-tuning open-source LLMs to integrate into SceneGenAgent. Experiments show that fine-tuning open-source LLMs on SceneInstruct yields significant performance improvements, with Llama3.1-70B approaching the capabilities of GPT-4o. Our code and data are available at https://github.com/THUDM/SceneGenAgent . | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 10:01:40 GMT"
},
{
"version": "v2",
"created": "Thu, 15 May 2025 16:40:39 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 06:24:08 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Xia",
"Xiao",
""
],
[
"Zhang",
"Dan",
""
],
[
"Liao",
"Zibo",
""
],
[
"Hou",
"Zhenyu",
""
],
[
"Sun",
"Tianrui",
""
],
[
"Li",
"Jing",
""
],
[
"Fu",
"Ling",
""
],
[
"Dong",
"Yuxiao",
""
]
] |
2410.22166 | Marta Picciau | Holger Gies and Marta Picciau | Mean-field theory for self-interacting relativistic Luttinger fermions | 16 pages, 3 figures | Phys. Rev. D 111, 085001 (2025) | 10.1103/PhysRevD.111.085001 | null | hep-th hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate a class of quantum field theories with relativistic Luttinger fermions and local self-interaction in scalar channels. For an understanding of possible low-energy phases, we first classify the set of mass terms arising from scalar fermion bilinears. For large flavor numbers, we show that each of our models features a coupling branch in which the theory is asymptotically free. In order to address the long-range behavior, we use mean-field theory which is exact in the limit of large flavor numbers. We identify two models which undergo dimensional transmutation, interconnecting the asymptotically free high-energy regime with an ordered low-energy phase sustaining a vacuum condensate. We also study the analytic structure of the Luttinger-fermionic propagator in the various possible gapped phases. | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 16:03:38 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 13:28:43 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Gies",
"Holger",
""
],
[
"Picciau",
"Marta",
""
]
] |
2410.22502 | Henan Xu | Henan Xu, Yeying Zhu and Donna L. Coffman | Gender disparities in rehospitalisations after coronary artery bypass grafting: evidence from a sparse functional causal mediation analysis of the MIMIC-IV data | Revised version and re-submitted to Journal of the Royal Statistical Society, Series C: Applied Statistics | null | null | null | stat.AP stat.ME | http://creativecommons.org/licenses/by/4.0/ | Hospital readmissions following coronary artery bypass grafting (CABG) not only impose a substantial cost burden on healthcare systems but also serve as a potential indicator of the quality of medical care. Previous studies of gender effects on complications after CABG surgery have consistently revealed that women tend to suffer worse outcomes. To better understand the causal pathway from gender to the number of rehospitalisations, we study the postoperative central venous pressure (CVP), recorded over the first 24 hours of patients' intensive care unit (ICU) stay after the CABG surgery, as sparse observations of a functional mediator. Confronted with time-varying CVP measurements and zero-inflated rehospitalisation counts within 60 days following discharge, we propose a parameter-simulating quasi-Bayesian Monte Carlo approximation method that accommodates a sparse functional mediator and a zero-inflated count outcome for causal mediation analysis. We find a causal relationship between the female gender and increased rehospitalisation counts after CABG, and that time-varying central venous pressure mediates this causal effect. | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 19:47:33 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 01:35:37 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Xu",
"Henan",
""
],
[
"Zhu",
"Yeying",
""
],
[
"Coffman",
"Donna L.",
""
]
] |
2410.23271 | Lara Bohnenblust | Lara Bohnenblust, Lucile Cangemi, Henrik Johansson, Paolo Pichini | Binary Kerr black-hole scattering at 2PM from quantum higher-spin Compton | 56 pages, 2 figures, ancillary files; v2: added references, minor corrections, updated ancillary file | null | null | UUITP-31/24 | hep-th gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum higher-spin theory applied to Compton amplitudes has proven to be surprisingly useful for elucidating Kerr black hole dynamics. Here we apply the framework to compute scattering amplitudes and observables for a binary system of two rotating black holes, at second post-Minkowskian order, and to all orders in the spin-multipole expansion for certain quantities. Starting from the established three-point and conjectured Compton quantum amplitudes, the infinite-spin limit gives classical amplitudes that serves as building block that we feed into the unitarity method to construct the 2-to-2 one-loop amplitude. We give scalar box, vector box, and scalar triangle coefficients to all orders in spin, where the latter are expressed in terms of Bessel-like functions. Using the Kosower-Maybee-O'Connell formalism, the classical 2PM impulse is computed, and in parallel we work out the scattering angle and eikonal phase. We give novel all-orders-in-spin formulae for certain contributions, and the remaining ones are given up to ${\cal O}(S^{11})$. Since Kerr 2PM dynamics beyond ${\cal O}(S^{\ge 5})$ is as of yet not completely settled, this work serves as a useful reference for future studies. | [
{
"version": "v1",
"created": "Wed, 30 Oct 2024 17:53:08 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 12:41:24 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Bohnenblust",
"Lara",
""
],
[
"Cangemi",
"Lucile",
""
],
[
"Johansson",
"Henrik",
""
],
[
"Pichini",
"Paolo",
""
]
] |
2410.23440 | Gregor Maier | Ben Adcock and Michael Griebel and Gregor Maier | The Sample Complexity of Learning Lipschitz Operators with respect to Gaussian Measures | Section 6 about pointwise sampling in v2 of this paper has been cut and will appear elsewhere | null | null | null | cs.LG cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | Operator learning, the approximation of mappings between infinite-dimensional function spaces using machine learning, has gained increasing research attention in recent years. Approximate operators, learned from data, can serve as efficient surrogate models for problems in computational science and engineering, complementing traditional methods. However, despite their empirical success, our understanding of the underlying mathematical theory is in large part still incomplete. In this paper, we study the approximation of Lipschitz operators with respect to Gaussian measures. We prove higher Gaussian Sobolev regularity of Lipschitz operators and establish lower and upper bounds on the Hermite polynomial approximation error. We then study general reconstruction strategies of Lipschitz operators from $m$ arbitrary (potentially adaptive) linear samples. As a key finding, we tightly characterize the corresponding sample complexity, that is, the smallest achievable worst-case error among all possible choices of (adaptive) sampling and reconstruction strategies in terms of $m$. As a consequence, we identify an inherent curse of sample complexity: No method to approximate Lipschitz operators based on $m$ linear samples can achieve algebraic convergence rates in $m$. On the positive side, we prove that a sufficiently fast spectral decay of the covariance operator of the underlying Gaussian measure guarantees convergence rates which are arbitrarily close to any algebraic rate. Overall, by tightly characterizing the sample complexity, our work confirms the intrinsic difficulty of learning Lipschitz operators, regardless of the data or learning technique. | [
{
"version": "v1",
"created": "Wed, 30 Oct 2024 20:32:30 GMT"
},
{
"version": "v2",
"created": "Tue, 07 Jan 2025 16:07:33 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 16:15:09 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Adcock",
"Ben",
""
],
[
"Griebel",
"Michael",
""
],
[
"Maier",
"Gregor",
""
]
] |
2411.00756 | Katharina Laubscher | Katharina Laubscher and Jay D. Sau | Detection of Majorana zero modes bound to Josephson vortices in planar superconductor-topological insulator-superconductor junctions | null | Phys. Rev. B 111, 235442 (2025) | 10.1103/PhysRevB.111.235442 | null | cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study signatures of Majorana zero modes (MZMs) bound to Josephson vortices in superconductor-three-dimensional topological insulator-superconductor (S-TI-S) Josephson junctions placed in a perpendicular magnetic field. First, using semiclassical analytical as well as numerical techniques, we calculate the spatially resolved supercurrent density carried by the low-energy Caroli-de Gennes-Matricon (CdGM) states in the junction. Motivated by a recent experiment [Yue et al., Phys. Rev. B 109, 094511 (2024)], we discuss if and how the presence of vortex MZMs is reflected in supercurrent measurements, showing that Fraunhofer signatures alone are not suitable to reliably detect vortex MZMs. Next, we propose two ways in which we believe supercurrent measurements could be complemented to further verify that the junction does indeed host MZMs. Explicitly, we discuss how additional Majorana signatures could be obtained by (1) mapping out the local density of states in the junction via scanning tunneling microscopy techniques, and (2) microwave spectroscopy of the spectrum of low-energy CdGM states in the junction. | [
{
"version": "v1",
"created": "Fri, 01 Nov 2024 17:30:19 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 19:46:09 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Laubscher",
"Katharina",
""
],
[
"Sau",
"Jay D.",
""
]
] |
2411.01252 | S.M. Yousuf Iqbal Tomal | S.M. Yousuf Iqbal Tomal, Abdullah Al Shafin | Quantum Token Obfuscation via Superposition: A Post-Quantum Security Framework Using Multi-Basis Verification and Entropy-Driven Evolution | 16 pages, Significant revisions based on reviewer feedback | null | null | null | quant-ph cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Traditional cryptographic techniques, including token obfuscation, are increasingly vulnerable to quantum attacks due to advancements in quantum computing. Quantum algorithms such as Shor's and Grover's pose significant threats to classical security methods, necessitating quantum-resistant alternatives. This study proposes a quantum-based approach to token obfuscation that leverages superposition and multi-basis verification to enhance security against quantum adversaries. Tokens are encoded in quantum superposition states, ensuring probabilistic concealment until measured. A multi-basis verification protocol strengthens authentication by requiring validation across multiple quantum measurement bases. Additionally, a quantum decay protocol and token refresh mechanism dynamically manage the token lifecycle to prevent prolonged exposure and replay attacks. The model was tested through quantum simulations, evaluating entropy quality, adversarial robustness, and token verification reliability. Experimental validation demonstrates an entropy quality score of 0.9996, a 0% attack success rate across five adversarial models, and a 67% false positive rate, indicating strict security constraints. These findings confirm the effectiveness of quantum-based token obfuscation in preventing unauthorized reconstruction. The proposed approach provides a foundation for post-quantum cryptographic security by integrating entropy-driven state transformations, dynamic token evolution, and multi-basis verification. Future work will focus on optimizing computational efficiency and testing real-world implementations on quantum hardware. | [
{
"version": "v1",
"created": "Sat, 02 Nov 2024 14:05:20 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Jan 2025 07:05:50 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 17:37:18 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Tomal",
"S. M. Yousuf Iqbal",
""
],
[
"Shafin",
"Abdullah Al",
""
]
] |
2411.01556 | Alexandre Trilles | Melih Emin Can, Alexandre Trilles | On the weakness of the vague specification property | 22 pages | null | null | null | math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that the vague specification property is strictly weaker than most of the specification-like properties, by establishing its equivalence with the asymptotic average shadowing property. In particular, we see that the weak specification property implies the vague specification property, but the converse does not hold, answering the question posed by Downarowicz and Weiss in [Ergod. Th. \& Dynam. Sys. 44(9) (2024), 2565--2580]. Additionally, we prove that, for surjective systems, the asymptotic average shadowing property is equivalent to the average shadowing property if the phase space is complete with respect to the dynamical Besicovitch pseudometric. We use the combination of both results to prove that the proximal and minimal shift spaces from [Ergod. Th. \& Dynam. Sys., 45(2) (2025), 396--426] possess the vague specification property (asymptotic average shadowing property). Our findings also allow us to address a couple of questions from [Fund. Math., 224(3) (2014), 241--278] about the asymptotic average shadowing property. | [
{
"version": "v1",
"created": "Sun, 03 Nov 2024 13:08:03 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:36:56 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Can",
"Melih Emin",
""
],
[
"Trilles",
"Alexandre",
""
]
] |
2411.01685 | Mohammad Hossein Moslemi | Mohammad Hossein Moslemi, Mostafa Milani | Reducing Biases in Record Matching Through Scores Calibration | null | null | null | null | cs.LG cs.CY cs.DB | http://creativecommons.org/licenses/by/4.0/ | Record matching is the task of identifying records that refer to the same real-world entity across datasets. While most existing models optimize for accuracy, fairness has become an important concern due to the potential for unequal outcomes across demographic groups. Prior work typically focuses on binary outcomes evaluated at fixed decision thresholds. However, such evaluations can miss biases in matching scores--biases that persist across thresholds and affect downstream tasks. We propose a threshold-independent framework for measuring and reducing score bias, defined as disparities in the distribution of matching scores across groups. We show that several state-of-the-art matching methods exhibit substantial score bias, even when appearing fair under standard threshold-based metrics. To address this, we introduce two post-processing score calibration algorithms. The first, calib, aligns group-wise score distributions using the Wasserstein barycenter, targeting demographic parity. The second, ccalib, conditions on predicted labels to further reduce label-dependent biases, such as equal opportunity. Both methods are model-agnostic and require no access to model training data. calib also offers theoretical guarantees, ensuring reduced bias with minimal deviation from original scores. Experiments across real-world datasets and matching models confirm that calib and ccalib substantially reduce score bias while minimally impacting model accuracy. | [
{
"version": "v1",
"created": "Sun, 03 Nov 2024 21:01:40 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 21:36:23 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Moslemi",
"Mohammad Hossein",
""
],
[
"Milani",
"Mostafa",
""
]
] |
2411.02136 | Robert Fonod | Robert Fonod and Haechan Cho and Hwasoo Yeo and Nikolas Geroliminis | Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery | null | null | 10.1016/j.trc.2025.105205 | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents a framework for extracting georeferenced vehicle trajectories from high-altitude drone imagery, addressing key challenges in urban traffic monitoring and the limitations of traditional ground-based systems. Our approach integrates several novel contributions, including a tailored object detector optimized for high-altitude bird's-eye view perspectives, a unique track stabilization method that uses detected vehicle bounding boxes as exclusion masks during image registration, and an orthophoto and master frame-based georeferencing strategy that enhances consistent alignment across multiple drone viewpoints. Additionally, our framework features robust vehicle dimension estimation and detailed road segmentation, enabling comprehensive traffic analysis. Conducted in the Songdo International Business District, South Korea, the study utilized a multi-drone experiment covering 20 intersections, capturing approximately 12TB of 4K video data over four days. The framework produced two high-quality datasets: the Songdo Traffic dataset, comprising approximately 700,000 unique vehicle trajectories, and the Songdo Vision dataset, containing over 5,000 human-annotated images with about 300,000 vehicle instances in four classes. Comparisons with high-precision sensor data from an instrumented probe vehicle highlight the accuracy and consistency of our extraction pipeline in dense urban environments. The public release of Songdo Traffic and Songdo Vision, and the complete source code for the extraction pipeline, establishes new benchmarks in data quality, reproducibility, and scalability in traffic research. Results demonstrate the potential of integrating drone technology with advanced computer vision for precise and cost-effective urban traffic monitoring, providing valuable resources for developing intelligent transportation systems and enhancing traffic management strategies. | [
{
"version": "v1",
"created": "Mon, 04 Nov 2024 14:49:01 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 09:25:50 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 20:45:19 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Fonod",
"Robert",
""
],
[
"Cho",
"Haechan",
""
],
[
"Yeo",
"Hwasoo",
""
],
[
"Geroliminis",
"Nikolas",
""
]
] |
2411.02398 | Hoang Nguyen | Hoang H Nguyen, Khyati Mahajan, Vikas Yadav, Julian Salazar, Philip S. Yu, Masoud Hashemi, Rishabh Maheshwary | Prompting with Phonemes: Enhancing LLMs' Multilinguality for Non-Latin Script Languages | Accepted to NAACL 2025 (Main Conference). This version contains minor improvements to the camera-ready | Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) | 10.18653/v1/2025.naacl-long.599 | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Although multilingual LLMs have achieved remarkable performance across benchmarks, we find they continue to underperform on non-Latin script languages across contemporary LLM families. This discrepancy arises from the fact that LLMs are pretrained with orthographic scripts, which are dominated by Latin characters that obscure their shared phonology with non-Latin scripts. We propose leveraging phonemic transcriptions as complementary signals to induce script-invariant representations. Our study demonstrates that integrating phonemic signals improves performance across both non-Latin and Latin script languages, with a particularly significant impact on closing the performance gap between the two. Through detailed experiments, we show that phonemic and orthographic scripts retrieve distinct examples for in-context learning (ICL). This motivates our proposed Mixed-ICL retrieval strategy, where further aggregation from both leads to our significant performance improvements for both Latin script languages (up to 12.6%) and non-Latin script languages (up to 15.1%) compared to randomized ICL retrieval. | [
{
"version": "v1",
"created": "Mon, 04 Nov 2024 18:59:51 GMT"
},
{
"version": "v2",
"created": "Thu, 06 Mar 2025 05:46:40 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 17:22:53 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Nguyen",
"Hoang H",
""
],
[
"Mahajan",
"Khyati",
""
],
[
"Yadav",
"Vikas",
""
],
[
"Salazar",
"Julian",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Hashemi",
"Masoud",
""
],
[
"Maheshwary",
"Rishabh",
""
]
] |
2411.04202 | Mohamad Kazma | Mohamad H. Kazma, Salma M. Elsherif, Ahmad F. Taha | Observability and Generalized Sensor Placement for Nonlinear Quality Models in Drinking Water Networks | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper studies the problem of optimal placement of water quality (WQ) sensors in water distribution networks (WDNs), with a focus on chlorine transport, decay, and reaction models. Such models are traditionally used as suitable proxies for WQ. The literature on this topic is inveterate, but has a key limitation: it utilizes simplified single-species decay and reaction models that do not capture WQ transients for nonlinear, multi-species interactions. This results in sensor placements (SP) that do not account for nonlinear WQ dynamics. Furthermore, as WQ simulations are parameterized by hydraulic profiles and demand patterns, the placement of sensors are often hydraulics-dependent. This study produces a greedy algorithm that addresses the two aforementioned limitations. The algorithm is grounded in nonlinear dynamic systems and observability theory, and yields SPs that are submodular and robust to hydraulic changes. Case studies on benchmark water networks are provided. The key findings provide practical recommendations for WDN operators. | [
{
"version": "v1",
"created": "Wed, 06 Nov 2024 19:02:16 GMT"
},
{
"version": "v2",
"created": "Tue, 26 Nov 2024 21:14:46 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 23:06:43 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Kazma",
"Mohamad H.",
""
],
[
"Elsherif",
"Salma M.",
""
],
[
"Taha",
"Ahmad F.",
""
]
] |
2411.05199 | Leitian Tao | Leitian Tao, Xiang Chen, Tong Yu, Tung Mai, Ryan Rossi, Yixuan Li, Saayan Mitra | CodeLutra: Boosting LLM Code Generation via Preference-Guided Refinement | TMLR 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) have revolutionized code generation but require significant resources and often over-generalize, limiting their task-specific efficiency. Fine-tuning smaller, open-source LLMs provides a cost-effective alternative. However, standard supervised approaches rely only on correct examples, missing valuable insights from failures. We introduce CodeLutra, a framework that leverages both correct and incorrect code attempts. Instead of using only correct solutions, CodeLutra applies iterative preference-based refinement, comparing successful and failed outputs to better approximate desired results. This approach narrows the performance gap with state-of-the-art larger models without requiring massive datasets or auxiliary models. For instance, on a challenging data science coding task, using only 500 samples improved Llama-3-8B's accuracy from 28.2% to 48.6%, approaching GPT-4's level. By learning from both successes and mistakes, CodeLutra provides a scalable and efficient path to high-quality code generation, making smaller open-source models more competitive with leading closed-source alternatives. | [
{
"version": "v1",
"created": "Thu, 07 Nov 2024 21:51:07 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Dec 2024 18:46:21 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 18:20:39 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Tao",
"Leitian",
""
],
[
"Chen",
"Xiang",
""
],
[
"Yu",
"Tong",
""
],
[
"Mai",
"Tung",
""
],
[
"Rossi",
"Ryan",
""
],
[
"Li",
"Yixuan",
""
],
[
"Mitra",
"Saayan",
""
]
] |
2411.05933 | Feng-Yu Yue | Feng-Yu Yue and Daniel Zelazo | A Passivity Analysis for Nonlinear Consensus on Balanced Digraphs | 9 pages, 2 figures, accepted by ECC 2025, and selected to EJC | null | null | null | math.OC cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work deals with the output consensus problem for multiagent systems over balanced digraphs by passivity analysis. As the standard diffusive coupling structure only models the undirected interconnection, we propose a general approach capable of processing directed coupling and performing passivity analysis. To mitigate the complexity arising from the nonlinearity and directed interconnections, we reformulate the output consensus problem as a convergence analysis on a submanifold. We provide passivity analysis and establish a sufficient condition based on passivity for achieving output agreement in multi-agent systems over balanced digraphs. The results are supported by a numerical example. | [
{
"version": "v1",
"created": "Fri, 08 Nov 2024 19:42:29 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 19:08:41 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 18:37:49 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Yue",
"Feng-Yu",
""
],
[
"Zelazo",
"Daniel",
""
]
] |
2411.06670 | Jhixon Mac\'ias | Jhixon Mac\'ias, Reyes Ortiz | A note on the Mac\'ias topology | null | null | null | null | math.AT | http://creativecommons.org/licenses/by/4.0/ | In this paper, we study some properties of the closure operator in the Mac\'ias topology on infinite integral domains. Moreover, under certain conditions, we present topological proofs of the infiniteness of maximal ideals and non-associated irreducible elements, taking advantage of the hyperconnectedness of the Mac\'ias topology. Additionally, some problems are proposed. | [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 02:35:06 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 02:44:45 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Macías",
"Jhixon",
""
],
[
"Ortiz",
"Reyes",
""
]
] |
2411.07925 | Chengchao Yuan | Chengchao Yuan, Walter Winter, B. Theodore Zhang, Kohta Murase, Bing Zhang | Revisiting X-ray Afterglows of Jetted Tidal Disruption Events with the External Reverse Shock | 11 pages, 3 figures, 1 table. Accepted for publication in ApJ | Astrophys.J. 982 (2025) 196 | 10.3847/1538-4357/adbbde | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | We investigate the external reverse shock region of relativistic jets as the origin of X-ray afterglows of jetted tidal disruption events (TDEs) that exhibit luminous jets accompanied by fast-declining non-thermal X-ray emissions. We model the dynamics of jet propagating within an external density medium, accounting for continuous energy injection driven by accretion activities. We compute the time-dependent synchrotron and inverse Compton emissions from the reverse shock region. Our analysis demonstrates that the reverse shock scenario can potentially explain the X-ray light curves and spectra of four jetted TDEs, AT 2022cmc, Swift J1644, Swift J2058, and Swift J1112. Notably, the rapid steepening of the late-stage X-ray light curves can be attributed jointly to the jet break and cessation of the central engine as the accretion rate drops below the Eddington limit. Using parameters obtained from X-ray data fitting, we also discuss the prospects for $\gamma$-ray and neutrino detection. | [
{
"version": "v1",
"created": "Tue, 12 Nov 2024 16:59:14 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Mar 2025 02:49:52 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Yuan",
"Chengchao",
""
],
[
"Winter",
"Walter",
""
],
[
"Zhang",
"B. Theodore",
""
],
[
"Murase",
"Kohta",
""
],
[
"Zhang",
"Bing",
""
]
] |
2411.08053 | Marie-Odile Fortier | Alyssa R. Pfadt-Trilling and Marie-Odile P. Fortier | We must re-evaluate assumptions about carbon trading for effective climate change mitigation | 18 pages, 2 figures | null | null | null | econ.GN physics.ao-ph q-fin.EC stat.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Effective climate action depends on dismantling the assumptions and oversimplifications that have become the basis of climate policy. The assumption that greenhouse gases (GHG) are fungible and the use of single-point values in normalizing GHG species to CO2-equivalents can propagate inaccuracies in carbon accounting and have already led to failures of carbon offset systems. Separate emission reduction targets and tracking by GHG species are recommended to achieve long-term climate stabilization. | [
{
"version": "v1",
"created": "Wed, 06 Nov 2024 18:03:32 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:58:56 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Pfadt-Trilling",
"Alyssa R.",
""
],
[
"Fortier",
"Marie-Odile P.",
""
]
] |
2411.08272 | Oguzhan Yigit | Oguzhan Yigit and Richard C. Wilson | LBONet: Supervised Spectral Descriptors for Shape Analysis | 15 pages, 14 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The Laplace-Beltrami operator has established itself in the field of non-rigid shape analysis due to its many useful properties such as being invariant under isometric transformation, having a countable eigensystem forming an orthornormal basis, and fully characterizing geodesic distances of the manifold. However, this invariancy only applies under isometric deformations, which leads to a performance breakdown in many real-world applications. In recent years emphasis has been placed upon extracting optimal features using deep learning methods,however spectral signatures play a crucial role and still add value. In this paper we take a step back, revisiting the LBO and proposing a supervised way to learn several operators on a manifold. Depending on the task, by applying these functions, we can train the LBO eigenbasis to be more task-specific. The optimization of the LBO leads to enormous improvements to established descriptors such as the heat kernel signature in various tasks such as retrieval, classification, segmentation, and correspondence, proving the adaption of the LBO eigenbasis to both global and highly local learning settings. | [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 00:49:05 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 22:45:51 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Yigit",
"Oguzhan",
""
],
[
"Wilson",
"Richard C.",
""
]
] |
2411.08747 | Tao Jing | Tao Jing and Cheng Li | Regression for Astronomical Data with Realistic Distributions, Errors and Non-linearity | 23 pages, 16 + 3 figures, 1 table; published in AJ; Code is avalible at https://github.com/astro-jingtao/raddest | null | 10.3847/1538-3881/add891 | null | astro-ph.IM astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We have developed a new regression technique, the maximum likelihood (ML)-based method and its variant, the KS-test based method, designed to obtain unbiased regression results from typical astronomical data. A normalizing flow model is employed to automatically estimate the unobservable intrinsic distribution of the independent variable as well as the unobservable correlation between uncertainty level and intrinsic value of both independent and dependent variables from the observed data points in a variational inference based empirical Bayes approach. By incorporating these estimated distributions, our method comprehensively accounts for the uncertainties associated with both independent and dependent variables. Our test on both mock data and real astronomical data from PHANGS-ALMA and PHANGS-JWST demonstrates that, given a sufficiently large sample size (> 1000), both the ML-based method and the KS-test based method significantly outperform the existing widely-used methods, particularly in cases of low signal-to-noise ratios. The KS-test based method exhibits remarkable robustness against deviations from underlying assumptions, complex intrinsic distributions, varying correlations between uncertainty levels and intrinsic values, inaccuracies in uncertainty estimations, outliers, and saturation effects. For sample sizes between 300 and 1000, the ML-based method yields the best performance. In the low data regime (< 300), the ML-based method maintains comparable performance to other state-of-the-art methods. A GPU-compatible Python implementation of our methods, nicknamed ``raddest'', will be made publicly available upon acceptance of this paper. | [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 16:26:58 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:04:59 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Jing",
"Tao",
""
],
[
"Li",
"Cheng",
""
]
] |
2411.09033 | Kseniia Telikova | K. Telikova, J. Gonz\'alez-L\'opez, M. Aravena, A. Posses, V. Villanueva, M. Baeza-Garay, G. C. Jones, M. Solimano, L. Lee, R. J. Assef, I. De Looze, T. Diaz Santos, A. Ferrara, R. Ikeda, R. Herrera-Camus, H. \"Ubler, I. Lamperti, I. Mitsuhashi, M. Relano, M. Perna, K. Tadaki | The ALMA-CRISTAL Survey: Complex kinematics of the galaxies at the end of the Reionization Era | Submitted to A&A, 16 pages, 17 figures, 1 table | A&A 699, A5 (2025) | 10.1051/0004-6361/202452990 | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The history of gas assembly in early galaxies is reflected in their complex kinematics. While a considerable fraction of galaxies at z~5 are consistent with rotating disks, current studies indicate that the dominant galaxy assembly mechanism corresponds to mergers. Despite the important progress, the dynamical classification of galaxies at these epochs is still limited by observations' resolution. We present a detailed morphological and kinematic analysis of the far-infrared bright main sequence galaxy HZ10 at z=5.65, making use of new high-resolution ($\lesssim0.3$") [CII] 158$\mu$m ALMA and rest-frame optical JWST/NIRSpec observations. These observations reveal a previously unresolved complex morphology and kinematics of the HZ10. We confirm that HZ10 is not a single galaxy but consists of at least three components in close projected separation along the east-to-west direction. We find a [CII] bright central component (C), separated by 1.5 and 4 kpc from the east (E) and west (W) components, respectively. Our [CII] observations resolve the HZ10-C component resulting in a velocity gradient, produced by either rotation or a close-in merger. We test the rotating disk possibility using DysmalPy kinematic modeling and propose three dynamical scenarios for the HZ10 system: (i) a double merger, in which the companion galaxy HZ10-W merges with the disturbed clumpy rotation disk formed by the HZ10-C and E components; (ii) a triple merger, where the companion galaxies, HZ10-W and HZ10-E, merge with the rotation disk HZ10-C; and (iii) a quadruple merger, in which the companion galaxies HZ10-W and HZ10-E merge with the close double merger HZ10-C. Comparing [CII] with JWST/NIRSpec data, we find that [CII] emission closely resembles the broad [OIII] 5007{\AA} emission. The latter reflects the interacting nature of the system and suggests that ionized and neutral gas phases in HZ10 are well mixed. | [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 21:25:16 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Telikova",
"K.",
""
],
[
"González-López",
"J.",
""
],
[
"Aravena",
"M.",
""
],
[
"Posses",
"A.",
""
],
[
"Villanueva",
"V.",
""
],
[
"Baeza-Garay",
"M.",
""
],
[
"Jones",
"G. C.",
""
],
[
"Solimano",
"M.",
""
],
[
"Lee",
"L.",
""
],
[
"Assef",
"R. J.",
""
],
[
"De Looze",
"I.",
""
],
[
"Santos",
"T. Diaz",
""
],
[
"Ferrara",
"A.",
""
],
[
"Ikeda",
"R.",
""
],
[
"Herrera-Camus",
"R.",
""
],
[
"Übler",
"H.",
""
],
[
"Lamperti",
"I.",
""
],
[
"Mitsuhashi",
"I.",
""
],
[
"Relano",
"M.",
""
],
[
"Perna",
"M.",
""
],
[
"Tadaki",
"K.",
""
]
] |
2411.09412 | Prajakta Tanaji Mane | Prajakta Mane, Anupreeta More, Surhud More | Rapid identification of lensed type Ia supernovae with color-magnitude selection | null | null | 10.33232/001c.141393 | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | Strongly lensed type Ia supernovae (SNe Ia) provide a unique cosmological probe to address the Hubble tension problem in cosmology. In addition to the sensitivity of the time delays to the value of the Hubble constant, the transient and standard candle nature of SNe Ia also enable valuable joint constraints on the model of the lens and the cosmological parameters. The upcoming Legacy Survey of Space and Time (LSST) with the Vera C. Rubin Observatory is expected to increase the number of observed SNe Ia by an order of magnitude in ten years of its lifetime. However, finding such systems in the LSST data is a challenge. In this work, we revisit the color-magnitude (CM) diagram used previously as a means to identify lensed SNe Ia and extend the work further as follows. We simulate LSST-like photometric data ($rizy$-bands) of lensed SNe Ia and analyze it in the CM parameter space. We find that a subset of lensed SNe Ia are redder compared to unlensed SNe Ia at a given magnitude, both in the rising and falling phases of their light curves and for SNe up to $z=3$. We propose a modified selection criterion based on these new results. We show that the contamination coming from unlensed core-collapse (CC) SNe is negligible, whereas a small fraction of lensed CC SNe types Ib and Ic may get selected by this criterion as potential lensed SNe. Finally, we demonstrate that our criterion works well on a wide sample of observed unlensed SNe Ia, a handful of known multiply-imaged lensed SNe systems, and a representative sample of observed CC SNe as well as super-luminous supernovae. | [
{
"version": "v1",
"created": "Thu, 14 Nov 2024 13:06:45 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 12:18:31 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Mane",
"Prajakta",
""
],
[
"More",
"Anupreeta",
""
],
[
"More",
"Surhud",
""
]
] |
2411.09521 | Nicholas Stephen Manton | N. S. Manton | A ${\mathbb{CP}}^2$ SMEFT | 22 pages. Discussion of model's phenomenology has new Section 9.4 giving couplings of gauge bosons to Higgs boson | Journal of High Energy Physics 04 (2025) 180 | null | null | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | An extension of the Standard Model is proposed, where the Higgs field is valued in the complex projective plane ${\mathbb{CP}}^2$, rather than ${\mathbb{C}}^2$. Its geometry is consistent with $U(2) \simeq (SU(2) \times U(1))/ \mathbb{Z}_2$ electroweak gauge symmetry. The leading terms in the Lagrangian, beyond those of the Standard Model, are much more tightly constrained than in general SMEFTs, and the custodial $SO(4)$ symmetry of the Higgs sector is mildly broken. The predicted tree-level deviations from Standard Model phenomenology depend on a single large mass parameter $M$. Current experimental data imply that $M$ is a few TeV or larger. | [
{
"version": "v1",
"created": "Thu, 14 Nov 2024 15:38:43 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 16:16:05 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 14:49:35 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Manton",
"N. S.",
""
]
] |
2411.09852 | Zhichen Zeng | Zhichen Zeng, Xiaolong Liu, Mengyue Hang, Xiaoyi Liu, Qinghai Zhou, Chaofei Yang, Yiqun Liu, Yichen Ruan, Laming Chen, Yuxin Chen, Yujia Hao, Jiaqi Xu, Jade Nie, Xi Liu, Buyun Zhang, Wei Wen, Siyang Yuan, Hang Yin, Xin Zhang, Kai Wang, Wen-Yen Chen, Yiping Han, Huayu Li, Chunzhi Yang, Bo Long, Philip S. Yu, Hanghang Tong, Jiyan Yang | InterFormer: Effective Heterogeneous Interaction Learning for Click-Through Rate Prediction | 11 pages, 6 figures | null | null | null | cs.IR cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Click-through rate (CTR) prediction, which predicts the probability of a user clicking an ad, is a fundamental task in recommender systems. The emergence of heterogeneous information, such as user profile and behavior sequences, depicts user interests from different aspects. A mutually beneficial integration of heterogeneous information is the cornerstone towards the success of CTR prediction. However, most of the existing methods suffer from two fundamental limitations, including (1) insufficient inter-mode interaction due to the unidirectional information flow between modes, and (2) aggressive information aggregation caused by early summarization, resulting in excessive information loss. To address the above limitations, we propose a novel module named InterFormer to learn heterogeneous information interaction in an interleaving style. To achieve better interaction learning, InterFormer enables bidirectional information flow for mutually beneficial learning across different modes. To avoid aggressive information aggregation, we retain complete information in each data mode and use a separate bridging arch for effective information selection and summarization. Our proposed InterFormer achieves state-of-the-art performance on three public datasets and a large-scale industrial dataset. | [
{
"version": "v1",
"created": "Fri, 15 Nov 2024 00:20:36 GMT"
},
{
"version": "v2",
"created": "Wed, 08 Jan 2025 01:44:07 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 21:48:04 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zeng",
"Zhichen",
""
],
[
"Liu",
"Xiaolong",
""
],
[
"Hang",
"Mengyue",
""
],
[
"Liu",
"Xiaoyi",
""
],
[
"Zhou",
"Qinghai",
""
],
[
"Yang",
"Chaofei",
""
],
[
"Liu",
"Yiqun",
""
],
[
"Ruan",
"Yichen",
""
],
[
"Chen",
"Laming",
""
],
[
"Chen",
"Yuxin",
""
],
[
"Hao",
"Yujia",
""
],
[
"Xu",
"Jiaqi",
""
],
[
"Nie",
"Jade",
""
],
[
"Liu",
"Xi",
""
],
[
"Zhang",
"Buyun",
""
],
[
"Wen",
"Wei",
""
],
[
"Yuan",
"Siyang",
""
],
[
"Yin",
"Hang",
""
],
[
"Zhang",
"Xin",
""
],
[
"Wang",
"Kai",
""
],
[
"Chen",
"Wen-Yen",
""
],
[
"Han",
"Yiping",
""
],
[
"Li",
"Huayu",
""
],
[
"Yang",
"Chunzhi",
""
],
[
"Long",
"Bo",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Tong",
"Hanghang",
""
],
[
"Yang",
"Jiyan",
""
]
] |
2411.11446 | Yu Rong | Shihong Liu, Yu Rong, Zichen Hua, Huijie Hu | Strong Correlation between Galactic HI-to-stellar Mass Ratio And Halo Spin Explored by HI-rich Galaxies | Accepted | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Using a semi-analytic approach, we estimate halo spins for a large sample of HI-rich galaxies from the Arecibo Legacy Fast Alfa Survey and examine the correlation between HI mass fractions and halo spins. Our analysis reveals a strong correlation between halo spin and the HI-to-stellar mass ratio in both low-mass and massive galaxy samples. This finding suggests a universal formation scenario: higher halo spin reduces angular momentum loss and gas condensation, leading to lower star formation rates and weaker feedback, which in turn helps retain gas within dark matter halos. | [
{
"version": "v1",
"created": "Mon, 18 Nov 2024 10:24:56 GMT"
},
{
"version": "v2",
"created": "Sun, 22 Jun 2025 08:02:54 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 04:11:38 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Liu",
"Shihong",
""
],
[
"Rong",
"Yu",
""
],
[
"Hua",
"Zichen",
""
],
[
"Hu",
"Huijie",
""
]
] |
2411.12558 | Ismail Nejjar | Ismail Nejjar, Hao Dong, Olga Fink | Recall and Refine: A Simple but Effective Source-free Open-set Domain Adaptation Framework | Accepted at TMLR 2025 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Open-set Domain Adaptation (OSDA) aims to adapt a model from a labeled source domain to an unlabeled target domain, where novel classes - also referred to as target-private unknown classes - are present. Source-free Open-set Domain Adaptation (SF-OSDA) methods address OSDA without accessing labeled source data, making them particularly relevant under privacy constraints. However, SF-OSDA presents significant challenges due to distribution shifts and the introduction of novel classes. Existing SF-OSDA methods typically rely on thresholding the prediction entropy of a sample to identify it as either a known or unknown class, but fail to explicitly learn discriminative features for the target-private unknown classes. We propose Recall and Refine (RRDA), a novel SF-OSDA framework designed to address these limitations by explicitly learning features for target-private unknown classes. RRDA employs a two-stage process. First, we enhance the model's capacity to recognize unknown classes by training a target classifier with an additional decision boundary,guided by synthetic samples generated from target domain features. This enables the classifier to effectively separate known and unknown classes. Second, we adapt the entire model to the target domain, addressing both domain shifts and distinguishability to unknown classes. Any off-the-shelf source-free domain adaptation method (e.g. SHOT, AaD) can be seamlessly integrated into our framework at this stage. Extensive experiments on three benchmark datasets demonstrate that RRDA significantly outperforms existing SF-OSDA and OSDA methods. | [
{
"version": "v1",
"created": "Tue, 19 Nov 2024 15:18:50 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:33:44 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Nejjar",
"Ismail",
""
],
[
"Dong",
"Hao",
""
],
[
"Fink",
"Olga",
""
]
] |
2411.13547 | Shirley Kokane | Shirley Kokane, Ming Zhu, Tulika Awalgaonkar, Jianguo Zhang, Thai Hoang, Akshara Prabhakar, Zuxin Liu, Tian Lan, Liangwei Yang, Juntao Tan, Rithesh Murthy, Weiran Yao, Zhiwei Liu, Juan Carlos Niebles, Huan Wang, Shelby Heinecke, Caiming Xiong, Silivo Savarese | ToolScan: A Benchmark for Characterizing Errors in Tool-Use LLMs | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Evaluating Large Language Models (LLMs) is one of the most critical aspects of building a performant compound AI system. Since the output from LLMs propagate to downstream steps, identifying LLM errors is crucial to system performance. A common task for LLMs in AI systems is tool use. While there are several benchmark environments for evaluating LLMs on this task, they typically only give a success rate without any explanation of the failure cases. To solve this problem, we introduce TOOLSCAN, a new benchmark to identify error patterns in LLM output on tool-use tasks. Our benchmark data set comprises of queries from diverse environments that can be used to test for the presence of seven newly characterized error patterns. Using TOOLSCAN, we show that even the most prominent LLMs exhibit these error patterns in their outputs. Researchers can use these insights from TOOLSCAN to guide their error mitigation strategies. | [
{
"version": "v1",
"created": "Wed, 20 Nov 2024 18:56:22 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 01:05:54 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Kokane",
"Shirley",
""
],
[
"Zhu",
"Ming",
""
],
[
"Awalgaonkar",
"Tulika",
""
],
[
"Zhang",
"Jianguo",
""
],
[
"Hoang",
"Thai",
""
],
[
"Prabhakar",
"Akshara",
""
],
[
"Liu",
"Zuxin",
""
],
[
"Lan",
"Tian",
""
],
[
"Yang",
"Liangwei",
""
],
[
"Tan",
"Juntao",
""
],
[
"Murthy",
"Rithesh",
""
],
[
"Yao",
"Weiran",
""
],
[
"Liu",
"Zhiwei",
""
],
[
"Niebles",
"Juan Carlos",
""
],
[
"Wang",
"Huan",
""
],
[
"Heinecke",
"Shelby",
""
],
[
"Xiong",
"Caiming",
""
],
[
"Savarese",
"Silivo",
""
]
] |
2411.14133 | Advik Raj Basani | Advik Raj Basani, Xiao Zhang | GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs | 38 pages, 8 tables, 18 figures | null | null | null | cs.LG cs.AI cs.CR cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | LLMs have shown impressive capabilities across various natural language processing tasks, yet remain vulnerable to input prompts, known as jailbreak attacks, carefully designed to bypass safety guardrails and elicit harmful responses. Traditional methods rely on manual heuristics but suffer from limited generalizability. Despite being automatic, optimization-based attacks often produce unnatural prompts that can be easily detected by safety filters or require high computational costs due to discrete token optimization. In this paper, we introduce Generative Adversarial Suffix Prompter (GASP), a novel automated framework that can efficiently generate human-readable jailbreak prompts in a fully black-box setting. In particular, GASP leverages latent Bayesian optimization to craft adversarial suffixes by efficiently exploring continuous latent embedding spaces, gradually optimizing the suffix prompter to improve attack efficacy while balancing prompt coherence via a targeted iterative refinement procedure. Through comprehensive experiments, we show that GASP can produce natural adversarial prompts, significantly improving jailbreak success over baselines, reducing training times, and accelerating inference speed, thus making it an efficient and scalable solution for red-teaming LLMs. | [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 14:00:01 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 19:01:33 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Basani",
"Advik Raj",
""
],
[
"Zhang",
"Xiao",
""
]
] |
2411.14278 | Pablo Moriano | Pablo Moriano, Steven C. Hespeler, Mingyan Li, Maria Mahbub | Adaptive Anomaly Detection for Identifying Attacks in Cyber-Physical Systems: A Systematic Literature Review | 30 pages, 4 figures, 6 tables | Artificial Intelligence Review 58, 283 (2025) | 10.1007/s10462-025-11292-w | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Modern cyberattacks in cyber-physical systems (CPS) rapidly evolve and cannot be deterred effectively with most current methods which focused on characterizing past threats. Adaptive anomaly detection (AAD) is among the most promising techniques to detect evolving cyberattacks focused on fast data processing and model adaptation. AAD has been researched in the literature extensively; however, to the best of our knowledge, our work is the first systematic literature review (SLR) on the current research within this field. We present a comprehensive SLR, gathering 397 relevant papers and systematically analyzing 65 of them (47 research and 18 survey papers) on AAD in CPS studies from 2013 to 2023 (November). We introduce a novel taxonomy considering attack types, CPS application, learning paradigm, data management, and algorithms. Our analysis indicates, among other findings, that reviewed works focused on a single aspect of adaptation (either data processing or model adaptation) but rarely in both at the same time. We aim to help researchers to advance the state of the art and help practitioners to become familiar with recent progress in this field. We identify the limitations of the state of the art and provide recommendations for future research directions. | [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 16:32:02 GMT"
},
{
"version": "v2",
"created": "Fri, 03 Jan 2025 21:56:28 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Moriano",
"Pablo",
""
],
[
"Hespeler",
"Steven C.",
""
],
[
"Li",
"Mingyan",
""
],
[
"Mahbub",
"Maria",
""
]
] |
2411.14366 | ALICE Publications | ALICE Collaboration | Direct-photon production in inelastic and high-multiplicity proton-proton collisions at $\sqrt{s} =$ 13 TeV | 28 pages, 6 captioned figures, authors from page 22, published version, figures at http://alice-publications.web.cern.ch/node/11776 | Phys. Lett. B 868 (2025) 139645 | 10.1016/j.physletb.2025.139645 | CERN-EP-2024-295 | hep-ex nucl-ex | http://creativecommons.org/licenses/by/4.0/ | In this letter, we present the first measurement of direct photons at the transverse momentum of $ 1 < p_{\rm T} < 6$ GeV/$c$ at midrapidity $|\eta| < 0.8$ in inelastic and high-multiplicity proton--proton collisions at a centre-of-mass energy of $\sqrt{s} =$ 13 TeV. The fraction of virtual direct photons in the inclusive virtual photon spectrum is obtained from a fit to the dielectron invariant mass spectrum. In the limit of zero invariant mass, this fraction is equal to the relative contribution of real direct photons in the inclusive real photon spectrum. Contributions from decays of light-flavour neutral mesons are estimated using independent measurements in proton-proton collisions at the same energy and the same event class. For the first time at the LHC energies, a direct-photon signal is observed at low $p_{\rm T}$ in both inelastic and high-multiplicity event classes, with a significance of 3.2$\sigma$ and 1.9$\sigma$ in terms of standard deviations, correspondingly. The yield of direct photons in inelastic pp collisions is compared to perturbative QCD calculations. The integrated photon yield is studied as a function of charged-particle multiplicity and is compared to the results from other experiments and theoretical calculations. The results show a significant increase of direct-photon yield with charged-particle multiplicity. | [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 18:07:14 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:13:47 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"ALICE Collaboration",
"",
""
]
] |
2411.15166 | Yue-Liang Wu | Yue-Liang Wu | Gravitization Equation and Zero Energy Momentum Tensor Theorem with Cancellation Law in Gravitational Quantum Field Theory | 28 pages, improved version to be published; three sections added, including: Distinct Foundations of GQFT and Poincare Gauge Theory, GQFT in Three-Dimensional Spacetime and Its Connection to Witten's Perspective on Three-Dimensional Gravity | null | null | null | physics.gen-ph gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | We investigate the essential properties of gravitational quantum field theory (GQFT) based on spin gauge symmetry, using the general theory of quantum electrodynamics as an example. A constraint equation for the field strength of the gravigauge field is derived, serving as a gravitization equation within the spin-related gravigauge spacetime. This equation reveals how gravitational effects emerge from the non-commutative relation of the gravigauge derivative operator. By transmuting the action from gravigauge spacetime to Minkowski spacetime, we demonstrate that translational invariance results in a vanishing energy-momentum tensor in GQFT when the equations of motion are applied to all fundamental fields, including the gravigauge field. This extends the conservation law of the energy-momentum tensor in quantum field theory to a cancellation law of the energy-momentum tensor in GQFT. As a result, an equivalence between the general gravitational equation and the zero energy-momentum tensor theorem naturally arises in GQFT. Certain aspects of the Poincar\'e gauge theory are also briefly discussed. Furthermore, a GQFT incorporating the Chern-Simons action in three-dimensional spacetime is developed, based on the inhomogeneous spin gauge symmetry WS(1,2) and the global Poincar\'e symmetry PO(1,2). This framework provides a basis for exploring its connection to Witten's perspective on three-dimensional gravity. | [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 07:58:56 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:39:01 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wu",
"Yue-Liang",
""
]
] |
2411.15188 | Pete Rigas | Pete Rigas | Poisson structure of the 4-vertex model, and the higher-spin XXX chain, and Yang-Baxter algebras | 72 pages (minor reorganization of computations with Poisson bracket to the Appendix). Presentations on related topics are at: https://youtu.be/tIO0AwbQp6Q, https://youtu.be/93KhYI8Fy_M, https://youtu.be/PQBB5ENUO-g, https://youtu.be/P1EF54ZF84s, https://youtu.be/SHYwtt81Q3E, https://youtu.be/LM9xakHYAog, https://youtu.be/sdfYV6ESEe8 | null | null | null | math-ph math.MP math.PR nlin.SI | http://creativecommons.org/licenses/by/4.0/ | We implement the quantum inverse scattering method for the 4-vertex model. In comparison to previous works of the author which examined the 6-vertex, and 20-vertex, models, the 4-vertex model exhibits different characteristics, ranging from L-operators expressed in terms of projectors and Pauli matrices to algebraic and combinatorial properties, including Poisson structure and boxed plane partitions. With far fewer computations with an L-operator provided for the 4-vertex model by Bogoliubov in 2007, in comparison to those for L-operators of the 6, and 20, vertex models, from lower order expansions of the transfer matrix we derive a system of relations from the structure of operators that can be leveraged for studying characteristics of the higher-spin XXX chain in the weak finite volume limit. In comparison to quantum inverse scattering methods for the 6, and 20, vertex models which can be used to further study integrability, and exact solvability, an adaptation of such an approach for the 4-vertex model can be used to approximate, asymptotically in the weak finite volume limit, sixteen brackets which generate the Poisson structure. From explicit relations for operators of the 4-vertex transfer matrix, we conclude by discussing corresponding aspects of the Yang-Baxter algebra, which is closely related to the operators obtained from products of L-operators for approximating the transfer, and quantum monodromy, matrices. The structure of computations from L-operators of the 4-vertex model directly transfers to L-operators of the higher-spin XXX chain, revealing a similar structure of another Yang-Baxter algebra of interest. | [
{
"version": "v1",
"created": "Tue, 19 Nov 2024 08:03:17 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 20:20:24 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Rigas",
"Pete",
""
]
] |
2411.15493 | M. M. Glazov | M. A. Semina, F. Tabataba-Vakili, A. Rupp, A.S. Baimuratov, A. H\"ogele, M.M. Glazov | Excitons and trions in CrSBr bilayers | 14 pages, 7+1 figures | Phys. Rev. B 111, 205301 (2025) | 10.1103/PhysRevB.111.205301 | null | cond-mat.mes-hall cond-mat.mtrl-sci cond-mat.other | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study theoretically the neutral and charged excitons in two-dimensional semiconductors with anisotropic dispersion of charge carriers. Such a situation is realized in CrSBr-based van der Waals heterostructures. We calculate the binding energies of excitons and trions and explore their dependence on the mass ratio, dielectric screening, and interlayer distance in bilayer structures. We also address the effects of exciton-light coupling, including the radiative decay and long-range electron-hole exchange interaction, and briefly analyze correlations between the excitons and the Fermi sea of resident electrons. The estimates for CrSBr bilayers are in reasonable agreement with recent experiments. | [
{
"version": "v1",
"created": "Sat, 23 Nov 2024 08:25:04 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 04:48:29 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Semina",
"M. A.",
""
],
[
"Tabataba-Vakili",
"F.",
""
],
[
"Rupp",
"A.",
""
],
[
"Baimuratov",
"A. S.",
""
],
[
"Högele",
"A.",
""
],
[
"Glazov",
"M. M.",
""
]
] |
2411.16933 | Omar Lakkis | Marcus J. Grote and Omar Lakkis and Carina Santos | A posteriori error estimates for the wave equation with mesh change in the leapfrog method | 30 pages with appendix, 4 figures with 10 subfigures | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We derive a fully computable aposteriori error estimator for a Galerkin finite element solution of the wave equation with explicit leapfrog time-stepping. Our discrete formulation accommodates both time evolving meshes and leapfrog based local time-stepping (Diaz & Grote, 2009), which overcomes the stringent stability restriction on the time-step due to local mesh refinement. Thus we account for adaptive time-stepping with mesh change in a fully explicit time integration while retaining its efficiency. The error analysis relies on elliptic reconstructors and abstract grid transfer operators, which allows for use-defined elliptic error estimators. Numerical results using the elliptic Babu\v{s}ka-Rheinboldt estimators illustrate the optimal rate of convergence with mesh size of the aposteriori error estimator. | [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 21:01:38 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:57:53 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Grote",
"Marcus J.",
""
],
[
"Lakkis",
"Omar",
""
],
[
"Santos",
"Carina",
""
]
] |
2411.16997 | Tianfeng Wu | Tianfeng Wu, Fang Yang, Tian Cao, Ling Cheng, Yupeng Chen, Jian Song, Julian Cheng, and Zhu Han | Channel Modeling for Ultraviolet Non-Line-of-Sight Communications Incorporating an Obstacle | Accepted by IEEE Global Communications Conference (GLOBECOM) 2024. arXiv admin note: substantial text overlap with arXiv:2411.15154 | null | 10.1109/GLOBECOM52923.2024.10901379 | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing studies on ultraviolet (UV) non-line-of-sight (NLoS) channel modeling primarily focus on scenarios without any obstacle, which makes them unsuitable for small transceiver elevation angles in most cases. To address this issue, a UV NLoS channel model incorporating an obstacle was investigated in this paper, where the impacts of atmospheric scattering and obstacle reflection on UV signals were both taken into account. To validate the proposed model, we compared it to the related Monte-Carlo photon-tracing (MCPT) model that had been verified by outdoor experiments. Numerical results manifest that the path loss curves obtained by the proposed model agree well with those determined by the MCPT model, while its computation complexity is lower than that of the MCPT model. This work discloses that obstacle reflection can effectively reduce the channel path loss of UV NLoS communication systems. | [
{
"version": "v1",
"created": "Fri, 08 Nov 2024 06:29:26 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wu",
"Tianfeng",
""
],
[
"Yang",
"Fang",
""
],
[
"Cao",
"Tian",
""
],
[
"Cheng",
"Ling",
""
],
[
"Chen",
"Yupeng",
""
],
[
"Song",
"Jian",
""
],
[
"Cheng",
"Julian",
""
],
[
"Han",
"Zhu",
""
]
] |
2411.17984 | Huiyang Hu | Huiyang Hu, Peijin Wang, Hanbo Bi, Boyuan Tong, Zhaozhi Wang, Wenhui Diao, Hao Chang, Yingchao Feng, Ziqi Zhang, Yaowei Wang, Qixiang Ye, Kun Fu, Xian Sun | RS-vHeat: Heat Conduction Guided Efficient Remote Sensing Foundation Model | 19 pages, 8 figures and 10 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Remote sensing foundation models largely break away from the traditional paradigm of designing task-specific models, offering greater scalability across multiple tasks. However, they face challenges such as low computational efficiency and limited interpretability, especially when dealing with large-scale remote sensing images. To overcome these, we draw inspiration from heat conduction, a physical process modeling local heat diffusion. Building on this idea, we are the first to explore the potential of using the parallel computing model of heat conduction to simulate the local region correlations in high-resolution remote sensing images, and introduce RS-vHeat, an efficient multi-modal remote sensing foundation model. Specifically, RS-vHeat 1) applies the Heat Conduction Operator (HCO) with a complexity of $O(N^{1.5})$ and a global receptive field, reducing computational overhead while capturing remote sensing object structure information to guide heat diffusion; 2) learns the frequency distribution representations of various scenes through a self-supervised strategy based on frequency domain hierarchical masking and multi-domain reconstruction; 3) significantly improves efficiency and performance over state-of-the-art techniques across 4 tasks and 10 datasets. Compared to attention-based remote sensing foundation models, we reduce memory usage by 84\%, FLOPs by 24\% and improves throughput by 2.7 times. The code will be made publicly available. | [
{
"version": "v1",
"created": "Wed, 27 Nov 2024 01:43:38 GMT"
},
{
"version": "v2",
"created": "Fri, 07 Mar 2025 13:24:25 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 01:51:21 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Hu",
"Huiyang",
""
],
[
"Wang",
"Peijin",
""
],
[
"Bi",
"Hanbo",
""
],
[
"Tong",
"Boyuan",
""
],
[
"Wang",
"Zhaozhi",
""
],
[
"Diao",
"Wenhui",
""
],
[
"Chang",
"Hao",
""
],
[
"Feng",
"Yingchao",
""
],
[
"Zhang",
"Ziqi",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Ye",
"Qixiang",
""
],
[
"Fu",
"Kun",
""
],
[
"Sun",
"Xian",
""
]
] |
2411.18309 | Xiwei Deng | Xiwei Deng, Xianchun He, Jianfeng Bao, Yudan Zhou, Shuhui Cai, Congbo Cai, Zhong Chen | MvKeTR: Chest CT Report Generation with Multi-View Perception and Knowledge Enhancement | Accepted for publication in IEEE Journal of Biomedical and Health Informatics | null | 10.1109/JBHI.2025.3581289 | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | CT report generation (CTRG) aims to automatically generate diagnostic reports for 3D volumes, relieving clinicians' workload and improving patient care. Despite clinical value, existing works fail to effectively incorporate diagnostic information from multiple anatomical views and lack related clinical expertise essential for accurate and reliable diagnosis. To resolve these limitations, we propose a novel Multi-view perception Knowledge-enhanced TansfoRmer (MvKeTR) to mimic the diagnostic workflow of clinicians. Just as radiologists first examine CT scans from multiple planes, a Multi-View Perception Aggregator (MVPA) with view-aware attention is proposed to synthesize diagnostic information from multiple anatomical views effectively. Then, inspired by how radiologists further refer to relevant clinical records to guide diagnostic decision-making, a Cross-Modal Knowledge Enhancer (CMKE) is devised to retrieve the most similar reports based on the query volume to incorporate domain knowledge into the diagnosis procedure. Furthermore, instead of traditional MLPs, we employ Kolmogorov-Arnold Networks (KANs) as the fundamental building blocks of both modules, which exhibit superior parameter efficiency and reduced spectral bias to better capture high-frequency components critical for CT interpretation while mitigating overfitting. Extensive experiments on the public CTRG-Chest-548 K dataset demonstrate that our method outpaces prior state-of-the-art (SOTA) models across almost all metrics. The code is available at https://github.com/xiweideng/MvKeTR. | [
{
"version": "v1",
"created": "Wed, 27 Nov 2024 12:58:23 GMT"
},
{
"version": "v2",
"created": "Mon, 06 Jan 2025 10:34:37 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 00:54:18 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Deng",
"Xiwei",
""
],
[
"He",
"Xianchun",
""
],
[
"Bao",
"Jianfeng",
""
],
[
"Zhou",
"Yudan",
""
],
[
"Cai",
"Shuhui",
""
],
[
"Cai",
"Congbo",
""
],
[
"Chen",
"Zhong",
""
]
] |
2411.18581 | Ishan Bansal | Ishan Bansal, Oktay G\"unl\"uk, Richard Shapley | Parallel Token Swapping for Qubit Routing | null | null | null | null | cs.DS cs.DM quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper we study a combinatorial reconfiguration problem that involves finding an optimal sequence of swaps to move an initial configuration of tokens that are placed on the vertices of a graph to a final desired one. This problem arises as a crucial step in reducing the depth of a quantum circuit when compiling a quantum algorithm. We provide the first known constant factor approximation algorithms for the parallel token swapping problem on graph topologies that are commonly found in modern quantum computers, including cycle graphs, subdivided star graphs, and grid graphs. We also study the so-called stretch factor of a natural lower bound to the problem, which has been shown to be useful when designing heuristics for the qubit routing problem. Finally, we study the colored version of this reconfiguration problem where some tokens share the same color and are considered indistinguishable. | [
{
"version": "v1",
"created": "Wed, 27 Nov 2024 18:26:16 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:44:00 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Bansal",
"Ishan",
""
],
[
"Günlük",
"Oktay",
""
],
[
"Shapley",
"Richard",
""
]
] |
2412.01787 | Rongkun Xue | Rongkun Xue, Jinouwen Zhang, Yazhe Niu, Dazhong Shen, Bingqi Ma, Yu Liu, Jing Yang | Pretrained Reversible Generation as Unsupervised Visual Representation Learning | Accepted by ICCV 2025 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent generative models based on score matching and flow matching have significantly advanced generation tasks, but their potential in discriminative tasks remains underexplored. Previous approaches, such as generative classifiers, have not fully leveraged the capabilities of these models for discriminative tasks due to their intricate designs. We propose Pretrained Reversible Generation (PRG), which extracts unsupervised representations by reversing the generative process of a pretrained continuous generation model. PRG effectively reuses unsupervised generative models, leveraging their high capacity to serve as robust and generalizable feature extractors for downstream tasks. This framework enables the flexible selection of feature hierarchies tailored to specific downstream tasks. Our method consistently outperforms prior approaches across multiple benchmarks, achieving state-of-the-art performance among generative model based methods, including 78% top-1 accuracy on ImageNet at a resolution of 64*64. Extensive ablation studies, including out-of-distribution evaluations, further validate the effectiveness of our approach. Code is available at https://github.com/opendilab/PRG. | [
{
"version": "v1",
"created": "Fri, 29 Nov 2024 08:24:49 GMT"
},
{
"version": "v2",
"created": "Sat, 08 Mar 2025 14:13:46 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 04:26:18 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Xue",
"Rongkun",
""
],
[
"Zhang",
"Jinouwen",
""
],
[
"Niu",
"Yazhe",
""
],
[
"Shen",
"Dazhong",
""
],
[
"Ma",
"Bingqi",
""
],
[
"Liu",
"Yu",
""
],
[
"Yang",
"Jing",
""
]
] |
2412.03083 | Marco Mordacci | Giacomo Belli, Marco Mordacci and Michele Amoretti | A Scalable Quantum Neural Network for Approximate SRBB-Based Unitary Synthesis | null | null | null | null | quant-ph cs.ET cs.LG | http://creativecommons.org/licenses/by/4.0/ | In this work, a scalable quantum neural network is introduced as a means to approximate any unitary evolution through the Standard Recursive Block Basis (SRBB) and, subsequently, redesigned with a number of CNOTs asymptotically reduced by an exponential contribution. This algebraic approach to the problem of unitary synthesis exploits Lie algebras and their topological features to obtain scalable parameterizations of unitary operators. First, the original SRBB-based scalability scheme, already known in the literature only from a theoretical point of view, is reformulated for efficient algorithm implementation and complexity management. Remarkably, 2-qubit operators emerge as a special case outside the original scaling scheme. Furthermore, an algorithm is proposed to reduce the number of CNOTs, thus deriving a new implementable scaling scheme that requires only one layer of approximation. The scalable CNOT-reduced quantum neural network is implemented and its performance is assessed with a variety of different unitary matrices, both sparse and dense, up to 6 qubits via the PennyLane library. The effectiveness of the approximation is measured with different metrics in relation to two optimizers: a gradient-based method and the Nelder-Mead method. The approximate CNOT-reduced SRBB-based synthesis algorithm is also tested on real hardware and compared with other valid approximation and decomposition methods available in the literature. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 07:21:23 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:43:45 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Belli",
"Giacomo",
""
],
[
"Mordacci",
"Marco",
""
],
[
"Amoretti",
"Michele",
""
]
] |
2412.03112 | Weikang Tang | W. Tang, M. Hoelzl, M. Lehnen, D. Hu, F. J. Artola, P. Halldestam, P. Heinrich, S. Jachmich, E. Nardon, G. Papp, A. Patel, the ASDEX Upgrade Team, the EUROfusion Tokamak Exploitation Team and the JOREK Team | Non-linear MHD modelling of shattered pellet injection in ASDEX Upgrade | null | null | null | null | physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Shattered pellet injection (SPI) is selected for the disruption mitigation system in ITER, due to deeper penetration, expected assimilation efficiency and prompt material delivery. This article describes non-linear magnetohydrodynamic (MHD) simulations of SPI in the ASDEX Upgrade tokamak to test the mitigation efficiency of different injection parameters for neon-doped deuterium pellets using the JOREK code. The simulations are executed as fluid simulations, while additional marker particles are used to evolve the charge state distribution and radiation property of impurities based on OpenADAS atomic data, i.e., a collisional-radiative model is used. Neon fraction scans between 0 - 10% are performed. Numerical results show that the thermal quench (TQ) occurs in two stages. In the first stage, approximately half of the thermal energy is abruptly lost, primarily through convective and conductive transport in the stochastic fields. This stage is relatively independent of the neon fraction. In the second stage, where the majority of the remaining thermal energy is lost, radiation plays a dominant role. In case of pure deuterium injection, this second stage may not occur at all. A larger fraction ($\sim $20%) of the total material in the pellet is assimilated in the plasma for low neon fraction pellets ($\leq 0.12\%$) due to the full thermal collapse of the plasma occurring later than in high neon fraction scenarios. Nevertheless, the total number of assimilated neon atoms increases with increasing neon fraction. The effects of fragment size and penetration speed are then numerically studied, showing that slower and smaller fragments promote edge cooling and the formation of a cold front. Faster fragments result in shorter TQ duration and higher assimilation as they reach the hotter plasma regions quicker. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 08:22:47 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 13:11:35 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Tang",
"W.",
""
],
[
"Hoelzl",
"M.",
""
],
[
"Lehnen",
"M.",
""
],
[
"Hu",
"D.",
""
],
[
"Artola",
"F. J.",
""
],
[
"Halldestam",
"P.",
""
],
[
"Heinrich",
"P.",
""
],
[
"Jachmich",
"S.",
""
],
[
"Nardon",
"E.",
""
],
[
"Papp",
"G.",
""
],
[
"Patel",
"A.",
""
],
[
"Team",
"the ASDEX Upgrade",
""
],
[
"Team",
"the EUROfusion Tokamak Exploitation",
""
],
[
"Team",
"the JOREK",
""
]
] |
2412.03359 | Chengwei Hu | Chengwei Hu, Jianhui Zheng, Yancheng He, Hangyu Guo, Junguang Jiang, Han Zhu, Kai Sun, Yuning Jiang, Wenbo Su, Bo Zheng | WiS Platform: Enhancing Evaluation of LLM-Based Multi-Agent Systems Through Game-Based Analysis | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in autonomous multi-agent systems (MAS) based on large language models (LLMs) have enhanced the application scenarios and improved the capability of LLMs to handle complex tasks. Despite demonstrating effectiveness, existing studies still evidently struggle to evaluate, analysis, and reproducibility of LLM-based MAS. In this paper, to facilitate the research on LLM-based MAS, we introduce an open, scalable, and real-time updated platform for accessing and analyzing the LLM-based MAS based on the games Who is Spy?" (WiS). Our platform is featured with three main worths: (1) a unified model evaluate interface that supports models available on Hugging Face; (2) real-time updated leaderboard for model evaluation; (3) a comprehensive evaluation covering game-winning rates, attacking, defense strategies, and reasoning of LLMs. To rigorously test WiS, we conduct extensive experiments coverage of various open- and closed-source LLMs, we find that different agents exhibit distinct and intriguing behaviors in the game. The experimental results demonstrate the effectiveness and efficiency of our platform in evaluating LLM-based MAS. Our platform and its documentation are publicly available at https://whoisspy.ai/. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 14:45:09 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:55:53 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Hu",
"Chengwei",
""
],
[
"Zheng",
"Jianhui",
""
],
[
"He",
"Yancheng",
""
],
[
"Guo",
"Hangyu",
""
],
[
"Jiang",
"Junguang",
""
],
[
"Zhu",
"Han",
""
],
[
"Sun",
"Kai",
""
],
[
"Jiang",
"Yuning",
""
],
[
"Su",
"Wenbo",
""
],
[
"Zheng",
"Bo",
""
]
] |
2412.03380 | Ibrahim Ekren | Ibrahim Ekren and Sergey Nadtochiy | Consistency of MLE in partially observed diffusion models on a torus | null | null | null | null | math.ST math.PR stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we consider a general partially observed diffusion model with periodic coefficients and with non-degenerate diffusion component. The coefficients of such a model depend on an unknown (static and deterministic) parameter which needs to be estimated based on the observed component of the diffusion process. We show that, given enough regularity of the diffusion coefficients, a maximum likelihood estimator of the unknown parameter converges to the true parameter value as the sample size grows to infinity. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 15:07:33 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Dec 2024 19:25:13 GMT"
},
{
"version": "v3",
"created": "Tue, 14 Jan 2025 19:57:24 GMT"
},
{
"version": "v4",
"created": "Thu, 26 Jun 2025 17:02:54 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Ekren",
"Ibrahim",
""
],
[
"Nadtochiy",
"Sergey",
""
]
] |
2412.03495 | Elvira Bilokon | Elvira Bilokon, Valeriia Bilokon, Dusty R. Lindberg, Lev Kaplan, Andrii Sotnikov, Denys I. Bondar | Few-fermion resonant tunneling and underbarrier trapping in asymmetric potentials | 13 pages, 9 figures | Commun Phys 8, 259 (2025) | 10.1038/s42005-025-02189-9 | null | quant-ph cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding quantum tunneling in many-body systems is crucial for advancing quantum technologies and nanoscale device design. Despite extensive studies of quantum tunneling, the role of interactions in determining directional transport through asymmetric barriers in discrete quantum systems remains unclear. Here we show that noninteracting fermions exhibit symmetric tunneling probabilities regardless of barrier orientation, while inter-particle interactions break this symmetry and create pronounced asymmetric tunneling behavior. We explore the dependence of tunneling behavior on the initial spin configurations of two spin-1/2 fermions: spin-triplet states preserve tunneling symmetry, while spin-singlet states show strong asymmetry. We identify regimes where interactions mediate tunneling through under-barrier resonant trapping and enhance tunneling via many-body resonant tunneling -- a phenomenon arising solely from inter-particle interactions and being fundamentally different from traditional single-particle resonant tunneling. Our results may be applied to the design of nanoscale devices with tailored transport properties, such as diodes and memristors. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 17:36:07 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:45:57 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Bilokon",
"Elvira",
""
],
[
"Bilokon",
"Valeriia",
""
],
[
"Lindberg",
"Dusty R.",
""
],
[
"Kaplan",
"Lev",
""
],
[
"Sotnikov",
"Andrii",
""
],
[
"Bondar",
"Denys I.",
""
]
] |
2412.03559 | Michael Schlosser | Michael J. Schlosser | A family of orthogonal functions on the unit circle and a new multilateral matrix inverse | 17 pp.; dedicated to Tom Koornwinder on the occasion of his 80th birthday | null | null | null | math.CA math.CO math.QA | http://creativecommons.org/licenses/by/4.0/ | Using Bailey's very-well-poised $_6\psi_6$ summation, we show that a specific sequence of well-poised bilateral basic hypergeometric $_3\psi_3$ series form a family of orthogonal functions on the unit circle. We further extract a bilateral matrix inverse from Dougall's ${}_2H_2$ summation which we use, in combination with the Pfaff--Saalsch\"utz summation, to derive a summation for a particular bilateral hypergeometric $_3H_3$ series. We finally provide multivariate extensions of the bilateral matrix inverse and the $_3H_3$ summation in the setting of hypergeometric series associated to the root system $A_r$. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 18:53:52 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 19:32:55 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Schlosser",
"Michael J.",
""
]
] |
2412.03934 | Yifan Lu | Yifan Lu, Xuanchi Ren, Jiawei Yang, Tianchang Shen, Zhangjie Wu, Jun Gao, Yue Wang, Siheng Chen, Mike Chen, Sanja Fidler, Jiahui Huang | InfiniCube: Unbounded and Controllable Dynamic 3D Driving Scene Generation with World-Guided Video Models | ICCV 2025. Project Page: https://research.nvidia.com/labs/toronto-ai/infinicube/ | null | null | null | cs.CV cs.AI cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present InfiniCube, a scalable method for generating unbounded dynamic 3D driving scenes with high fidelity and controllability. Previous methods for scene generation either suffer from limited scales or lack geometric and appearance consistency along generated sequences. In contrast, we leverage the recent advancements in scalable 3D representation and video models to achieve large dynamic scene generation that allows flexible controls through HD maps, vehicle bounding boxes, and text descriptions. First, we construct a map-conditioned sparse-voxel-based 3D generative model to unleash its power for unbounded voxel world generation. Then, we re-purpose a video model and ground it on the voxel world through a set of carefully designed pixel-aligned guidance buffers, synthesizing a consistent appearance. Finally, we propose a fast feed-forward approach that employs both voxel and pixel branches to lift the dynamic videos to dynamic 3D Gaussians with controllable objects. Our method can generate controllable and realistic 3D driving scenes, and extensive experiments validate the effectiveness and superiority of our model. | [
{
"version": "v1",
"created": "Thu, 05 Dec 2024 07:32:20 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:10:09 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Lu",
"Yifan",
""
],
[
"Ren",
"Xuanchi",
""
],
[
"Yang",
"Jiawei",
""
],
[
"Shen",
"Tianchang",
""
],
[
"Wu",
"Zhangjie",
""
],
[
"Gao",
"Jun",
""
],
[
"Wang",
"Yue",
""
],
[
"Chen",
"Siheng",
""
],
[
"Chen",
"Mike",
""
],
[
"Fidler",
"Sanja",
""
],
[
"Huang",
"Jiahui",
""
]
] |
2412.04283 | Spiros Cotsakis | Spiros Cotsakis | Structural stability and general relativity | v2: 87 pages, various additions and improvements, slightly changed title, matches published version | Universe 11(7) (2025) 209 | 10.3390/universe11070209 | null | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We review recent developments in structural stability as applied to key topics in general relativity. For a nonlinear dynamical system arising from the Einstein equations by a symmetry reduction, bifurcation theory fully characterizes the set of all stable perturbations of the system, known as the `versal unfolding'. This construction yields a comprehensive classification of qualitatively distinct solutions and their metamorphoses into new topological forms, parametrized by the codimension of the bifurcation in each case. We illustrate these ideas through bifurcations in the simplest Friedmann models, the Oppenheimer-Snyder black hole, the evolution of causal geodesic congruences in cosmology and black-hole spacetimes, crease flow on event horizons, and the Friedmann-Lema\^itre equations. Finally, we list open problems and briefly discuss emerging aspects such as partial differential equation stability of versal families, the general relativity landscape, and potential connections between gravitational versal unfoldings and those of the Maxwell, Dirac, and Schr\"{o}dinger equations. | [
{
"version": "v1",
"created": "Thu, 05 Dec 2024 16:03:40 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:30:49 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Cotsakis",
"Spiros",
""
]
] |
2412.04292 | Zhenglin Huang | Zhenglin Huang, Jinwei Hu, Xiangtai Li, Yiwei He, Xingyu Zhao, Bei Peng, Baoyuan Wu, Xiaowei Huang, Guangliang Cheng | SIDA: Social Media Image Deepfake Detection, Localization and Explanation with Large Multimodal Model | This version revises and corrects the metric calculations in the tables | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid advancement of generative models in creating highly realistic images poses substantial risks for misinformation dissemination. For instance, a synthetic image, when shared on social media, can mislead extensive audiences and erode trust in digital content, resulting in severe repercussions. Despite some progress, academia has not yet created a large and diversified deepfake detection dataset for social media, nor has it devised an effective solution to address this issue. In this paper, we introduce the Social media Image Detection dataSet (SID-Set), which offers three key advantages: (1) extensive volume, featuring 300K AI-generated/tampered and authentic images with comprehensive annotations, (2) broad diversity, encompassing fully synthetic and tampered images across various classes, and (3) elevated realism, with images that are predominantly indistinguishable from genuine ones through mere visual inspection. Furthermore, leveraging the exceptional capabilities of large multimodal models, we propose a new image deepfake detection, localization, and explanation framework, named SIDA (Social media Image Detection, localization, and explanation Assistant). SIDA not only discerns the authenticity of images, but also delineates tampered regions through mask prediction and provides textual explanations of the model's judgment criteria. Compared with state-of-the-art deepfake detection models on SID-Set and other benchmarks, extensive experiments demonstrate that SIDA achieves superior performance among diversified settings. The code, model, and dataset will be released. | [
{
"version": "v1",
"created": "Thu, 05 Dec 2024 16:12:25 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Mar 2025 11:03:16 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 21:47:50 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Huang",
"Zhenglin",
""
],
[
"Hu",
"Jinwei",
""
],
[
"Li",
"Xiangtai",
""
],
[
"He",
"Yiwei",
""
],
[
"Zhao",
"Xingyu",
""
],
[
"Peng",
"Bei",
""
],
[
"Wu",
"Baoyuan",
""
],
[
"Huang",
"Xiaowei",
""
],
[
"Cheng",
"Guangliang",
""
]
] |
2412.04515 | Pete Rigas | Pete Rigas | Intertwining vectors, and Boltzmann weight matrices, of a Solid-on-Solid model from the 20-vertex model | Template version (30 pages). Related topics of discussion are at: https://youtu.be/tIO0AwbQp6Q, https://youtu.be/93KhYI8Fy_M, https://youtu.be/PQBB5ENUO-g, https://youtu.be/P1EF54ZF84s, https://youtu.be/SHYwtt81Q3E, https://youtu.be/sdfYV6ESEe8, https://youtu.be/LM9xakHYAog | null | null | null | math-ph math.MP math.PR | http://creativecommons.org/licenses/by/4.0/ | We initiate a new study on the correspondence between the 20-vertex model and a SOS (Solid-on-Solid) model. In comparison to two previous works of the author in 2024 which characterized properties of the transfer, and quantum monodromy, matrices of the 20-vertex model from the perspective of the quantum inverse scattering method, in addition to the structure of nonlocal correlations, the forthcoming approach is an adaptation of a study on the rational 7-vertex model. For the rational 7-vertex model, Antoenneko and Valinevich demonstrated that the intertwining vectors can be used to transform the R-matrix of a vertex model into the Boltzman weight matrix of an SOS model. To further develop perspectives on classes of higher dimensional SOS models, by leveraging previous computations with L-operators due to the author, and by also manipulating q-exponentials, and other related factors from a factorization of the universal R-matrix due to Boos et al, analogs of intertwining vectors for the 20- vertex model can be obtained. In comparison to the system of equations that have been obtained for the rational 7-vertex model, the system that one obtains for the 20-vertex model consists of 9 equations for each entry of the R-matrix. Furthermore, from the fact that the 20-vertex model contains more possible configurations in the sample space than the rational 7-vertex model does, the Boltzman weight matrix of the corresponding SOS model asymptotically depends upon several contributions, a few of which include tensor products of basis elements, and spectral parameters over the triangular lattice. | [
{
"version": "v1",
"created": "Thu, 05 Dec 2024 03:51:49 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 20:10:25 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Rigas",
"Pete",
""
]
] |
2412.04833 | Peiyan Hu | Peiyan Hu, Rui Wang, Xiang Zheng, Tao Zhang, Haodong Feng, Ruiqi Feng, Long Wei, Yue Wang, Zhi-Ming Ma, Tailin Wu | Wavelet Diffusion Neural Operator | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Simulating and controlling physical systems described by partial differential equations (PDEs) are crucial tasks across science and engineering. Recently, diffusion generative models have emerged as a competitive class of methods for these tasks due to their ability to capture long-term dependencies and model high-dimensional states. However, diffusion models typically struggle with handling system states with abrupt changes and generalizing to higher resolutions. In this work, we propose Wavelet Diffusion Neural Operator (WDNO), a novel PDE simulation and control framework that enhances the handling of these complexities. WDNO comprises two key innovations. Firstly, WDNO performs diffusion-based generative modeling in the wavelet domain for the entire trajectory to handle abrupt changes and long-term dependencies effectively. Secondly, to address the issue of poor generalization across different resolutions, which is one of the fundamental tasks in modeling physical systems, we introduce multi-resolution training. We validate WDNO on five physical systems, including 1D advection equation, three challenging physical systems with abrupt changes (1D Burgers' equation, 1D compressible Navier-Stokes equation and 2D incompressible fluid), and a real-world dataset ERA5, which demonstrates superior performance on both simulation and control tasks over state-of-the-art methods, with significant improvements in long-term and detail prediction accuracy. Remarkably, in the challenging context of the 2D high-dimensional and indirect control task aimed at reducing smoke leakage, WDNO reduces the leakage by 78% compared to the second-best baseline. The code can be found at https://github.com/AI4Science-WestlakeU/wdno.git. | [
{
"version": "v1",
"created": "Fri, 06 Dec 2024 07:56:25 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Apr 2025 12:34:20 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 13:39:47 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Hu",
"Peiyan",
""
],
[
"Wang",
"Rui",
""
],
[
"Zheng",
"Xiang",
""
],
[
"Zhang",
"Tao",
""
],
[
"Feng",
"Haodong",
""
],
[
"Feng",
"Ruiqi",
""
],
[
"Wei",
"Long",
""
],
[
"Wang",
"Yue",
""
],
[
"Ma",
"Zhi-Ming",
""
],
[
"Wu",
"Tailin",
""
]
] |
2412.05217 | Lorenzo Portinale | Peter Gladbach, Jan Maas, Lorenzo Portinale | Stochastic homogenisation of nonlinear minimum-cost flow problems | null | null | null | null | math.AP math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper deals with the large-scale behaviour of nonlinear minimum-cost flow problems on random graphs. In such problems, a random nonlinear cost functional is minimised among all flows (discrete vector-fields) with a prescribed net flux through each vertex. On a stationary random graph embedded in $\mathbb{R}^d$, our main result asserts that these problems converge, in the large-scale limit, to a continuous minimisation problem where an effective cost functional is minimised among all vector fields with prescribed divergence. Our main result is formulated using $\Gamma$-convergence and applies to multi-species problems. The proof employs the blow-up technique by Fonseca and M\"uller in a discrete setting. One of the main challenges overcome is the construction of the homogenised energy density on random graphs without a periodic structure. | [
{
"version": "v1",
"created": "Fri, 06 Dec 2024 17:48:10 GMT"
},
{
"version": "v2",
"created": "Wed, 09 Apr 2025 19:42:45 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 15:23:25 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Gladbach",
"Peter",
""
],
[
"Maas",
"Jan",
""
],
[
"Portinale",
"Lorenzo",
""
]
] |
2412.05782 | Yongjie Deng | Yongjie Deng, Tianbo Liu, Ya-jin Zhou | The Sivers asymmetry of vector meson production in semi-inclusive deep inelastic scattering | null | Chinese Phys. Lett. 42 (2025) 061301 | 10.1088/0256-307X/42/6/061301 | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | The transverse single-spin asymmetry for $\rho^0$ production in semi-inclusive deep inelastic scattering was recently reported by the COMPASS Collaboration. Using the Sivers functions extracted from pion and kaon productions, we perform a calculation of the Sivers asymmetry within the transverse momentum dependent factorization. Our results are consistent with the COMPASS data, confirming the universality of the Sivers functions within current experimental uncertainties. While various global analyses of Sivers functions can equally well describe the current data, we obtain very different predictions on the Sivers asymmetry of $\rho$ and $K^*$ productions at electron-ion colliders, which therefore are expected to provide further constraints. | [
{
"version": "v1",
"created": "Sun, 08 Dec 2024 02:28:27 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 01:31:44 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Deng",
"Yongjie",
""
],
[
"Liu",
"Tianbo",
""
],
[
"Zhou",
"Ya-jin",
""
]
] |
2412.06934 | Victor Hugo Nagahama | Victor Hugo Nagahama, James Sweeney, Niamh Cahill | A Scalable Bayesian Spatiotemporal Model for Water Level Predictions using a Nearest Neighbor Gaussian Process Approach | null | null | null | null | stat.AP | http://creativecommons.org/licenses/by/4.0/ | Obtaining accurate water level predictions are essential for water resource management and implementing flood mitigation strategies. Several data-driven models can be found in the literature. However, there has been limited research with regard to addressing the challenges posed by large spatio-temporally referenced hydrological datasets, in particular, the challenges of maintaining predictive performance and uncertainty quantification. Gaussian Processes (GPs) are commonly used to capture complex space-time interactions. However, GPs are computationally expensive and suffer from poor scaling as the number of locations increases due to required covariance matrix inversions. To overcome the computational bottleneck, the Nearest Neighbor Gaussian Process (NNGP) introduces a sparse precision matrix providing scalability without having to make inferential compromises. In this work we introduce an innovative model in the hydrology field, specifically designed to handle large datasets consisting of a large number of spatial points across multiple hydrological basins, with daily observations over an extended period. We investigate the application of a Bayesian spatiotemporal NNGP model to a rich dataset of daily water levels of rivers located in Ireland. The dataset comprises a network of 301 stations situated in various basins across Ireland, measured over a period of 90 days. The proposed approach allows for prediction of water levels at future time points, as well as the prediction of water levels at unobserved locations through spatial interpolation, while maintaining the benefits of the Bayesian approach, such as uncertainty propagation and quantification. Our findings demonstrate that the proposed model outperforms competing approaches in terms of accuracy and precision. | [
{
"version": "v1",
"created": "Mon, 09 Dec 2024 19:24:18 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:55:16 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Nagahama",
"Victor Hugo",
""
],
[
"Sweeney",
"James",
""
],
[
"Cahill",
"Niamh",
""
]
] |
2412.07122 | Hai-Zhou Lu | Lei Shi, Xiaoxiong Liu, C. M. Wang, Tianyu Liu, Hai-Zhou Lu, and X. C. Xie | Quantum oscillation in Hopf-link semimetals | 6 pages, 3 figures, 2 tables | Phys. Rev. B 111, L201103 (2025) | 10.1103/PhysRevB.111.L201103 | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | Since the discovery of the relation between the Chern number and quantum Hall effect, searching for observables of topological invariants has been an intriguing topic. Topological Hopf-link semimetals have attracted tremendous interest, in which the conduction and valence energy bands touch at linked nodal lines. However, it is challenging to identify this sophisticated topology. We propose to use the quantum oscillation in strong magnetic fields to probe the Hopf links. For a generic model of Hopf-link semimetal that captures the linked-trivial phase transition, we figure out the phase shifts of oscillation for all Fermi pockets in all magnetic-field directions, by presenting self-consistent results from the Fermi surface tomography, Landau fan diagram, and electrical resistivity. As the magnetic field is rotated, the phase shifts exhibit a unique pattern, which could help to identify Hopf links in real materials, such as those in Li$_2$NaN. | [
{
"version": "v1",
"created": "Tue, 10 Dec 2024 02:30:11 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:57:08 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Shi",
"Lei",
""
],
[
"Liu",
"Xiaoxiong",
""
],
[
"Wang",
"C. M.",
""
],
[
"Liu",
"Tianyu",
""
],
[
"Lu",
"Hai-Zhou",
""
],
[
"Xie",
"X. C.",
""
]
] |
2412.07229 | Wan Jiang | Wan Jiang, He Wang, Xin Zhang, Dan Guo, Zhaoxin Fan, Yunfeng Diao, Richang Hong | Moderating the Generalization of Score-based Generative Model | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Score-based Generative Models (SGMs) have demonstrated remarkable generalization abilities, e.g. generating unseen, but natural data. However, the greater the generalization power, the more likely the unintended generalization, and the more dangerous the abuse. Research on moderated generalization in SGMs remains limited. To fill this gap, we first examine the current 'gold standard' in Machine Unlearning (MU), i.e., re-training the model after removing the undesirable training data, and find it does not work in SGMs. Further analysis of score functions reveals that the MU 'gold standard' does not alter the original score function, which explains its ineffectiveness. Based on this insight, we propose the first Moderated Score-based Generative Model (MSGM), which introduces a novel score adjustment strategy that redirects the score function away from undesirable data during the continuous-time stochastic differential equation process. Extensive experimental results demonstrate that MSGM significantly reduces the likelihood of generating undesirable content while preserving high visual quality for normal image generation. Albeit designed for SGMs, MSGM is a general and flexible MU framework that is compatible with diverse diffusion architectures (SGM and DDPM) and training strategies (re-training and fine-tuning), and enables zero-shot transfer of the pre-trained models to downstream tasks, e.g. image inpainting and reconstruction. The code will be shared upon acceptance. | [
{
"version": "v1",
"created": "Tue, 10 Dec 2024 06:41:18 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 12:06:00 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Jiang",
"Wan",
""
],
[
"Wang",
"He",
""
],
[
"Zhang",
"Xin",
""
],
[
"Guo",
"Dan",
""
],
[
"Fan",
"Zhaoxin",
""
],
[
"Diao",
"Yunfeng",
""
],
[
"Hong",
"Richang",
""
]
] |
2412.09587 | Constantine Lignos | Chester Palen-Michel, Maxwell Pickering, Maya Kruse, Jonne S\"alev\"a, and Constantine Lignos | OpenNER 1.0: Standardized Open-Access Named Entity Recognition Datasets in 50+ Languages | Under review | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present OpenNER 1.0, a standardized collection of openly-available named entity recognition (NER) datasets. OpenNER contains 36 NER corpora that span 52 languages, human-annotated in varying named entity ontologies. We correct annotation format issues, standardize the original datasets into a uniform representation with consistent entity type names across corpora, and provide the collection in a structure that enables research in multilingual and multi-ontology NER. We provide baseline results using three pretrained multilingual language models and two large language models to compare the performance of recent models and facilitate future research in NER. We find that no single model is best in all languages and that significant work remains to obtain high performance from LLMs on the NER task. | [
{
"version": "v1",
"created": "Thu, 12 Dec 2024 18:55:53 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 17:51:40 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Palen-Michel",
"Chester",
""
],
[
"Pickering",
"Maxwell",
""
],
[
"Kruse",
"Maya",
""
],
[
"Sälevä",
"Jonne",
""
],
[
"Lignos",
"Constantine",
""
]
] |
2412.09925 | Andy Yang | Andy Yang, Lena Strobl, David Chiang, Dana Angluin | Simulating Hard Attention Using Soft Attention | 19 pages | null | null | null | cs.LG cs.CL cs.FL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study conditions under which transformers using soft attention can simulate hard attention, that is, effectively focus all attention on a subset of positions. First, we examine several subclasses of languages recognized by hard-attention transformers, which can be defined in variants of linear temporal logic. We demonstrate how soft-attention transformers can compute formulas of these logics using unbounded positional embeddings or temperature scaling. Second, we demonstrate how temperature scaling allows softmax transformers to simulate general hard-attention transformers, using a temperature that depends on the minimum gap between the maximum attention scores and other attention scores. | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 07:27:42 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 13:41:24 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Yang",
"Andy",
""
],
[
"Strobl",
"Lena",
""
],
[
"Chiang",
"David",
""
],
[
"Angluin",
"Dana",
""
]
] |
2412.09967 | Zhaoxuan Zhang | Zhaoxuan Zhang, Le Zhang, Pengjie Zhang | Reconstructing the Thermal Sunyaev Zeldovich Power Spectrum from Planck using the ABS Method | 27 pages, 10 figures | null | null | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study employs a novel approach for reconstructing the thermal Sunyaev-Zeldovich (tSZ) effect power spectrum from Planck data using the Analytical Blind Separation (ABS) method. The ABS method improves the recovery of weak signals, by applying eigenmode exclusion for low signal-to-noise ratio regimes and introducing a shift parameter to stabilize calculations. Validation through simulated Planck data demonstrates the robustness of ABS in reconstructing the tSZ power spectrum, even under challenging conditions. In the analysis of the {\it Planck} PR3 full-mission data, ABS shows lower amplitudes at $\ell \gtrsim 300$ compared to the {\it Planck} 2015 band powers using the MILCA and NILC foreground cleaning methods. In our analysis, we include or exclude the trispectrum contribution to the statistical uncertainty to enable comparison with previous studies and to test the robustness of our results. When the trispectrum contribution is included, and after marginalizing over residual foreground components, we find that the overall amplitude of the tSZ power spectrum is 34\% lower than the ``Planck 2015'' best-fit amplitude, 24\% lower than the ``Battaglia 2012'' model, and 12\% lower than the ``Bolliet 2018'' best-fit amplitude. These differences correspond to $3.0\sigma$, $1.78\sigma$, and $0.75\sigma$, respectively, in terms of the associated statistical uncertainties. Across all cases, the tSZ band power remains unaffected by template choice. These findings highlight the potential of the ABS method as a promising alternative for tSZ power spectrum analysis. | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 08:51:15 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:43:09 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhang",
"Zhaoxuan",
""
],
[
"Zhang",
"Le",
""
],
[
"Zhang",
"Pengjie",
""
]
] |
2412.10028 | Chang-Bin Zhang | Chang-Bin Zhang, Yujie Zhong, Kai Han | Mr. DETR++: Instructive Multi-Route Training for Detection Transformers with Mixture-of-Experts | Under review. Extended version of our CVPR 2025 paper, see arXiv:2412.10028v3 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing methods enhance the training of detection transformers by incorporating an auxiliary one-to-many assignment. In this work, we treat the model as a multi-task framework, simultaneously performing one-to-one and one-to-many predictions. We investigate the roles of each component in the transformer decoder across these two training targets, including self-attention, cross-attention, and feed-forward network. Our empirical results demonstrate that any independent component in the decoder can effectively learn both targets simultaneously, even when other components are shared. This finding leads us to propose a multi-route training mechanism, featuring a primary route for one-to-one prediction and two auxiliary training routes for one-to-many prediction. We propose a novel instructive self-attention mechanism, integrated into the first auxiliary route, which dynamically and flexibly guides object queries for one-to-many prediction. For the second auxiliary route, we introduce a route-aware Mixture-of-Experts (MoE) to facilitate knowledge sharing while mitigating potential conflicts between routes. Additionally, we apply an MoE to low-scale features in the encoder, optimizing the balance between efficiency and effectiveness. The auxiliary routes are discarded during inference. We conduct extensive experiments across various object detection baselines, achieving consistent improvements as demonstrated in Fig. 1. Our method is highly flexible and can be readily adapted to other tasks. To demonstrate its versatility, we conduct experiments on both instance segmentation and panoptic segmentation, further validating its effectiveness. Project page: https://visual-ai.github.io/mrdetr/ | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 10:39:27 GMT"
},
{
"version": "v2",
"created": "Tue, 01 Apr 2025 08:32:36 GMT"
},
{
"version": "v3",
"created": "Wed, 02 Apr 2025 05:38:32 GMT"
},
{
"version": "v4",
"created": "Thu, 26 Jun 2025 15:06:14 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhang",
"Chang-Bin",
""
],
[
"Zhong",
"Yujie",
""
],
[
"Han",
"Kai",
""
]
] |
2412.10245 | Arthur Chatton | Arthur Chatton and Michael Schomaker and Miguel-Angel Luque-Fernandez and Robert W. Platt and Mireille E. Schnitzer | Is checking for sequential positivity violations getting you down? Try sPoRT! | 10 pages, 4 Tables, 2 Figures. Accepted for publication in Epidemiology | null | null | null | stat.ME | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Background: Sequential positivity is often a necessary assumption for drawing causal inferences, such as through marginal structural modeling. Unfortunately, verification of this assumption can be challenging because it usually relies on multiple parametric propensity score models, unlikely all correctly specified. Therefore, we propose a new algorithm, called sequential Positivity Regression Tree (sPoRT), to overcome this issue and identify the subgroups found to be violating this assumption, allowing for insights about the nature of the violations and potential solutions.
Methods: We present different versions of sPoRT based on either stratifying or pooling over time under static or dynamic treatment strategies. This methodological development was motivated by a real-life application of the impact of the timing of initiation of HIV treatment with and without smoothing over time, which we also use to demonstrate the method.
Results: The illustration of sPoRT demonstrates its easy use and the interpretability of the results for applied epidemiologists. Furthermore, an R notebook showing how to use sPoRT in practice is available at github.com/ArthurChatton/sPoRT-notebook.
Conclusions: The sPoRT algorithm provides interpretable subgroups violating the sequential positivity violation, allowing patterns and trends in the confounders to be easily identified. We finally provided practical implications and recommendations when positivity violations are identified. | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 16:14:27 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2025 20:21:00 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 14:41:55 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Chatton",
"Arthur",
""
],
[
"Schomaker",
"Michael",
""
],
[
"Luque-Fernandez",
"Miguel-Angel",
""
],
[
"Platt",
"Robert W.",
""
],
[
"Schnitzer",
"Mireille E.",
""
]
] |
2412.10943 | Yiping Li | Zhangjun Zhou, Yiping Li, Chunlin Zhong, Jianuo Huang, Jialun Pei, Hua Li and He Tang | Rethinking Detecting Salient and Camouflaged Objects in Unconstrained Scenes | 18 pages, 11 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While the human visual system employs distinct mechanisms to perceive salient and camouflaged objects, existing models struggle to disentangle these tasks. Specifically, salient object detection (SOD) models frequently misclassify camouflaged objects as salient, while camouflaged object detection (COD) models conversely misinterpret salient objects as camouflaged. We hypothesize that this can be attributed to two factors: (i) the specific annotation paradigm of current SOD and COD datasets, and (ii) the lack of explicit attribute relationship modeling in current models. Prevalent SOD/COD datasets enforce a mutual exclusivity constraint, assuming scenes contain either salient or camouflaged objects, which poorly aligns with the real world. Furthermore, current SOD/COD methods are primarily designed for these highly constrained datasets and lack explicit modeling of the relationship between salient and camouflaged objects. In this paper, to promote the development of unconstrained salient and camouflaged object detection, we construct a large-scale dataset, USC12K, which features comprehensive labels and four different scenes that cover all possible logical existence scenarios of both salient and camouflaged objects. To explicitly model the relationship between salient and camouflaged objects, we propose a model called USCNet, which introduces two distinct prompt query mechanisms for modeling inter-sample and intra-sample attribute relationships. Additionally, to assess the model's ability to distinguish between salient and camouflaged objects, we design an evaluation metric called CSCS. The proposed method achieves state-of-the-art performance across all scenes in various metrics. The code and dataset will be available at https://github.com/ssecv/USCNet. | [
{
"version": "v1",
"created": "Sat, 14 Dec 2024 19:37:17 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:53:27 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhou",
"Zhangjun",
""
],
[
"Li",
"Yiping",
""
],
[
"Zhong",
"Chunlin",
""
],
[
"Huang",
"Jianuo",
""
],
[
"Pei",
"Jialun",
""
],
[
"Li",
"Hua",
""
],
[
"Tang",
"He",
""
]
] |
2412.11521 | Andrea Perin | Andrea Perin and Stephane Deny | On the Ability of Deep Networks to Learn Symmetries from Data: A Neural Kernel Theory | JMLR accepted version, including an extension of the theory to general finite groups (including non-abelian groups) | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Symmetries (transformations by group actions) are present in many datasets, and leveraging them holds considerable promise for improving predictions in machine learning. In this work, we aim to understand when and how deep networks -- with standard architectures trained in a standard, supervised way -- learn symmetries from data. Inspired by real-world scenarios, we study a classification paradigm where data symmetries are only partially observed during training: some classes include all transformations of a cyclic group, while others -- only a subset. In the infinite-width limit, where kernel analogies apply, we derive a neural kernel theory of symmetry learning. The group-cyclic nature of the dataset allows us to analyze the Gram matrix of neural kernels in the Fourier domain; here we find a simple characterization of the generalization error as a function of class separation (signal) and class-orbit density (noise). This characterization reveals that generalization can only be successful when the local structure of the data prevails over its non-local, symmetry-induced structure, in the kernel space defined by the architecture. We extend our theoretical treatment to any finite group, including non-abelian groups. Our framework also applies to equivariant architectures (e.g., CNNs), and recovers their success in the special case where the architecture matches the inherent symmetry of the data. Empirically, our theory reproduces the generalization failure of finite-width networks (MLP, CNN, ViT) trained on partially observed versions of rotated-MNIST. We conclude that conventional deep networks lack a mechanism to learn symmetries that have not been explicitly embedded in their architecture a priori. Our framework could be extended to guide the design of architectures and training procedures able to learn symmetries from data. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 07:56:54 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 15:02:44 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Perin",
"Andrea",
""
],
[
"Deny",
"Stephane",
""
]
] |
2412.11579 | Jingqian Wu | Jingqian Wu, Shuo Zhu, Chutian Wang, Boxin Shi, and Edmund Y. Lam | SweepEvGS: Event-Based 3D Gaussian Splatting for Macro and Micro Radiance Field Rendering from a Single Sweep | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in 3D Gaussian Splatting (3D-GS) have demonstrated the potential of using 3D Gaussian primitives for high-speed, high-fidelity, and cost-efficient novel view synthesis from continuously calibrated input views. However, conventional methods require high-frame-rate dense and high-quality sharp images, which are time-consuming and inefficient to capture, especially in dynamic environments. Event cameras, with their high temporal resolution and ability to capture asynchronous brightness changes, offer a promising alternative for more reliable scene reconstruction without motion blur. In this paper, we propose SweepEvGS, a novel hardware-integrated method that leverages event cameras for robust and accurate novel view synthesis across various imaging settings from a single sweep. SweepEvGS utilizes the initial static frame with dense event streams captured during a single camera sweep to effectively reconstruct detailed scene views. We also introduce different real-world hardware imaging systems for real-world data collection and evaluation for future research. We validate the robustness and efficiency of SweepEvGS through experiments in three different imaging settings: synthetic objects, real-world macro-level, and real-world micro-level view synthesis. Our results demonstrate that SweepEvGS surpasses existing methods in visual rendering quality, rendering speed, and computational efficiency, highlighting its potential for dynamic practical applications. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 09:09:42 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 03:39:49 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wu",
"Jingqian",
""
],
[
"Zhu",
"Shuo",
""
],
[
"Wang",
"Chutian",
""
],
[
"Shi",
"Boxin",
""
],
[
"Lam",
"Edmund Y.",
""
]
] |
2412.12641 | Konstantin Avrachenkov | Konstantin Avrachenkov, Vivek S. Borkar, Pratik Shah | Lagrangian Index Policy for Restless Bandits with Average Reward | null | null | null | null | cs.LG cs.AI math.OC math.PR | http://creativecommons.org/licenses/by/4.0/ | We study the Lagrange Index Policy (LIP) for restless multi-armed bandits with long-run average reward. In particular, we compare the performance of LIP with the performance of the Whittle Index Policy (WIP), both heuristic policies known to be asymptotically optimal under certain natural conditions. Even though in most cases their performances are very similar, in the cases when WIP shows bad performance, LIP continues to perform very well. We then propose reinforcement learning algorithms, both tabular and NN-based, to obtain online learning schemes for LIP in the model-free setting. The proposed reinforcement learning schemes for LIP require significantly less memory than the analogous schemes for WIP. We calculate analytically the Lagrange index for the restart model, which applies to the optimal web crawling and the minimization of the weighted age of information. We also give a new proof of asymptotic optimality in case of homogeneous arms as the number of arms goes to infinity, based on exchangeability and de Finetti's theorem. | [
{
"version": "v1",
"created": "Tue, 17 Dec 2024 08:03:53 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 14:00:55 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Avrachenkov",
"Konstantin",
""
],
[
"Borkar",
"Vivek S.",
""
],
[
"Shah",
"Pratik",
""
]
] |
2412.13918 | Matthias Barkowsky | Matthias Barkowsky and Holger Giese | Localized RETE for Incremental Graph Queries with Nested Graph Conditions | arXiv admin note: substantial text overlap with arXiv:2405.01145 | null | null | null | cs.LO cs.DB | http://creativecommons.org/licenses/by/4.0/ | The growing size of graph-based modeling artifacts in model-driven engineering calls for techniques that enable efficient execution of graph queries. Incremental approaches based on the RETE algorithm provide an adequate solution in many scenarios, but are generally designed to search for query results over the entire graph. However, in certain situations, a user may only be interested in query results for a subgraph, for instance when a developer is working on a large model of which only a part is loaded into their workspace. In this case, the global execution semantics can result in significant computational overhead.
To mitigate the outlined shortcoming, in this article we propose an extension of the RETE approach that enables local, yet fully incremental execution of graph queries, while still guaranteeing completeness of results with respect to the relevant subgraph.
We empirically evaluate the presented approach via experiments inspired by a scenario from software development and with queries and data from an independent social network benchmark. The experimental results indicate that the proposed technique can significantly improve performance regarding memory consumption and execution time in favorable cases, but may incur a noticeable overhead in unfavorable cases. | [
{
"version": "v1",
"created": "Wed, 18 Dec 2024 14:58:06 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 12:52:26 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Barkowsky",
"Matthias",
""
],
[
"Giese",
"Holger",
""
]
] |
2412.14501 | Yuzuki Arai | Yuzuki Arai and Sho Tsugawa | Do Large Language Models Advocate for Inferentialism? | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The emergence of large language models (LLMs) such as ChatGPT and Claude presents new challenges for philosophy of language, particularly regarding the nature of linguistic meaning and representation. While LLMs have traditionally been understood through distributional semantics, this paper explores Robert Brandom's inferential semantics as an alternative foundational framework for understanding these systems. We examine how key features of inferential semantics -- including its anti-representationalist stance, logical expressivism, and quasi-compositional approach -- align with the architectural and functional characteristics of Transformer-based LLMs. Through analysis of the ISA (Inference, Substitution, Anaphora) approach, we demonstrate that LLMs exhibit fundamentally anti-representationalist properties in their processing of language. We further develop a consensus theory of truth appropriate for LLMs, grounded in their interactive and normative dimensions through mechanisms like RLHF. While acknowledging significant tensions between inferentialism's philosophical commitments and LLMs' sub-symbolic processing, this paper argues that inferential semantics provides valuable insights into how LLMs generate meaning without reference to external world representations. Our analysis suggests that LLMs may challenge traditional assumptions in philosophy of language, including strict compositionality and semantic externalism, though further empirical investigation is needed to fully substantiate these theoretical claims. | [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 03:48:40 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 11:03:13 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Arai",
"Yuzuki",
""
],
[
"Tsugawa",
"Sho",
""
]
] |
2412.15028 | Gatien Verley | Paul Raux, Christophe Goupil, Gatien Verley | Thermodynamic Circuits: Modeling chemical reaction networks with nonequilibrium conductance matrices | 11 pages, 2 figures, last paper of the series | null | null | null | cond-mat.stat-mech physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We derive the nonequilibrium conductance matrix for open stationary Chemical Reaction Networks (CRNs) described by a deterministic mass action kinetic equation. As an illustration, we determine the nonequilibrium conductance matrix of a CRN made of two pseudo-linear sub-networks, called chemical modules, in two different ways: First by computing the nonequilibrium conductances of the modules that are then serially connected. Second by computing the nonequilibrium conductance of the CRN directly. The two approaches coincide, as expected from our theory of thermodynamic circuits. | [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 16:39:33 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 16:59:50 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Raux",
"Paul",
""
],
[
"Goupil",
"Christophe",
""
],
[
"Verley",
"Gatien",
""
]
] |
2412.15910 | Alexander Katsevich | Alexander Katsevich | Local analysis of iterative reconstruction from discrete generalized Radon transform data in the plane | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Local reconstruction analysis (LRA) is a powerful and flexible technique to study images reconstructed from discrete generalized Radon transform (GRT) data, $g=\mathcal R f$. The main idea of LRA is to obtain a simple formula to accurately approximate an image, $f_\epsilon(x)$, reconstructed from discrete data $g(y_j)$ in an $\epsilon$-neighborhood of a point, $x_0$. The points $y_j$ lie on a grid with step size of order $\epsilon$ in each direction. In this paper we study an iterative reconstruction algorithm, which consists of minimizing a quadratic cost functional. The cost functional is the sum of a data fidelity term and a Tikhonov regularization term. The function $f$ to be reconstructed has a jump discontinuity across a smooth surface $\mathcal S$. Fix a point $x_0\in\mathcal S$ and any $A>0$. The main result of the paper is the computation of the limit $\Delta F_0(\check x;x_0):=\lim_{\epsilon\to0}(f_\epsilon(x_0+\epsilon\check x)-f_\epsilon(x_0))$, where $f_\epsilon$ is the solution to the minimization problem and $|\check x|\le A$. A numerical experiment with a circular GRT demonstrates that $\Delta F_0(\check x;x_0)$ accurately approximates the actual reconstruction obtained by the cost functional minimization. | [
{
"version": "v1",
"created": "Fri, 20 Dec 2024 14:00:23 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 20:05:02 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Katsevich",
"Alexander",
""
]
] |
2412.16390 | Ashley Melvin | Ashley Melvin, J. C. Mandal | Development of a Weakly Compressible Solver for Incompressible Two-Phase Flows | 39 pages, 22 figures, partial work presented at ECCOMAS-2024 Lisbon, submitted to Journal of Computational Physics | null | 10.1016/j.jcp.2025.114195 | null | physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, a novel fully-explicit weakly compressible solver is developed for solving incompressible two-phase flows. The two-phase flow is modelled by coupling the general pressure equation, momentum conservation equations and the conservative level set advection equation. A HLLC-type Riemann solver is proposed to evaluate the convective fluxes along with a simple, consistent and oscillation-free discretization for the non-conservative terms. The solver is tested against several two-phase flow problems for its robustness and adaptability on structured as well as unstructured meshes. | [
{
"version": "v1",
"created": "Fri, 20 Dec 2024 22:57:13 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Melvin",
"Ashley",
""
],
[
"Mandal",
"J. C.",
""
]
] |
2412.16703 | Barun Kumar Pal | Barun Kumar Pal | Returning Back to Mukhanov Parametrization of Inflationary Equation of State | Accepted for publication in EPJC | EPJC, 2025 | 10.1140/epjc/s10052-025-14259-x | Volume 85, article number 547, (2025) | gr-qc astro-ph.CO hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We have re-examined Mukhanov parametrization for inflationary equation of state, $1+\omega=\frac{\beta}{({N}+1)^\alpha}$, in the light of Planck 2018 results and latest bound of tensor-to-scalar ratio employing Hamilton-Jacobi formalism. We have found that the current observational values of scalar spectral index and tensor-to-scalar ratio can be used efficiently to constrain the model parameters. The recent bound of $r<0.032$ has been used to put an upper bound on one of the model parameter. Whereas 1-$\sigma$ bound of the scalar spectral index $0.9607\leq n_{_S}\leq 0.9691$ along with the upper bound of tensor-to-scalar ratio provided restriction on the other model parameter $1.50<\alpha\leq2.20$. These bounds however depend on the number of e-foldings still left before the end of inflation and whenever $1.50<\alpha\leq2.20$ we can find appropriate values of the other model parameter $\beta$ so that the observational predictions are in tune with the latest available inflationary observables. We have further utilized the predictions from forthcoming CMB missions in the likes of CMB-S4 and LiteBIRD in order to obtain bounds on the model parameters. We find that detection of gravity waves would help us constrain the model parameters further. But in the absence of detection of primordial gravity wave signal by these CMB missions may rule out Mukhanov parametrization. | [
{
"version": "v1",
"created": "Sat, 21 Dec 2024 17:02:17 GMT"
},
{
"version": "v2",
"created": "Sat, 03 May 2025 16:59:19 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Pal",
"Barun Kumar",
""
]
] |
2412.17003 | Roni Con | Roni Con | Anonymous Shamir's Secret Sharing via Reed-Solomon Codes Against Permutations, Insertions, and Deletions | null | null | null | null | cs.IT cs.CR math.IT | http://creativecommons.org/licenses/by/4.0/ | In this work, we study the performance of Reed-Solomon codes against an adversary that first permutes the symbols of the codeword and then performs insertions and deletions. This adversarial model is motivated by the recent interest in fully anonymous secret-sharing schemes [EBG+24],[BGI+24]. A fully anonymous secret-sharing scheme has two key properties: (1) the identities of the participants are not revealed before the secret is reconstructed, and (2) the shares of any unauthorized set of participants are uniform and independent. In particular, the shares of any unauthorized subset reveal no information about the identity of the participants who hold them.
In this work, we first make the following observation: Reed-Solomon codes that are robust against an adversary that permutes the codeword and then deletes symbols from the permuted codeword can be used to construct ramp threshold secret-sharing schemes that are fully anonymous. Then, we show that over large enough fields of size, there are $[n,k]$ Reed-Solomon codes that are robust against an adversary that arbitrary permutes the codeword and then performs $n-2k+1$ insertions and deletions to the permuted codeword. This implies the existence of a $(k-1, 2k-1, n)$ ramp secret sharing scheme that is fully anonymous. That is, any $k-1$ shares reveal nothing about the secret, and, moreover, this set of shares reveals no information about the identities of the players who hold them. On the other hand, any $2k-1$ shares can reconstruct the secret without revealing their identities. We also provide explicit constructions of such schemes based on previous works on Reed-Solomon codes correcting insertions and deletions. The constructions in this paper give the first gap threshold secret-sharing schemes that satisfy the strongest notion of anonymity together with perfect reconstruction. | [
{
"version": "v1",
"created": "Sun, 22 Dec 2024 12:51:16 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 12:45:57 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Con",
"Roni",
""
]
] |
2412.17587 | Aswini Kumar Patra | Aswini Kumar Patra, Tejashwini Gajurel | Improved Cotton Leaf Disease Classification Using Parameter-Efficient Deep Learning Framework | 4 figures, 3 Tables | null | 10.1109/ECTIDAMTNCON64748.2025.10961983 | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cotton crops, often called "white gold," face significant production challenges, primarily due to various leaf-affecting diseases. As a major global source of fiber, timely and accurate disease identification is crucial to ensure optimal yields and maintain crop health. While deep learning and machine learning techniques have been explored to address this challenge, there remains a gap in developing lightweight models with fewer parameters which could be computationally effective for agricultural practitioners. To address this, we propose an innovative deep learning framework integrating a subset of trainable layers from MobileNet, transfer learning, data augmentation, a learning rate decay schedule, model checkpoints, and early stopping mechanisms. Our model demonstrates exceptional performance, accurately classifying seven cotton disease types with an overall accuracy of 98.42% and class-wise precision ranging from 96% to 100%. This results in significantly enhanced efficiency, surpassing recent approaches in accuracy and model complexity. The existing models in the literature have yet to attain such high accuracy, even when tested on data sets with fewer disease types. The substantial performance improvement, combined with the lightweight nature of the model, makes it practically suitable for real-world applications in smart farming. By offering a high-performing and efficient solution, our framework can potentially address challenges in cotton cultivation, contributing to sustainable agricultural practices. | [
{
"version": "v1",
"created": "Mon, 23 Dec 2024 14:01:10 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Patra",
"Aswini Kumar",
""
],
[
"Gajurel",
"Tejashwini",
""
]
] |
2412.18108 | Junjia Guo | Jing Bi, Junjia Guo, Yunlong Tang, Lianggong Bruce Wen, Zhang Liu, Chenliang Xu | Unveiling Visual Perception in Language Models: An Attention Head Analysis Approach | null | CVPR 2025 (IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025) | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated remarkable progress in visual understanding. This impressive leap raises a compelling question: how can language models, initially trained solely on linguistic data, effectively interpret and process visual content? This paper aims to address this question with systematic investigation across 4 model families and 4 model scales, uncovering a unique class of attention heads that focus specifically on visual content. Our analysis reveals a strong correlation between the behavior of these attention heads, the distribution of attention weights, and their concentration on visual tokens within the input. These findings enhance our understanding of how LLMs adapt to multimodal tasks, demonstrating their potential to bridge the gap between textual and visual understanding. This work paves the way for the development of AI systems capable of engaging with diverse modalities. | [
{
"version": "v1",
"created": "Tue, 24 Dec 2024 02:31:24 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Bi",
"Jing",
""
],
[
"Guo",
"Junjia",
""
],
[
"Tang",
"Yunlong",
""
],
[
"Wen",
"Lianggong Bruce",
""
],
[
"Liu",
"Zhang",
""
],
[
"Xu",
"Chenliang",
""
]
] |
2501.00524 | Shuai Zhang | Shuai Zhang, Aliza Panjwani, Penghao Xiao, Maitrayee Ghosh, Tadashi Ogitsu, Yuan Ping, and S. X. Hu | Thermal Induced Structural Competitiveness and Metastability of Body-centered Cubic Iron under Non-Equilibrium Conditions | 5 pages, 4 figures | Phys. Rev. B 111, 224106 (2025) | 10.1103/ws64-kmzr | null | cond-mat.mtrl-sci physics.geo-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The structure and stability of iron near melting at multi-megabar pressures are of significant interest in high pressure physics and earth and planetary sciences. While the body-centered cubic (BCC) phase is generally recognized as unstable at lower temperatures, its stability relative to the hexagonal close-packed (HCP) phase at high temperatures (approximately 0.5 eV) in the Earth's inner core (IC) remains a topic of ongoing theoretical and experimental debate. Our ab initio calculations show a significant drop in energy, the emergence of a plateau and a local minimum in the potential energy surface, and stabilization of all phonon modes at elevated electron temperatures (>1-1.5 eV). These effects increase the competition among the BCC, HCP, and the face-centered cubic (FCC) phases and lead to the metastability of the BCC structure. Furthermore, the thermodynamic stability of BCC iron is enhanced by its substantial lattice vibration entropy. This thermally induced structural competitiveness and metastability under non-equilibrium conditions provide a clear theoretical framework for understanding iron phase relations and solidification processes, both experimentally and in the IC. | [
{
"version": "v1",
"created": "Tue, 31 Dec 2024 16:11:39 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Zhang",
"Shuai",
""
],
[
"Panjwani",
"Aliza",
""
],
[
"Xiao",
"Penghao",
""
],
[
"Ghosh",
"Maitrayee",
""
],
[
"Ogitsu",
"Tadashi",
""
],
[
"Ping",
"Yuan",
""
],
[
"Hu",
"S. X.",
""
]
] |
2501.02469 | Atonu Ghosh | Atonu Ghosh, Sudip Misra | LoRaConnect: Unlocking HTTP Potential on LoRa Backbones for Remote Areas and Ad-Hoc Networks | 10 pages | null | null | null | cs.NI cs.CY cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Minimal infrastructure requirements make LoRa suitable for service delivery in remote areas. Additionally, web applications have become a de-facto standard for modern service delivery. However, Long Range (LoRa) fails to enable HTTP access due to its limited bandwidth, payload size limitations, and high collisions in multi-user setups. We propose LoRaConnect to enable HTTP access over LoRa. The LoRaWeb hardware tethers a WiFi hotspot to which client devices connect and access HTTP resources over LoRa backhaul. It implements caching and synchronization mechanisms to address LoRa's aforementioned limitations. It also implements a message-slicing method in the application layer to overcome LoRa's payload limitations. We evaluate the proposed system using actual hardware in three experimental setups to assess the baseline performance, ideal scenario, and practical application scenario with Frequency Hopping Spread Spectrum (FHSS). Additionally, it implements a ping operation to demonstrate Internet capability and extensible nature. LoRaWeb achieves an average throughput of 1.18 KB/S approximately, with an access delay of only 1.3 S approximately for a 1.5KB webpage in the baseline setup. Moreover, it achieves an access delay of approximately 6.7 S for a 10KB webpage in the ideal case and an average end-to-end delay of only 612 ms approximately in the FHSS-based setup. Comparison with benchmark suggests multi-fold improvement. | [
{
"version": "v1",
"created": "Sun, 05 Jan 2025 07:41:53 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 07:58:19 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 05:12:22 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Ghosh",
"Atonu",
""
],
[
"Misra",
"Sudip",
""
]
] |
2501.02648 | David Restrepo | David Restrepo, Chenwei Wu, Yueran Jia, Jaden K. Sun, Jack Gallifant, Catherine G. Bielick, Yugang Jia, Leo A. Celi | Representation Learning of Lab Values via Masked AutoEncoders | 14 pages of main text, 11 appendix | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Accurate imputation of missing laboratory values in electronic health records (EHRs) is critical to enable robust clinical predictions and reduce biases in AI systems in healthcare. Existing methods, such as XGBoost, softimpute, GAIN, Expectation Maximization (EM), and MICE, struggle to model the complex temporal and contextual dependencies in EHR data, particularly in underrepresented groups. In this work, we propose Lab-MAE, a novel transformer-based masked autoencoder framework that leverages self-supervised learning for the imputation of continuous sequential lab values. Lab-MAE introduces a structured encoding scheme that jointly models laboratory test values and their corresponding timestamps, enabling explicit capturing temporal dependencies. Empirical evaluation on the MIMIC-IV dataset demonstrates that Lab-MAE significantly outperforms state-of-the-art baselines such as XGBoost, softimpute, GAIN, EM, and MICE across multiple metrics, including root mean square error (RMSE), R-squared (R2), and Wasserstein distance (WD). Notably, Lab-MAE achieves equitable performance across demographic groups of patients, advancing fairness in clinical predictions. We further investigate the role of follow-up laboratory values as potential shortcut features, revealing Lab-MAE's robustness in scenarios where such data is unavailable. The findings suggest that our transformer-based architecture, adapted to the characteristics of EHR data, offers a foundation model for more accurate and fair clinical imputation. In addition, we measure and compare the carbon footprint of Lab-MAE with the a XGBoost model, highlighting its environmental requirements. | [
{
"version": "v1",
"created": "Sun, 05 Jan 2025 20:26:49 GMT"
},
{
"version": "v2",
"created": "Thu, 09 Jan 2025 11:17:01 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 15:34:13 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Restrepo",
"David",
""
],
[
"Wu",
"Chenwei",
""
],
[
"Jia",
"Yueran",
""
],
[
"Sun",
"Jaden K.",
""
],
[
"Gallifant",
"Jack",
""
],
[
"Bielick",
"Catherine G.",
""
],
[
"Jia",
"Yugang",
""
],
[
"Celi",
"Leo A.",
""
]
] |
2501.03236 | Haonan Gu | Haonan Gu | P-adic integrals: A solution to the p-adic Schrodinger Equation | null | null | null | null | math.GM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper first introduces the concept of p-adic number and field. Then it develops the p-adic integration and applied it to solve p-adic Schrodinger equations. | [
{
"version": "v1",
"created": "Sat, 14 Dec 2024 06:51:52 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Jun 2025 19:11:57 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Gu",
"Haonan",
""
]
] |
2501.03717 | Lezhong Wang | Lezhong Wang, Duc Minh Tran, Ruiqi Cui, Thomson TG, Anders Bjorholm Dahl, Siavash Arjomand Bigdeli, Jeppe Revall Frisvad, Manmohan Chandraker | Materialist: Physically Based Editing Using Single-Image Inverse Rendering | Add acknowledgements, more authors and more results. Project website: https://lez-s.github.io/materialist_project/ | null | null | null | cs.CV cs.AI cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Achieving physically consistent image editing remains a significant challenge in computer vision. Existing image editing methods typically rely on neural networks, which struggle to accurately handle shadows and refractions. Conversely, physics-based inverse rendering often requires multi-view optimization, limiting its practicality in single-image scenarios. In this paper, we propose Materialist, a method combining a learning-based approach with physically based progressive differentiable rendering. Given an image, our method leverages neural networks to predict initial material properties. Progressive differentiable rendering is then used to optimize the environment map and refine the material properties with the goal of closely matching the rendered result to the input image. Our approach enables a range of applications, including material editing, object insertion, and relighting, while also introducing an effective method for editing material transparency without requiring full scene geometry. Furthermore, Our envmap estimation method also achieves state-of-the-art performance, further enhancing the accuracy of image editing task. Experiments demonstrate strong performance across synthetic and real-world datasets, excelling even on challenging out-of-domain images. Project website: https://lez-s.github.io/materialist_project/ | [
{
"version": "v1",
"created": "Tue, 07 Jan 2025 11:52:01 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 16:22:07 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Wang",
"Lezhong",
""
],
[
"Tran",
"Duc Minh",
""
],
[
"Cui",
"Ruiqi",
""
],
[
"TG",
"Thomson",
""
],
[
"Dahl",
"Anders Bjorholm",
""
],
[
"Bigdeli",
"Siavash Arjomand",
""
],
[
"Frisvad",
"Jeppe Revall",
""
],
[
"Chandraker",
"Manmohan",
""
]
] |
2501.03917 | Walter Lambrecht | Claudio Garcia and Walter R. L. Lambrecht | Optical absorption and luminescence of $\alpha$-LiV$_2$O$_5$ from the Bethe Salpeter Equation | 9 pages 7 figures | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | $\alpha$-Li$_x$V$_2$O$_5$ is obtained by intercalating Li between the layers of V$_2$O$_5$. The partial filling of the split-off conduction band by electron donation from Li leads to significant changes in optical properties. Here we study the electronic band structure of $\alpha$-LiV$_2$O$_5$ using quasiparticle self-consistent (QS) $GW$ calculations and the optical dielectric function by means of the Bethe Salpeter Equation (BSE). The half-filling of the narrow split-off band leads to a spin-splitting and formation of magnetic moments of 0.5 $\mu_B$ per V which order antiferromagnetically along the chain or $b$-direction.
The imaginary part of the dielectric function shows a very strong optical absorption band near 2 eV for polarization along the $a$-direction. This absorptions stems from a localized transition between the occupied V-$d_{xy}$ derived band, which is odd with respect to the $a$-mirrorplane to the higher lying empty band formed from the same V-$d_{xy}$
orbitals but even with respect to that mirror-plane, which explains its polarization and large oscillator strength.
We relate this to a recent experimental study of
cathodoluminescence (CL) in which a suppression of the lowest CL peak was observed upon addition of Li. The lowest CL peak near 1.8 eV, which lies well below the indirect absorption onset of V$_2$O$_5$ is proposed to be related to recombination of a self-trapped electron polaron, resulting from oxygen vacancies, with a hole at the valence band maximum and is suppressed in LiV$_2$O$_5$ by the strong self--absorption from the Li induced occupied band to the higher empty bands at about the same energy. | [
{
"version": "v1",
"created": "Tue, 07 Jan 2025 16:31:16 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Feb 2025 21:01:41 GMT"
},
{
"version": "v3",
"created": "Mon, 28 Apr 2025 14:56:10 GMT"
},
{
"version": "v4",
"created": "Wed, 25 Jun 2025 20:49:00 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Garcia",
"Claudio",
""
],
[
"Lambrecht",
"Walter R. L.",
""
]
] |
2501.04092 | Pablo Pais | Fabrizio Canfora, Pablo Pais | A Bogomol'nyi-Prasad-Sommerfield bound with a first-order system in the $2D$ Gross-Pitaevskii equation | 21 pages, 4 Figures. Version revised and accepted in European Physical Journal C | null | null | null | cond-mat.quant-gas hep-ph hep-th nlin.SI nucl-th | http://creativecommons.org/licenses/by/4.0/ | A novel Bogomol'nyi-Prasad-Sommerfield (BPS) bound for the Gross-Pitaevskii equations in two spatial dimensions is presented. The energy can be bounded from below in terms of the combination of two boundary terms, one related to the vorticity (but ``dressed'' by the condensate profile) and the second to the ``skewness'' of the configurations. The bound is saturated by configurations that satisfy a system of two first-order partial differential equations. When such a BPS system is satisfied, the Gross-Pitaevskii equations are also satisfied. The analytic solutions of this BPS system in the present manuscript represent configurations with fractional vorticity living in an annulus. Using these techniques, we present the first analytic examples of this kind. The hydrodynamical interpretation of the BPS system is discussed, and the implications of these results are outlined. | [
{
"version": "v1",
"created": "Tue, 07 Jan 2025 19:00:06 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Feb 2025 03:28:39 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 01:43:55 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Canfora",
"Fabrizio",
""
],
[
"Pais",
"Pablo",
""
]
] |
2501.04334 | Piyabut Burikham | Theppawan Rukpakawong and Piyabut Burikham | The effective phase space and $e$-folding of the Starobinsky and extended Starobinsky model of inflation | 17 pages, 12 figures | null | null | null | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | For zero spatial curvature, cosmological phase space of Starobinsky and extended Starobinsky inflationary model show three apparent attractors; the fixed angle attractor in the large field limit, the final attractor representing reheating phase in the small field region, and the apparent attractor corresponding to the slow-roll condition connecting between the large-field and small-field region. To consider the total $e$-folding likelihood of the model, Remmen-Carroll conserved measure is constructed and normalized. Using the measure, the total e-folding number $N$ and its expectation value $\left\langle N \right\rangle$ are calculated. Our results show that most classical slow-roll trajectories which intersect the Planck surface have $N<60$, and $\phi_{\rm UV}>5.5 M^{*}_{\rm Pl}$ is required for $N>60$. It is found that for $\phi_{\rm UV}\in [5.22,5.50]M^{*}_{\rm Pl}$ which satisfies the constraint on the spectral index, $n_s = 0.9658 \pm 0.0040\,\,\,(68\%\,\,{\rm CL})$, the expectation value $\langle N \rangle \simeq 3.5 - 4$ for trajectories intersecting the Planck surface in the Starobinsky model. For extended Starobinsky model with additional $R^3$ term parametrized by a coupling parameter $\alpha$, the expectation value when inflation starts from the top of the potential shifts to $\left\langle N \right\rangle = 4.025,4.336$ for $\alpha = 10^{-4},6.5\times 10^{-5}$ respectively. In the Starobinsky model even at very large inflaton cutoff $\phi_{\rm UV}$ where the field value is super-Planckian, the energy density from the (saturating) inflaton potential and the Hubble parameter are still sub-Planckian and therefore the inflation occurs within the semi-classical regime. | [
{
"version": "v1",
"created": "Wed, 08 Jan 2025 08:11:02 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Jun 2025 07:19:22 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Rukpakawong",
"Theppawan",
""
],
[
"Burikham",
"Piyabut",
""
]
] |
2501.04502 | Dipendu Halder | Dipendu Halder and Saurabh Basu | Controlled probing of localization effects in the non-Hermitian Aubry-Andr\'e model via topolectrical circuits | Published in Phys. Rev. B; 12 pages & 7 figures | Phys. Rev. B 111, 235447 (2025) | 10.1103/7t7k-qg49 | null | cond-mat.dis-nn cond-mat.mes-hall physics.app-ph quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Anderson localization and the non-Hermitian skin effect are two distinct confinement phenomena of the eigenfunctions that are driven, respectively, by disorder and nonreciprocity. Understanding their interplay within a unified framework offers valuable insights into the localization properties of low-dimensional systems. To this end, we investigate a non-Hermitian version of the celebrated Aubry-Andr\'e model, which serves as an ideal platform due to its unique self-dual properties and ability to demonstrate a delocalization-localization transition in one dimension. Interestingly, in our setting, the competition between Anderson localization and the skin effect can be precisely controlled via the complex phase of the quasiperiodic disorder. Additionally, by analyzing the time evolution, we demonstrate that quantum jumps between the skin states and the Anderson-localized states occur in the theoretical model. Further, to gain support for our theoretical predictions in an experimental platform, we propose a topolectrical circuit featuring an interface that separates two distinct electrical circuit networks. The voltage profile of the circuit exhibits confinement at the interface, analogous to the skin effect, while the phenomenon of Anderson localization in the circuit can be perceived via a predicted localization behavior near the excitation node, rather than exhibiting sudden non-Hermitian jumps, as observed in the tight-binding framework. This interplay leads to a spatially tunable localization of the output voltage of the circuit. Our findings provide deeper insights into the controlled confinement of the eigenstates of the non-Hermitian Aubry-Andr\'e model by designing analogous features in topolectrical circuits, opening avenues in the fabrication of advanced electronic systems such as highly sensitive sensors and efficient devices for information transfer and communication. | [
{
"version": "v1",
"created": "Wed, 08 Jan 2025 13:42:47 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Apr 2025 16:26:30 GMT"
},
{
"version": "v3",
"created": "Thu, 26 Jun 2025 16:45:46 GMT"
}
] | 2025-06-27T00:00:00 | [
[
"Halder",
"Dipendu",
""
],
[
"Basu",
"Saurabh",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.