Peer Revenue Impact factor: volatile

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Monday 9 November

Artificial Intelligence

Trust, but verify, audit and attribute: a human-in-the-loop framework for checking AI-produced research results

Artificial intelligence (AI) systems now produce research results in bulk: in October 2026 a single laboratory released more than 700 AI-generated mathematics manuscripts, only part of them with machine-checked proofs. Such volumes will be checked mostly by AI agents that call specialised software. A passing check certifies only the artefact that was checked. For proofs translated into the proof assistant Lean, it has been shown that this artefact may state something other than the author's text, and that deciding whether a translation is faithful is undecidable in general. The human researcher therefore stays in the loop, and human and AI contributions must be recorded. We propose a domain-independent, human-in-the-loop framework for checking AI-produced results. Validated, agent-usable tools with different grounds of trust check the results; a fidelity audit asks whether each check addressed the claim as stated; the human decides at fixed points; and a dialog report attributes the work to the human and to the AI. We describe two automatic implementations of the tool layer, for mathematics (Wolfram Language, MATLAB, Lean) and statistics (Python and R), and run the mathematical one on a two-claim example. An ambiguous claim was split into two readings, one refuted and one proved.
Read the full abstract

Software Engineering

Code-Grounded Agentic Explanations for Static Analysis Vulnerability Findings

Static application security testing (SAST) tools surface vulnerabilities early and their verdicts are rule-based and auditable. Yet developers routinely dismiss findings they cannot interpret. Large language models can supply the missing explanation, but designs that also hand the model the detection give up the auditability that makes static analysis defensible. We keep detection with the analyser and give the model only the explaining. A preliminary user study with 20 participants and a workshop with 19 practitioners yielded five requirements. Guided by these we redesigned SAFE, an open-source IntelliJ IDEA plugin that explains findings from any analyser reporting in the Static Analysis Results Interchange Format. It tiers explanations by demand, ties them to the reported data flow, and replaces the single-prompt backend with an agent that must consult the repository and the rule that fired before answering. A within-subjects crossover experiment with 24 participants compared the redesign against its predecessor. Perceived usability rose from 77.3 to 85.0 on the System Usability Scale. Actionability gained most among the explanation-quality items, the aspect both earlier studies flagged as weakest, and the odds of a correct remediation were 2.6 times higher. Triage accuracy and time per finding were indistinguishable, which we read as uninformative rather than as equivalence. Trust was the one measure that did not move, and the closing interviews indicate that grounding made findings easier to understand rather than easier to believe. Participants also drew a boundary no requirement anticipated, wanting the tool to find and explain while they judge and apply.
Read the full abstract

All authors Oshando Johnson, Ranjith Krishnamurthy, Rohith Kumar Shanmuganathan, Goran Piskachev, Eric Bodden

Differential Geometry

On the Concavity of Rényi Entropy and Rényi Entropy Power along Heat Flows on Manifolds

In this paper, we establish the concavity of Rényi entropy and Rényi entropy power along heat flow on Euclidean space of arbitrary dimension, for Rényi orders $\alpha\in[1,2]$. More generally, we prove the corresponding result along the heat flow generated by the Witten Laplacian on complete Riemannian manifolds under a suitable curvature-dimension condition. Our argument builds on novel functional inequalities that were recently developed for the study of Tsallis entropy concavity along the heat flow. We extend both these inequalities and the resulting concavity theorem from the Euclidean setting to the Riemannian manifold setting. In addition, we establish the complete monotonicity of Tsallis entropy of order $2$ along the heat flow and derive its spectral decomposition.
Read the full abstract

Robotics

Robo-MLT: Dual-System Robotic Manipulation for Long-Horizon Medical Laboratory Tasks

Reliable medical laboratory automation requires robots to complete entire workflows in protocol order while handling changes during execution. A robot may execute each motion correctly yet fail the workflow by confusing visually similar stages or stalling after human intervention. We present Robo-MLT, a dual-system framework that couples explicit workflow monitoring with learned action generation. A vision–language planner composes an ordered atomic-skill sequence from the protocol and the initial workspace observation. A multiframe execution monitor detects when the active skill has completed or become infeasible, then advances the workflow or triggers replanning. A skill-conditioned vision–language–action policy generates the manipulation trajectories, and an asynchronous runtime coordinates it with the monitor. We evaluate Robo-MLT against four policy baselines and a VLM-based monitor baseline on blood gas analysis and blood sample tube sorting, with 50 nominal configurations per method and 20 sample-removal configurations. Robo-MLT achieves task success rates of 82% and 70%, exceeding $\pi_{0.5}$ by 30 and 22 percentage points, respectively. After an operator removes the sample being handled, the system revises the remaining plan and continues processing the other samples. These results indicate that, under the evaluated conditions, explicit workflow monitoring makes long-horizon laboratory manipulation more reliable and recoverable.
Read the full abstract

All authors Zequan Zeng, Kai Wang, Delvin Ce Zhang, Huaping Liu, Zhipeng Yuan, Junqin Yang, YunLong Zhang

Materials Science

Point defects and primary radiation damage in Co-free Fe-Ni-Cr-Mn high-entropy alloys from molecular dynamics

Co-free Fe-Ni-Cr-Mn high-entropy alloys (HEAs) are candidates for nuclear structural applications. We used molecular dynamics with the EAM-21 potential, stiffened at short range, to study five HEAs that have been irradiated experimentally: Fe37Ni34Cr16Mn13 (ES1), Fe35Ni35Cr15Mn15 (P35), Fe46Ni22Cr15Mn17 (Y3), Fe27Ni28Cr18Mn27 (Kumar) and Fe30Ni30Cr10Mn30 (C30). They were compared with pure Ni, Fe60Ni40, the Fe70Ni10Cr20 model alloy of 316L, and Fe50Ni34Cr16, which is ES1 with Mn replaced by Fe. We calculated stacking fault energies, threshold displacement energies, point-defect formation and migration energies, and 1015 displacement cascades from 5 to 40 keV at 300 and 823 K. The stacking fault energy follows the Ni content: 19-44 mJ/m^2 in the alloys and 113 mJ/m^2 in Ni. Compared with the Mn-free alloys, the HEAs have lower threshold displacement energies (48-56 against 64-74 eV) and lower dumbbell formation energies, and Mn is the element most transported by interstitials. All the concentrated alloys leave about the same number of Frenkel pairs, 56-72 at 40 keV, half the number in Ni. At 40 keV, 54-58% of the interstitials are in clusters in the HEAs, against 67-71% in Fe60Ni40, Fe70Ni10Cr20 and Fe50Ni34Cr16 and 82% in Ni. At the same cluster size, clusters of 4-12 interstitials are faulted about half as often in the HEAs, and this does not follow the stacking fault energy. The surviving interstitials are enriched in Mn in the HEAs and in Ni in the Mn-free alloys, and depleted in Cr. The results support direct loop formation in cascades in Fe60Ni40 and Fe70Ni10Cr20 and less of it in ES1.
Read the full abstract

All authors Ayobami Daramola, Giovanni Bonny, Gilles Adjanor, Christophe Domain, Ghiath Monnet, Anna Fraczkiewicz, Duc Nguyen-Manh

Statistics Theory

Spectral-Domain Information Geometry of Stationary Covariance Models: Gaussian and Elliptical Extensions

This paper develops an information-geometric framework for stationary covariance structures. We begin with the infinite-dimensional setting, where stationary covariance structures can be characterized without finite-window boundary effects. By formulating the covariance structure as a bilinear form on a suitable class of test functions, we establish its spectral representation and show how finite-dimensional Toeplitz covariance matrices arise as restrictions to nested observation subspaces. For zero-mean proper complex Gaussian wide-sense stationary (WSS) processes, we show that the normalized finite-dimensional Fisher information converges to a spectral-domain metric. In logarithmic spectral coordinates, this metric becomes the standard $L^2$ inner product, yielding an explicit Fisher–Rao distance that is also recovered as the asymptotic normalized affine-invariant Riemannian distance between Toeplitz covariance matrices. We then extend the framework to elliptical observation models by separating global covariance scale from shape. The shape Fisher information is orthogonal to radial nuisance directions and is characterized by a dimension- and radial-model-dependent coefficient. We derive a condition under which this coefficient converges to its Gaussian value as the observation dimension tends to infinity. The resulting limiting spectral-shape geometry is the centered $L^2$ geometry of log-spectra and is invariant to positive rescaling. Numerical experiments illustrate the theoretical convergence results and the recovery of the Gaussian shape coefficient.
Read the full abstract

Operating Systems

CapBuddy: Efficient Physical Memory Management for Capability-based Resource-controlling Microkernel

Capability-based resource-controlling microkernels securely multiplex hardware by exposing low-level resource abstractions guarded by capability systems. Such a design helps provide an application-specific resource management policy to achieve high performance without sacrificing security. However, the misuse of capability-protected secure abstractions can introduce performance overhead in user-level resource management, affecting the runtime performance of dynamic applications in emerging scenarios. In this paper, we focus on the secure abstraction of physical memory in microkernels with capability-based resource management delegation to investigate an appropriate design for dynamic memory management. We propose CapBuddy, an approach to building an efficient buddy system on capability-protected secure abstractions of memory, to bypass the performance overhead of the capability system by offloading functionalities that can run offline from the critical path of serving a memory request. We implemented a prototype system on the seL4 library OS, and the evaluation results show that our approach improves the performance of the seL4 user-level buddy system by 20 to 54.5 times and achieves the same order of magnitude as Linux. Further burst experiments evaluate CapBuddy's scalability and show that it maintains low median allocation latency as the number of concurrent clients increases.
Read the full abstract

All authors Zhuang Lu, Guangtao Zhu, Shijun Zhao, Qianying Zhang, Lixin Chen, Zekun Xu, Wu Luo, Rui Hou

High Energy Astrophysical Phenomena

The Ordinary Days of Extraordinary Blazar 3C 454.3

The blazar 3C 454.3 (z = 0.859) is one of the brightest and most variable flat-spectrum radio quasars in the sky. Here, we present a multi-wavelength analysis of this enigmatic object in its quiescent states during June 2022 and May 2023 coincident with its first X-ray polarimetric and hard X-ray observations with International X-ray Polarimetric Explorer (IXPE) and Nuclear Spectroscopic Telescope Array (NuSTAR). Though no X-ray polarization was detected, NuSTAR detected the source on all occasions during this period. We generated five broadband spectral energy distributions (SEDs)of the blazar using quasi-simultaneous observations taken with the Fermi-Large Area Telescope, NuSTAR, and Neil Gehrels Swift Observatory. The hard X-ray flux measured by NuSTAR exhibited mild flux variations. The SEDs were modeled with a one-zone leptonic emission framework using the JetSet package. The optical-ultraviolet spectrum of 3C 454.3 revealed an upturn which we reproduced with the standard accretion disk model. The X-ray to gamma-ray spectrum of 3C 454 was found to be well explained by the external Compton (EC) process with the seed photons provided by the BLR and accretion disk. These results provide intriguing clues about the multi-wavelength behavior of 3C 454.3 during its quiescent activity state.
Read the full abstract

All authors Sajeesh George, P N Naseef Mohammed, Aminabi Thekkoth, Vaidehi S. Paliya, C D Ravikumar

Robotics

RoboACE: Adversarial Co-Evolution for Robot Self-Improvement

Robot agents can improve reusable skills, planning procedures, and recovery strategies from execution feedback, yet a fixed practice source leaves the search for new weaknesses unchanged. Useful practice depends on the agent's current capabilities, so the procedure that chooses practice must also be revisable. We present RoboACE (Robot Adversarial Co-Evolution), a framework that jointly evolves curriculum and execution harnesses around frozen foundation models. The curriculum harness uses accumulated capability evidence to propose informative challenges, while the execution harness responds with reusable revisions that expand execution capabilities. Alternating updates hold one side fixed as the other changes, and admitted challenges form fixed snapshots for comparing execution revisions. Starting from a compact set of LIBERO practice conditions, the resulting frozen deployment outperforms the evaluated controllers and agent systems across four LIBERO-family panels. On LIBERO-PRO, it reaches 74.0% success, exceeding the initial deployment by 15.0 percentage points and direct execution by its VLA controller by 13.5 points. Case studies document changes to evidence-guided practice selection, drawer-skill grounding, and controller intervention.
Read the full abstract

All authors Delin Zhao, Zhengrong Yue, Junlin He, Yali Wang

Instrumentation and Methods for Astrophysics

Mitigating oscillatory systematics in MeerKLASS L-band H i intensity maps

HI intensity mapping is a promising probe of the large-scale structure of the Universe, but is limited by foregrounds and instrumental systematics. For the MeerKAT Large Area Synoptic Survey (MeerKLASS) L-band data, it is found that emission line stacking measurement reveals a type of systematics related to the instrument beam. Using simulations with a full-sky electromagnetic beam model, we show that the chromaticity and sidelobes of the MeerKAT beam modulate the frequency-smooth foreground emission and introduce oscillating structures that persist after foreground removal. The beam rotates with respect to the sky between two types of scans, and combining datasets with different beam orientations before foreground cleaning makes the residuals stronger. In MeerKLASS L-band data, processing these two types of scans separately significantly suppresses the oscillating residuals while preserving the stacked HI signal. The ratio of the HI signal amplitude to the systematic fluctuation amplitude increases by 81%. We find an oscillating structure in PCA mode 11 in our previous data processing, but simply removing this mode does not improve the stacked spectrum due to the signal loss. Our strategy can help mitigate spectral ripples that are induced by the beam in future MeerKAT and SKA-Mid surveys.
Read the full abstract

All authors Boyan Zhao, Eloy de Lera Acedo, Zhaoting Chen, John Cumner, Vijay Mahatma, Marta Spinelli, Dirk I. L. de Villiers, Matilde Barberi-Squarotti, Phil Bull, Stefano Camera, Isabella Paola Carucci, Xuelei Chen, Steven Cunnington, Furen Deng, Brandon N. Engelbrecht, José Fonseca, Keith Grainge, Wenkai Hu, Melis O. Irfan, Yichao Li, Mario G. Santos, Jingying Wang, Yougang Wang, Laura Wolz

Quantum Gases

Universal Fold Caustics in Anharmonic Matter-wave Lenses

Caustics provide a universal connection between classical ray focusing and structured wave patterns. We theoretically investigate this connection in non-linear matter waves focused by a short-duration anharmonic lens acting on a Bose-Einstein condensate. A semi-classical description shows how the leading cubic aberration transforms the harmonic focal point into a fold caustic, whose local wave field is described by an Airy canonical integral. Three-dimensional Gross-Pitaevskii simulations show that mean-field interactions modify the near-focus density and longitudinal momentum distribution, producing a secondary momentum-space feature absent from the non-interacting evolution. We quantify the information about the effective interaction strength carried by the $1D$ distribution using classical Fisher information and maximum-likelihood estimation. We obtain a relative uncertainty of $2.7\times10^{-3}$ for a sample of $10^5$ detected points along $z$. These results establish a framework for studying interaction-driven departures from single-particle caustic universality and promote matter-wave lenses as a sensitive probe of mean-field interactions.
Read the full abstract

All authors Joachim Guyomard, Julia Pahl, Ernst Rasel, Naceur Gaaloul

Dynamical Systems

Explicit resonance and bifurcation of symmetric periodic solutions for differential systems with multiple distributed delays

We investigate the bifurcation of symmetric periodic solutions for differential systems with multiple distributed delays. For general multiple-delay systems with a quadratic potential, we establish a Hamiltonian formulation with a \(P_{n,m}\)-boundary condition and derive explicit Maslov index and nullity formulas. For the special case of two distributed delays, we extend the analysis to general nonlinear potentials. These results yield explicit resonance criteria, bifurcation conditions, and Rabinowitz type and Fadell-Rabinowitz type alternatives for $\mathbb{R} $-distinct solutions.
Read the full abstract

Artificial Intelligence

APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models

Vision-language-action (VLA) policies have demonstrated strong generalization in robotic manipulation, yet processing dense visual tokens often incurs high inference latency. Visual token pruning offers a practical way to reduce this cost, but identifying a compact set of tokens that preserves reliable action generation remains challenging. Existing methods largely rely on static or immediate heuristic cues to estimate token importance. We introduce APS-VLA, a framework that learns visual token selection from robot demonstrations while keeping the pretrained VLA backbone frozen. A lightweight Decision Influence Estimator scores tokens conditioned on current visual observations, language instructions, and robot proprioception. Its training combines two complementary signals: action supervision matches both the full-input policy predictions and demonstrated actions, while an auxiliary visual transition objective uses demonstrated actions to predict subsequent visual features from the selected representation. At deployment, the estimator performs a single hard selection and packing step before the policy's native action-generation procedure; the auxiliary predictor is removed. Experiments with OpenVLA-OFT on LIBERO and with $\pi_{0.5}$ and XR-1 on RoboCasa365 demonstrate inference acceleration with modest changes in task success. Across five real-world tasks on two dual-arm platforms, APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline. Together, these results support APS-VLA as a modular, plug-in framework for efficient robotic control across VLA architectures, preserving the underlying policy parameters and native action-generation procedure.
Read the full abstract

All authors Xiaoyang Xing, Chun-Mei Feng, Anthony Cohn, Xiaoliang Gong, Asoke Nandi

Image and Video Processing

Artifact Filtering versus Artifact-Compensated Training in JPEG to JPEG AI Recompression

In this paper, we consider the problem of recompression of images from JPEG to JPEG AI format. We show that direct recompression from JPEG to JPEG~AI trained on uncompressed images is accompanied by the preservation of the JPEG image artifacts in the resulting JPEG~AI image. In order to avoid it, we evaluate an approach where an artifacts removal neural network is first applied to the input JPEG image as an external pre-filter. However, in this case, the processing time is increased by 29 times. Therefore, we propose an artifact-compensated training for JPEG AI and show that it is able to combine pre-filtering against target distortions with efficient compression without increase in the processing time.
Read the full abstract

All authors Igor Podtsepko, Yuriy Kim, Karam Mahfod, Evgeny Belyaev, Kristian Kirs, Jianping Lin, Ilya Papiev

Computer Vision

GTAGym: AAA Game Tasks Environment for Evaluation and Reinforcement Learning

AAA games offer a rich testbed for evaluating agent perception, reasoning, and control. However, evaluating whether agents can complete tasks in AAA games is challenging and often human-annotated because task outcomes are often difficult to verify automatically, let alone applying it for Reinforcement Learning (RL). We introduce GTAGym, an environment and framework for evaluation and RL on GTA V tasks. Its evaluation benchmark, GTAVBench, derived from GTA V main-story missions, contains 103 online tasks reconstructed across nine task families, alongside 47 image-based shooting and sniping tasks using GTA V imagery. A web server provides access to the online environment and evaluation interface, supporting parallel testing through screenshots and keyboard-mouse actions with automatic engine-state outcome checks. We evaluate six frontier LLMs on GTAVBench with textual task instructions and on its GTAVBench-AGI branch, where instructions appear only in the first screenshot and a harness automatically compresses interaction history. The best configuration achieves 70.9% overall accuracy in GTAVBench-AGI, yet aerial navigation, evasion and interception remain difficult. GTAGym also provides a scalable generator that samples task parameters within user-specified ranges and option sets under fixed templates, creating batches of 3D and image-based shooting environments for RL with automatic reward and termination signals. LoRA RL of Qwen3.8-27B on generated on-foot navigation tasks improves the in-domain performance and transfers to out-of-domain task categories, and native GTA V tasks from 15% to 23%.
Read the full abstract

All authors Jiajun Chen, Shicheng Fang, Jiahao Lu, Yuxin Wang, Xipeng Qiu

Human-Computer Interaction

TypedInk: Knowing When Not To Draw

Can a scientific figure generator know when it does not have enough evidence to draw? Complete-source illustrators map a finished methodology section to a schematic, or a table to plotting code, and are never asked this question because their input is assumed complete. The research record is rarely that complete: an outline, a draft paragraph, a table, or a one-sentence intent, in any combination, and painting measured numbers with an image model when the table is missing is a hallucinated result. TypedInk infers a figure type from {diagram, plot, plate} from a possibly incomplete source and dispatches to the renderer that can honor that type's evidence contract, refusing dispatch when the contract cannot be met. On the six gold plots in TypedFigBench (5 views, n=30), TypedInk's code path inks every gold table token when a table is observed (12/12), refuses the 12 hidden-table views instead of inventing numbers, and marks 6 intent-only views illustrative, with no execution failures. An always-draw coder with no such gate inks 30/30 charts, including all 18 with no table, and recovers zero gold numbers on any of them; two vision-language raters (not humans; human replication is future work) still judge those always-draw charts submittable 17/36 times, against TypedInk's refusal 0/36 times. A diagnostic field ablation of the type router (16 illustrations, 5 views each, n=80) locates where this failure would originate if unchecked: an outline alone yields type accuracy 0.31 and sends every gold plot into an integrity violation, and a chat-model planner narrows but does not close the gap (overall accuracy 0.85, still integrity-failing on 25% of outlines). Diagram-label recall is 0.47 on the full record and 0.19 on an outline. Gold plates are typed at dispatch (oracle integrity-fail 0.00) and rendered as stills.
Read the full abstract

Software Engineering

Repository-Context-Guided Image Understanding for Agentic Software Development

Coding agents increasingly receive UI screenshots as part of bug reports, yet degraded image quality - caused by low resolution, compression artifacts, or small rendered text - can prevent them from identifying the relevant UI elements and source files. Vision-language models become less reliable as image quality deteriorates, while OCR-assisted approaches may propagate recognition errors from degraded inputs. Both approaches are limited when the visual evidence required for localization is no longer clearly observable in the query screenshot. We propose augmenting coding agents with repository-rendered reference images: crisp screenshots generated from the same codebase that the agent is tasked with modifying. These references provide readable visual evidence of the application's UI while preserving the current, potentially bug-containing implementation state. Because they are generated from the available repository rather than requiring a known bug-free version, they can be produced automatically for reproducible application states without manual image annotation. We evaluate the approach on a controlled benchmark derived from SWE-bench Multimodal by injecting visual UI bugs into a real web application. We vary screenshot degradation and reference-image provision and measure whether the coding agent identifies the correct source file. The results show that reference images improve file localization especially for degraded image quality, increasing the gold file match rate under fourfold downsampling with JPEG compression. We confirmed with further analysis that the proposed approach recognizes text approximately 12 percentage points better than language biased OCR under the same degradation.
Read the full abstract

All authors Fumiki Yohena, Toshiaki Sota, Shota Suzuki, Tomoko Okamoto, Michiaki Tatsubori

Representation Theory

Representations of the Loop mirror BMS-Kac-Moody algebra

In this paper, we construct a new family of non-weight modules $\Omega(\lambda,\zeta,\alpha,\beta,\gamma,\sigma, h)$ over the loop mirror BMS-Kac-Moody algebra $\mathcal{G}$, which is a Takiff-type extension of the loop mirror Heisenberg–Virasoro algebra. We establish criteria for irreducibility, determine their isomorphism classes, and classify all $\mathcal{G}$-modules that are free of rank one. Moreover, we determine all self-extensions of these modules: for every such module $M$, $\operatorname{Ext}^{1}_{U(\mathcal{G})}(M,M)\cong \mathbb{C}^{6}$, and we construct explicit $1$-cocycles realizing every cohomology class. Every nonsplit self-extension of a simple module is indecomposable of composition length two. Restricting these modules to the loop mirror Heisenberg–Virasoro subalgebra $\mathcal{H}$ yields new $\mathcal{H}$-modules. Among them the $\sigma$-type modules $\Psi(\lambda,\zeta,\sigma,\alpha,\beta,\gamma,h)$ constitute a new family with no counterpart in the non-loop or non-Takiff case. For the degenerate case $\sigma=0$, we establish criteria for irreducibility and isomorphism, and describe certain maximal submodules and their simple quotients. Finally, we construct and classify a new class of tensor product $\mathcal{H}$-modules.
Read the full abstract

Artificial Intelligence

GRAP: Gender Rubric-Aware Probing for Analyzing Gender Bias in LLMs

As LLMs increasingly mediate everyday interactions, subtle gender biases in open-ended responses pose growing societal risks. Existing probing methods have two key limitations: they rely on fixed-format tasks, such as pronoun resolution or occupation association, that alignment-tuned models can easily handle; and they compare raw group-level rubric scores, ignoring per-dimension variance and obscuring which bias dimensions reliably distinguish gender groups. We introduce GRAP (Gender Rubric-Aware Probing), a framework that uses a lightweight logistic probe to identify the most discriminative bias dimensions from responses scored across 10 rubric dimensions, operating task-agnostically on open-ended instruction following.
Read the full abstract

All authors Seona Seo, Geonhee Ahn, Hyunsoo Cho, Sookyung Kim

High Energy Physics (Phenomenology)

Multi-TeV vector dark matter in the $U(1)_X$ Higgs portal model

We investigate heavy-mass vector dark matter within the framework of the $U(1)_X$ gauge Higgs portal vector model, by jointly integrating experimental constraints from relic density measurements, direct detection searches, and indirect detection observations alongside the perturbative criterion to derive bounds on the model parameter space. Within the weakly interacting massive particle scenario, we assess the relative contributions of $s$-channel and $t/u$-channel processes to thermal dark matter annihilation. For parameter configurations lying in the $s$-channel resonance region, each fixed value of the scalar mixing angle $\sin\alpha$ corresponds to a well-defined upper limit on the heavy dark scalar mass $m_{h_2}$. This upper limit decreases monotonically with increasing $\sin\alpha$, with a critical threshold of $\sin\alpha \lesssim 0.05$ beyond which perturbative constraints entirely exclude the full resonance corridor. By contrast, the non-resonant parameter regime dominated by $t/u$-channel annihilation processes is found to be excluded for $m_V \simeq 1$ TeV and above under the combined joint constraints from relic density requirements and perturbative consistency. Our results indicate that the perturbative consistency of the scalar sector, enforced via the two-loop renormalisation group equations, imposes bounds on the parameter space in the heavy mass regime that are comparable to, and in some benchmark regions more restrictive than, current experimental limits. Nevertheless, even under the application of these stringent perturbativity constraints, the model still admits a viable set of parameter points spanning a broad multi-TeV mass range.
Read the full abstract

All authors Kun-Peng Chen, Han Hu, Hao-Yu Luo, Yun Wei, Ming Zhong

Artificial Intelligence

Understanding-First AI Engineering: Purposes, Reference Architecture, and Architectural Patterns for Human Understanding

AI systems can produce reliable results without developing human understanding. However, human understanding remains valuable for shaping subsequent AI-supported work, forming preferences, governing collective choices and satisfying curiosity, even when human participation is unnecessary for correcting errors. We propose understanding-first AI engineering, which treats purpose-specific human understanding as a first-class design principle guiding requirements, design and operation. The purposes of understanding are organised into four complementary rationales: reflective agency, effective action, collective governance and epistemic value. A reference architecture connects production, verification, use, and learning and exploration through shared records, supporting understanding through explanation and participation. Five candidate architectural patterns guide choices about parallel learning, explanatory models, goal exploration, audience-specific views and distributed expertise. Qualitative scenario analysis of two hypothetical use cases demonstrates how the purpose classification, reference architecture and selected patterns guide investigation, participation and decision processes while retaining operational controls. The cases illustrate design consequences and limitations rather than establish empirical effectiveness. The approach makes human understanding an explicit engineering commitment whose justification need not depend on improving the original AI task.
Read the full abstract

Statistical Methodology

Cluster-weighted modeling for count data with excess zeros and mixed-type covariates

Count data may exhibit excess zeros, within-group overdispersion, and heterogeneous relationships between the response and its covariates. Conventional zero-inflated regression models typically impose a single count-regression structure, whereas Poisson count components cannot accommodate residual within-component overdispersion. We propose a zero-inflated negative binomial cluster-weighted model that combines a structural-zero component with multiple negative binomial regression components and jointly models the response and mixed-type covariates.We establish sufficient conditions for identifiability of the proposed model and its Poisson counterpart and develop an expectation-maximization algorithm for estimation. Numerical studies demonstrate improved regression coefficient estimation and clustering performance relative to the Poisson counterpart under within-component overdispersion.
Read the full abstract

Artificial Intelligence

ChipOmni: A General-Purpose AI Agent for Electronic Design Automation

Electronic design automation (EDA) workflows remain highly fragmented, requiring engineers to coordinate heterogeneous representations, specialised software tools, and iterative verification across multiple design stages. Although large language models have brought more general reasoning and code-generation capabilities to EDA, many existing approaches remain focused on isolated tasks and struggle to plan, verify, and repair designs autonomously within executable engineering environments. Here we introduce ChipOmni, a knowledge-enhanced, context-evolving AI agent framework for EDA. ChipOmni integrates multimodal understanding, dynamic planning, multi-agent collaboration, and verification-driven tool interaction into a unified autonomous system. By continually incorporating feedback from external engineering tools, the framework maintains an evolving task context, integrates domain knowledge, and iteratively refines intermediate designs. Across benchmarks for analog circuit, RTL, testbench, and formal-assertion generation, ChipOmni improves multiple task-specific measures over the evaluated baselines and solves all 30 AnalogCoder-Pro tasks. In a long-horizon case study, ChipOmni advances a natural-language RISC-V specification through three design levels to a micro-SoC GDSII layout that passes physical sign-off, without human intervention after initialization. Together, these results show that ChipOmni can coordinate verification-driven design across heterogeneous tasks and sustain long-horizon engineering workflows, advancing a shift from fragmented, tool-centred processes towards agent-centred, closed-loop systems.
Read the full abstract

All authors Pengkun Wang, Rui Mao, Zhe Zhao, Dongting Xie, Haibin Wen, Song Xu, Di Wu, Xu Wang, Zhengyang Zhou, Binwu Wang, Xiaofeng Cao, Yang Wang

Distributed Computing

Two-Tier Task Offloading for Satellite-Assisted Marine Networks: A Hybrid Stackelberg-Bargaining Game Approach

The proliferation of maritime activities has spurred the emergence of numerous computation-intensive and delay-sensitive marine applications and services. Given the inherent rationality, selfish nature, and limited computational abilities of marine devices, devising effective strategies to incentivize their participation in task processing become a critical challenge. In this paper, we investigate the satellite-assisted marine multi-access edge computing (MEC) and propose a two-tier task offloading scheme through a hybrid Stackelberg-Bargaining game approach to enhance offloading efficiency and maximize the utility of marine devices. Specifically, for the underwater acoustic communication, we consider the scenario where multiple autonomous underwater vehicles (AUVs), managed by maritime autonomous surface ships (MASSs), upload their collected data using non-orthogonal multiple access(NOMA) to optimize channel utilization. For the data transmission above the sea surface, we consider the scenario where a low earth orbit satellite (LEOS) functions as a space edge server to provide computing services, and MASS offloads workloads to LEOS through frequency division multiple access(FDMA) to prevent co-channel interference. we define the utility of AUVs, MASSs and LEOSs, and model the offloading process between AUVs and MASSs as a Stackelberg game, while representing the offloading interaction between MASSs and LEOSs as a Bargaining game. Additionally, we propose efficient algorithms to optimize AUV offloading strategies and MASS pricing strategies, while refining the bidding strategies for both MASSs and LEOSs. Simulation results demonstrate that the proposed algorithms significantly outperform benchmark schemes in achieving optimal solutions.
Read the full abstract

All authors Zhen Wang, Bin Lin, Qiang Ye, Haixia Peng

Analysis of PDEs

Counterexamples to the buckling eigenvalue-ratio conjecture in every dimension

In 1999, Ashbaugh formulated the conjecture that the ball maximizes the ratio of the first two clamped buckling eigenvalues among bounded domains in Euclidean space. We disprove this conjecture in every dimension $d\ge2$ by constructing smooth connected domains with finitely many interior holes. The key observation is that a zero-value constraint on a concentric sphere close to the boundary of a ball increases the first two eigenvalues by the same relative amount to first order, whereas their ratio increases strictly at second order. A quantitative surface penalty and logarithmic cutoffs transfer this strict inequality to smooth perforated domains. The construction gives explicit geometric parameters depending only on the dimension \(d\).
Read the full abstract

Artificial Intelligence

Multi-Future Drafting: Mixture Block Drafters for Tree-Verified Speculative Decoding

Block drafters such as DFlash draft a block of future tokens in one forward pass, and draft trees such as DDTree verify the prefixes ranked highest by the product of the drafter's per-position distributions. Under tree verification, the expected number of accepted tokens is the probability mass the tree covers. We show that a product misallocates it: when the continuation can follow several futures, it splices them into chimeras, prefixes stitched together from different futures that the target never produces, and gives them as much budget as the futures themselves. Barring ties, no product-ranked tree can hold two futures that disagree in $m$ positions without $2^{m-1}-1$ chimeras, however well the drafter is trained, and in practice chimeras make up 27-33% of a 64-node tree. We propose Multi-Future Drafting (MFD), which restores the coupling with a latent variable instead of the chain rule: $K$ learned mode embeddings at the drafter's input draft $K$ coherent blocks in the same pass, and a mixture DDTree lets the verifier pick the future the target takes, losslessly and with $K+1$ extra vectors. On four Qwen targets from 4B to 27B parameters, at batch size 1, MFD reaches up to 7.8$\times$ speedup, 5-8% faster than the best single-mode tree trained identically and, on Qwen3-4B and Qwen3-8B, 9-16% faster than the released DARTree trained on the same prompts.
Read the full abstract

All authors Haoyu Huang, Jialiang Kang, Linlin Yang, Xiao Yang, Baochang Zhang

Machine Learning

Permutation-Based Pre-trained Model Merging via Reciprocal Orthogonality

Ito et al. (2026) showed that sufficiently wide, independently trained neural networks can exhibit linear mode connectivity (LMC) with softmax temperature calibration, even without permutation. LMC means that a low-loss linear path connects the models' parameters. This behavior is explained by layerwise exponentially weighted connectivity (LEWC), whereby each merged layer outputs an exponentially weighted sum of the original models' corresponding representations, yielding ensemble-like predictions. However, prior work observed LEWC with substantial widening, such as 8192 units per hidden layer for MLPs and $16\times$ width for VGG-11, raising the question of whether such widths are necessary. We therefore search for permutations that encourage reciprocal orthogonality (RO), which suppresses interference between each model's weights and the other model's representations. This allows both models' computations to coexist in the merged network, promoting LEWC at smaller widths. We formulate the search as independent layerwise quadratic assignment problems and approximately solve their continuous relaxations using Frank–Wolfe. Experiments show that our method facilitates LMC with less width expansion. It also merges models trained on different data distributions, where weight matching, which maximizes weight similarity through permutations, struggles. The merged predictions approximate those of a logit ensemble of the original models.
Read the full abstract

Artificial Intelligence

From Questions to Decisions: Benchmarking How LLM Agents Acquire Decision-Relevant Information from Human Beings

LLM agents are increasingly asked to make plans for people, such as scheduling drivers or assigning nurses. However, people's requests are often incomplete, and the supporting information for decision-making may be outdated, requiring agents to ask for clarification. Existing benchmarks typically provide fully specified problems or rely on uncalibrated LLM-simulated users, leaving this ability insufficiently tested. We introduce Ask2Decide, a benchmark of 43 planning tasks that evaluates whether agents ask for missing information that matters for their decisions. An operator holds facts absent from the agent's request and records, and optimization solvers determine which facts change the optimal plan. We also include similarly worded facts that leave the optimal plan unchanged to test whether agents distinguish useful information from irrelevant details. We collect 79 interaction sessions from 19 participants and find that, after an incorrect plan, human replies omit the correcting fact far more often than replies from an uncalibrated simulated user (43% versus 7%). Using these interactions, we calibrate a simulated operator that more closely matches human reply patterns. Evaluating four frontier LLMs, we find that their initial plans are infeasible on 58-98% of tasks when information is missing, compared with 5-28% when all facts are provided upfront. Agents uncover facts that change the optimal plan only slightly more often than facts that leave it unchanged (39-52% versus 32-41%). These results show that obtaining decision-relevant information through interaction remains a major bottleneck for LLM planning. This is ongoing work and will be updated. The date of this version is shown on the first page.
Read the full abstract

All authors Yuhan Tang, Ke Fan, Xinkai Zou, Baichuan Mo, Kangxin Cui, Ao Qu, Ou Zheng, Jiangbo Yu, Xuan Jiang, Haris N. Koutsopoulos, Jinhua Zhao, Yuan Zhou

Image and Video Processing

Is a CNN-Transformer Ensemble Worth It When a Chest X-Ray Foundation Model Is Available?

Ensembles of ImageNet-pretrained convolutional networks and Vision Transformers are a common way to improve multi-label chest X-ray classification. We ask whether such an ensemble is still worth its cost when a model pretrained on chest radiographs is available. On the official patient-disjoint split of NIH ChestX-ray14 (25,596 test images, 2,797 patients), we train five ImageNet-pretrained backbones and the EVA-X-S foundation model with validation-only model selection, compare five fusion rules, and estimate every difference with a patient-level paired bootstrap. EVA-X-S alone reached a test macro-AUROC of 0.828; the five-model ensemble reached 0.814 (difference -0.0136, 95% CI -0.0178 to -0.0098) at about 7.8 times the summed backbone latency and ten times the parameters. Averaging the five ImageNet models with EVA-X-S changed AUROC by -0.0056 (-0.0095 to -0.0021). With all five ImageNet backbones trained with the EVA-X-S recipe, their ensemble scored 0.817, still -0.0109 (-0.0142 to -0.0076) below EVA-X-S, and across three seeds EVA-X-S varied by at most 0.001 AUROC. Under the ImageNet models' recipe, however, EVA-X-S fell to 0.798. Our full-stacking configuration never outperformed calibrated averaging. In this setting, the tested ImageNet ensembles did not improve on a single chest X-ray foundation model fine-tuned with its own recipe.
Read the full abstract

Machine Learning

FaST: Factored Synthesis of Transitions for Offline Reinforcement Learning

Synthetic data augmentation promises to scale offline reinforcement learning beyond fixed logged datasets, but leading diffusion-based generators model the transition tuple (state, action, next state, reward, terminal) as a monolith: no single factor can be natively redrawn in isolation. We present FaST (Factored Synthesis of Transitions), which factors transitions via the chain rule into state, action, dynamics, reward, and terminal stages, fitting each with an estimator matched to its statistical properties. Across twelve D4RL locomotion and maze benchmarks, TD3+BC and IQL agents trained on FaST transitions match the SynthER diffusion baseline while generating transitions over 90x faster. The explicit action-given-state conditional can also be re-queried for a fixed state, enabling targeted action synthesis: drawing N candidate actions per generated state and keeping the highest-advantage one under a value function fit to the same data, thereby synthesizing improved behavior during generation rather than filtering for it post-hoc. Best-of-10 synthesis beats training on real data on 11 of 12 medium and medium-replay settings (+8.9 D4RL points on average, up to +28.4), and is competitive with advantage-filtered SynthER at a fraction of its generation cost. On expert mixtures both mechanisms degrade identically, tracing the failure to the advantage estimate itself rather than to either targeting mechanism.
Read the full abstract