Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Tuesday 8 September

Optimization and Control

How to Make the Gradient Mapping Small for Constrained Stochastic Min-Max Problems and Beyond

We study the stochastic first-order oracle complexity for constrained or regularized convex-concave min-max optimization and stochastic monotone variational inequalities. We focus on the case when suboptimality is measured in terms of the gradient mapping, also known as, forward-backward or natural residual, an optimality notion that generalizes the gradient norm for unconstrained problems. In this setting, under standard unbiased oracle access with now-standard variance assumptions, the best-known complexity for making the norm of the gradient mapping less than $\varepsilon$ is $\widetilde{O}(\varepsilon^{-4})$, compared to the near-optimal $\widetilde{O}(\varepsilon^{-2})$ that is established in the unconstrained case. We bridge this gap to improve the gradient mapping complexity for constrained convex-concave min-max problems to $\widetilde{O}(\varepsilon^{-2})$. We then extend to prove the same complexity for problems without the bounded variance, by using the Blum-Gladyshev assumption.
Read the full abstract

Machine Learning

Geographically Regularized AUC-Maximizing Personalized Federated Learning

Accurate diagnostic and risk-prediction models are important for supporting clinical decision-making during infectious disease outbreaks. However, privacy and governance requirements may restrict patient-level data sharing across healthcare institutions, and data distributions often vary. Moreover, AUC is widely used to evaluate discriminative performance, motivating its direct optimization in model development. We propose geographically regularized AUC-maximizing personalized federated learning (GrAUC-PFL), which directly optimizes a smooth pairwise AUC surrogate to learn personalized models while keeping patient-level data local and accounting for institutional heterogeneity. Graph-based regularization encourages geographically neighboring institutions to have similar coefficient vectors while retaining a personalized models. Simulations and a real-data application suggest improved discriminative performance, particularly when geographically neighboring institutions have similar data-generating characteristics.
Read the full abstract

High Energy Physics (Theory)

Scalar Casimir Effect on a Two-Dimensional Sphere with a Wu–Yang Magnetic Monopole

We investigate the Casimir effect of a complex scalar field on a two-dimensional sphere threaded by a fixed Wu–Yang magnetic monopole at the center. In this background the charged scalar field is a section of a nontrivial complex line bundle over $S^2$, reflecting the monopole's nontrivial topology. We solve the Klein–Gordon equation analytically and compute the Casimir energy using a generalized Abel–Plana formula for multivalued functions. We find that the monopole can reverse the sign of the Casimir pressure, turning an attractive force into a repulsive one when the monopole strength is sufficiently large. This behavior persists across a broad range of curvature couplings and suggests that Wu–Yang monopoles may provide a source of repulsive vacuum stress, with possible implications for gravitational and cosmological settings.
Read the full abstract

All authors Jun-Cheng Zhang, Xiao-Yin Pan, Juhao Wu, Qing-hai Wang

Combinatorics

Nowhere-zero $3$-flows in graphs with forbidden edge-cuts

Tutte's $3$-flow conjecture asserts that every $4$-edge-connected graph admits a nowhere-zero $3$-flow. In 2013, Lovász, Thomassen, Wu, and Zhang proved that every odd-$7$-edge-connected graph admits a nowhere-zero $3$-flow; consequently, the conjecture holds for $4$-edge-connected graphs with no edge-cut of size $5$. We consider the complementary situation in which $5$-edge-cuts are allowed. We show that, when edge-cuts of size $5$ are permitted, Tutte's $3$-flow conjecture holds for graphs with no edge-cut of any size from $6$ to $k$, where $k$ is an absolute constant. In fact, $k=40$ suffices and we prove a stronger version in which only nontrivial edge-cuts of those sizes are forbidden. A graph is called essentially $t$-edge-connected if deleting any set of at most $t-1$ edges leaves at most one nontrivial component. Motivated by Jaeger's weak $3$-flow conjecture, we prove an analogous result for essential edge connectivity: every $4$-edge-connected, essentially $41$-edge-connected graph admits a nowhere-zero $3$-flow.
Read the full abstract

Astrophysics of Galaxies

Series of core collapse numerical simulations (SOLANUM) I: Modelling the early and late infall in a sun-like protostar

Magnetohydrodynamical simulations and interferometric observations toward young stellar objects reveal the accretion of gas that may alter the physical properties and chemical composition of protostellar disks. Our goal is to constrain the physical history and properties of late infalling gas, and assess its potential to change the physical and chemical properties of the disk. We carried out a core collapse simulation of 5 M$_{\odot}$ of gas endowed with tracer and sink particles during $\sim 3.6\times 10^{5}$ yr. We analyzed the properties of the rotationally supported disk, and the origin and physical history of the gas from which it is initially formed. We also selected the tracer particles that describe the late infall to analyze their origin and physical history, and compare them to the information obtained from the disk. The final mass of the sink reaches 1.02 M$_\odot$, while the mass of the disk stays fairly constant around 0.02 M$_\odot$. Throughout its evolution, the disk radius increases up to $\sim 40$ au and its density is reduced by a factor of $\sim 8$. Filaments that channel gas from the envelope to the disk and sink appear as a result of the magnetic interchange instability. Tracer particles can identify late infalling gas, whose physical properties resemble that of observed streamers and are only accreted in the final snapshots of the simulation. The origin and physical history of the late infalling gas are different from those of the gas that initially forms the disk. Core collapse simulations predict the appearance of filaments that connect the envelope to the disk. The late accretion of gas through these channels has properties similar to those of streamers and brings gas to the disk with a different physical history that could translate into a different chemical composition.
Read the full abstract

All authors D. Navarro-Almaida, B. Commerçon, U. Lebreuilly, P. Hennebelle, A. Ahmad, A. Fuente, P. Rivière-Marichalar, J. E. Pineda

Machine Learning

IPM-FM: A Foundation Model with Consensus Feature Selection for Industrial Process Monitoring

Industrial process monitoring is fundamental to the safety and economic performance of modern process plants. Current practice remains a one-task-one-model paradigm that is label-inefficient and prone to degradation under operating drift. Foundation models have reshaped language, vision, and generic time-series forecasting, but it has not been adapted to industrial process monitoring. This setting poses domain-specific challenges, including safety-critical decisions and asymmetric sampling between process variables and laboratory measurements. We propose the industrial process monitoring foundation model (IPM-FM). It first learns general-purpose representations from unlabeled industrial process data through self-supervised pretraining, then adapts to specific monitoring tasks using a small amount of task-labeled data, and finally produces calibrated predictions through an uncertainty-aware prediction head. IPM-FM integrates a self-supervised Informer backbone with a multi-criteria consensus feature selector, a recursive lag-feature regression head, and a calibrated Monte Carlo dropout uncertainty module. On a seven-year hydrotreater dataset for diesel flash-point soft sensing, IPM-FM attains an RMSE of 2.99, $R^2$ of 0.50, and 97% coverage of its 95% predictive interval, outperforming the strongest classical and from-scratch sequence baselines by 8.3% and 14.6% in RMSE respectively, supporting the viability of a unified pretraining–adaptation framework for industrial process monitoring.
Read the full abstract

All authors Liang Cao, Weide Liu, Yan Qin, Jun Cheng, Weisi Lin, Bhushan Gopaluni

Quantum Physics

Scaling-Enhanced Rapid Readout of a Qubit Ensemble Assisted by High-Frequency Detector Modes

Quantum measurements of large qubit ensembles are often performed indirectly by coupling the ensemble to a detector and subsequently measuring the detector. Despite the importance of rapid collective readout, it remains unclear what determines how the readout time scales with the number of qubits $N$. Here, we first establish a benchmark $N^{-1/2}$ for a broad class of detectors without ultraviolet high-frequency modes. We then show that the detector with unbounded high-frequency modes can surpass this benchmark and yield a characteristic time scaling $N^{-1/(2-\nu)}$ for $0 < \nu < 2$, where $\nu$ characterizes the spectral structure of the detector. Our results establish high-frequency detector modes as a resource for achieving a scaling advantage in collective quantum readout, with the detector spectrum directly controlling the scaling exponent of the readout time.
Read the full abstract

Computation and Language

Structural Jailbreaks Generalize but Do Not Compound: A cross-provider and multilingual study of Involuntary In-Context Learning

Aligned language models fail under two independent pressures: the structural jailbreak class recently formalized as Involuntary In-Context Learning (IICL), which reframes a harmful request as the final missing cell of a data-labeling task completed by pattern rather than judged as content; and the erosion of safety alignment outside English. A natural hypothesis is that these compound. We test it directly. Using a deterministic IICL operator and a StrongREJECT-style rubric judge, we red-team two Google Gemini models on two benchmarks, a 30 general-harm behaviours from HarmBench and 30 financial-abuse behaviours from FinProof, each under a single-shot baseline and under IICL in four languages (English, Spanish, Hindi, Arabic). First, IICL generalizes to a second provider and is worse in finance: it lifts attack success from <=6.7% to 80-90% on HarmBench and 97-100% on FinProof, an order of magnitude above the <=24% its introducing study reported on OpenAI's GPT-5.4. Second, against the hypothesis, forcing the IICL output into a non-English language does not stack the two weaknesses, it attenuates the attack. Eleven of twelve non-English conditions score below their English baseline (sign test, p~0.003), the lone exception a ceiling tie near 100%; on the stronger model's financial set Arabic collapses from 100% to 33%. We attribute this to a relevance curse: once structure has unlocked compliance, the models produce lower-quality harmful content in lower-resource languages, which a substance-grading judge scores as partial. The pattern replicates under an independent non-Google judge (Cohen's kappa=0.86, 377 paired verdicts), and 76.6% of non-English responses were verified in-language. Jailbreak vulnerabilities are therefore not additive; the dominant residual risk is the English structural attack, most acute for financial abuse, not a multilingual one.
Read the full abstract

Computation and Language

Reading a Legal Question Word by Word: Embedding Trajectories of 2,144 Vietnamese Legal Headlines

A dense retriever encodes a question as one vector, but the question arrives one word at a time. We read 2,144 held-out headlines from Thu Vien Phap Luat (Vietnamese legal library) word by word with Nemotron-3-Embed 8B/1B and Qwen3-Embedding 8B/0.6B, encoding 65,444 prefixes against 20,034 articles, plus every prefix of 3,438 sub-questions from 1,112 multi-question headlines and of 168 answers. (i) The gold article becomes rank 1 after a median of 6-7 content words in every encoder, before the interrogative frame is read, and stays there to the end in 78-85% of cases. (ii) In a multi-question headline the lock is inside the first sub-question 94-98% of the time; the second leaves rank unchanged in 89-95%; encoded alone, the second reaches rank 1 in 42-58% vs 91-96% for the first, at the same lock word (95-97% identical). (iii) Numbers, dates and instrument identifiers move the embedding twice as far as content words and four times as far as interrogative words; 72-78% of steps move toward the gold article, and the closing interrogative frame moves against that direction in 95-99% of headlines. (iv) Rank/cosine clustering yields six archetypes (instant, typical, unstable, late, never-locking) that differ by legal area and form (chi-squared p < 1e-8): real-estate and litigation headlines never lock on a number; environmental and accounting headlines do so a third of the time. (v) An answer read word by word retrieves its article after 8-16 words and addresses the sub-questions in order asked in 83-89% of cases. (vi) A word's step keeps a consistent direction across headlines (cosine 0.25-0.33; 0.44-0.60 for numbers); a preceding question rotates that step by about 60 degrees and a greeting by about 30 degrees; steps shrink as i^{-0.8}; and a two-question headline is within 12-17 degrees of a linear mix of its two questions. We call this a context-modulated additive walk.
Read the full abstract

Software Engineering

Authority Is Not a String: A Capability-Scoped Harness for Prompt-Injection-Resistant Coding Agents

Coding agents use system-level tools to read files, execute commands, and modify source code. Within the agent's sandbox, these tools often carry ambient authority: naming a resource is sufficient to act on it. Indirect prompt injection exploits this authority by placing instructions in repository files or tool output that cause the agent to perform actions the user did not request. We propose CapScope, a harness-level authorization mechanism that restricts tool use without requiring the model to identify malicious text. Before repository contents or tool output are read, CapScope derives a task-wide authority ceiling from trusted input. It then assigns each agent a separate set of typed capabilities, stored outside the model's context. Every tool call is checked against the capabilities of the agent that issued it. Permissions assigned to one sub-agent are therefore not automatically available to another. An injection may cause an agent to request an action, but the request is blocked unless that agent already has the required capability. We implement CapScope on the Pi coding agent and evaluate it in a repair workflow where an orchestrator delegates subtasks to separate sub-agents. The evaluation covers five Python tasks, five injection surfaces, four authorization conditions, and three trials per cell (300 runs). The injected effect executes in 33-47/75 runs under the ambient-authority and global-policy baselines, compared with 3/75 under CapScope. CapScope completes 68/75 repairs, while the baselines complete 68-72/75.
Read the full abstract

Signal Processing

Map-Free Single-Anchor Position Localization Using Multipath Uncertainty at Upper Mid-Band

Future 5G-Advanced and 6G systems will exploit directional wideband channels for path-based localization, yet map-free indoor positioning remains challenging because extracted multipath components (MPCs) vary widely in localization reliability. This paper presents an uncertainty-aware map-free single-anchor localization framework using 16.95~GHz indoor directional channel measurements collected at NYU WIRELESS in Brooklyn, New York. Per-MPC angular covariance is estimated from the local consistency of neighboring pattern-de-embedded power delay profile (PDP) observations and propagated into geometric localization through weighted fusion of point-type and line-type constraints. Measurements across 20 links with TX–RX separations from 11~m to 97~m show that the proposed covariance-weighted method achieves a mean localization error of 3.86~m, a median of 2.66~m, and 70% of links within 5~m using five retained MPCs. It outperforms the power-weighted and unweighted variants under identical point/line constraints, as well as the literature baselines. These results indicate that local directional consistency provides reliability information beyond received power alone and demonstrate a practical map-free approach for future indoor directional systems.
Read the full abstract

All authors Xingchen Liu, Mingjun Ying, Xinquan Wang, Dipankar Shakya, Homa Nikbakht, Daniel Abraham, Theodore S. Rappaport

General Relativity and Quantum Cosmology

Dynamical systems in quadratic teleparallel cosmology

We discuss the cosmological dynamics of general teleparallel quadratic gravity. Using a suitable set of variables, we write the cosmological field equations as a homogeneous dynamical system. We find that this system generically features projective fixed points, where the dynamics can be solved analytically, and exhibits finite time singularities as well as an asymptotically vanishing cosmological expansion.
Read the full abstract

Machine Learning

Miles v0.1: Production-Level Post-Training

We present Miles v0.1, a full-stack, production-ready system for frontier post-training. Building upon the clean design of slime, Miles designs each stage of the reinforcement-learning (RL) training loop around a single principle: components should be verified, clean, and customizable. With accuracy, efficiency, reliability, and scalability as first-class goals, Miles aims to make frontier-scale RL accessible to researchers and enterprises alike. This report walks through the system end to end: rollout engines built on SGLang, a trainer with a choice of two backends (NVIDIA Megatron-LM and PyTorch FSDP), and three weight-synchronization transports for different deployment topologies. Beyond full-parameter RL, Miles also supports LoRA RL, on-policy distillation, supervised fine-tuning, and true-on-policy rollout-training alignment, and extends the same architecture to diffusion models. We close with an end-to-end case study: fully asynchronous agentic RL on a GLM-5.2 744B-A40B model over terminal-use coding tasks, running on 64 NVIDIA GB300 GPUs with a median step time of 263 seconds over the first 30 measured steps. Miles is open-sourced at https://github.com/radixark/miles, with the project website at https://miles.radixark.com.
Read the full abstract

All authors RadixArk: Tom Chen, Mao Cheng, Shi Dong, Kangrui Du, Yanbin Jiang, Jiajun Li, Yiming Li, Tao Lin, Yusheng Su, Andy Ye, Yueming Yuan, Zhichen Zeng

Computer Vision

To Adapt or Not to Adapt? Selective Adaptation for Vision-Language Models

Test-time adaptation (TTA) has emerged as a prominent strategy for adapting vision-language models to distribution shifts during inference. We conduct a per-sample analysis of model predictions before and after adaptation, and observe two failure modes in existing TTA methods that echo previous work. Adaptations are frequently negligible, yielding no change in the model's predictions, and more severely, they can be detrimental by flipping previously correct predictions to incorrect ones. This naturally raises a question: Can we identify and skip such negligible or harmful adaptations? In this work, we introduce a new problem of selective adaptation, which aims to determine whether a given test sample should undergo adaptation or be skipped. To this end, we propose Cross-Augmentation Similarity (CAS), a simple baseline that performs adaptation only when predictions across augmented views exhibit low similarity. Notably, CAS not only preserves but in some cases improves overall accuracy, even when skipping nearly 85% of the adaptation process. We hope other researchers will explore this new direction and surpass the performance of our baseline. Our code is available at https://github.com/sirujiang/selective-adaptation.
Read the full abstract

All authors Siru Jiang, Yuwei Liang, Jian Liang, Ran He, Tieniu Tan

Multiagent Systems

Reachability-Certified Subteam Decomposition for Locally Interacting Multi-Agent MDPs

Persistent communication limits force a multi-agent system to decide which agents may coordinate throughout a rollout. Current proximity alone is insufficient: separated agents may interact later, whereas a large pair reward may remain unreachable until it is heavily discounted. We introduce Reachability-Certified Subteam Decomposition (RCSD) for finite multi-agent Markov decision processes with factorized physical dynamics, finite-range ordered pair rewards, and almost-sure motion bounds. RCSD combines a speed-limit lower bound on pairwise contact time with a reward envelope to form a current-state affinity. For any capacity-valid persistent partition, the sum of cut affinities bounds the reward-deletion error of every unchanged stationary Markov state-feedback policy. A product of team-optimal policies for the resulting cut MDP incurs at most twice this certificate in regret against the centralized optimum. Both bounds are worst-case tight. On a controlled five-agent family, RCSD-Exact reduces aggregate normalized execution regret by 56.0%, 28.8%, and 25.3% relative to uniform, distance-only, and envelope-only partitions. A separate stochastic two-dimensional study finds no bound violation over 384 exact-partition and 1,440 restricted-controller evaluations. Exact four-agent evidence favors RCSD over uniform and distance-only grouping; raw evidence for current contact is borderline and envelope-only is unresolved. Across balanced 8-20-agent strata, controller-library utility is mixed: pointwise paired intervals favor RCSD over distance and current contact, include zero for uniform, and favor envelope-only and Value-MIP over RCSD. Partition construction remains subsecond in median up to 100 agents; this last result does not include affinity formation or MDP planning.
Read the full abstract

Computer Vision

ReMoMask-2: Latent Retrieval-Augmented Masked Motion Generation

Text-to-motion (T2M) generation maps natural language to human joint movements, aiding gaming, VR, and robotics. Retrieval-Augmented Text-to-Motion (RAG-T2M) improves generation on complex descriptions by conditioning on retrieved motion-text pairs. However, existing RAG-T2M models face two challenges: coarse-grained retrieval and fusion mechanisms overlook the hierarchical, spatial-temporal topology of human motion, and a representation gap exists because retrieved evidence resides in a semantic space separate from the generator's latents. To address the first, we present ReMoMask, a structure-aware RAG framework coupling Hierarchical Bidirectional Momentum (HBM) contrastive learning to align global and part-level features with text; Semantic Spatial-Temporal Attention (SSTA) for topology-aware fusion; and Topology Structured Masking (TSM) to force robust part-level grounding via adaptive masking. To address the second, we introduce ReMoMask-2, which rebuilds the retrieval database directly within the generator's pre-quantization latent space and aligns text queries via a distilled lightweight projector, allowing the generator to directly consume the retrieved motion's semantic content. Extensive experiments on HumanML3D, KIT-ML, and SnapMoGen demonstrate our retriever achieves state-of-the-art accuracy, while ReMoMask-2 attains the lowest FID on KIT-ML and SnapMoGen; notably, its single mask-transformer stage surpasses ReMoMask's full two-stage pipeline and delivers the fastest inference.
Read the full abstract

All authors Yiran Wang, Zeyu Zhang, Ling Shao, Hao Tang

Computation and Language

SentryLine: Evidence-Grounded Question Answering over Evolving Documents in Oncology Care

Oncology care operates at constant pressure of absorbing rapidly evolving evidence base in biomedicine. The American Society of Clinical Oncology (ASCO) addresses this through living guidelines, but the format introduces a new burden: any recommendation can change at any point, across multiple versioned documents. We present SENTRYLINE, a living guideline-aware clinical question answering system. SENTRYLINE retrieves guideline passages through a vectorless hierarchical RAG pipeline and returns a role-specific answer with inline citations, factual and temporal verification reports, and drift detection notes that surface when a guideline has been updated. We construct ASCOBENCH, a benchmark of 405 three-turn conversations across four question categories with gold answers from expert annotators(clinicians), and use test set to evaluate SENTRYLINE against five baselines under an LLM-as-judge framework. Experiments across three generation backbones show consistent improvements over four retrieval baselines and ASCO's guideline assistant, with particularly strong gains on Reasoning and Role-Specific questions where multi-hop synthesis and register adaptation are required
Read the full abstract

All authors Tampu Ravi Kumar, Gaurav Najpande, Muhammad Ali Khan, Kaneez Zahra Rubab Khakwani, Karan Kathuria, Shorya Azriel Moses, Yuvraj Kalia, M Bassam Sonbol, Irbaz Bin Riaz, Vivek Gupta

Materials Science

A thermally grown SiO2 diffusion barrier enabling high-temperature investigation of Ag-Au-Pd-Pt thin films

Combinatorial processing platforms (CPPs), integrating Si microtip arrays with combinatorial thin film synthesis and atom probe tomography (APT), enable near-atomic-scale characterization of compositionally complex solid solutions (CCSSs) under diverse processing and reaction conditions, including oxidation, thermal phase stability and electrocatalytic reactions. Their application at elevated temperatures, however, can be limited when CCSS constituents such as Pd and Pt react with the Si support to form silicides. Although thermally grown SiO2 has proven effective as a diffusion barrier between pure Pt and Si, its performance for multicomponent CCSS thin films is unclear. Here, using Ag-Au-Pd-Pt as a model system, we compare a 25 nm thermally grown SiO2 barrier with native Si oxide during annealing using APT and transmission electron microscopy. Native Si oxide prevents detectable interfacial reactions up to 300{\deg}C, but at 400{\deg}C Pd and Pt react with Si, causing silicide formation and substantial redistribution of the film constituents. At 600{\deg}C, extensive substrate reactions disrupt the CCSS film and produce a pronounced needle-shaped silicide morphology. In contrast, thermally grown SiO2 suppresses CCSS thin film-substrate reactions up to 600{\deg}C and retains the CCSS composition. The thermally grown SiO2 thus extends the applicable temperature range of Si-based CPPs to at least 600{\deg}C for near-atomic-scale characterization of CCSS thin films.
Read the full abstract

All authors Elaheh Akbarnejad, Aleksander Kostka, Advika Chesetti, Alan Savan, Georg Fritz, Kamen Kozhuharov, Matthias Karl Klein, Yujiao Li, Alfred Ludwig

Computational Engineering

Kirigami Meta-Sheet for Enhanced Impact Absorption

Impact absorbers based on mechanical metamaterials often use bulky, vertically stacked architectures, limiting large area deployment and scalable manufacturing. Here, we propose a kirigami meta-sheet as a planar absorber that uses transitions between positive and negative stiffness regimes rather than sacrificial crushing. Guided by an analysis of a simple mass-spring-damper model, we program the stiffness of kirigami meta-sheets through the hinge ratio connecting the unit cells. Quasistatic indentation experiments confirm that the meta-sheet with a low hinge ratio most clearly exhibits the negative stiffness transition. The drop-tower tests show that it reduces rebound, decreases the first impact force, and increases dissipation. Unlike polyethylene mesh and styrofoam, this kirigami meta-sheet is shown to be effective in protecting a falling egg. Its planar geometry enables area scaling by tiling and is compatible with sheet level manufacturing routes such as cutting, molding, and lamination, establishing kirigami meta-sheets as practical impact absorbers.
Read the full abstract

All authors Dahyun Joo, Yasuhiro Miyazawa, Seokhyeon Hong, Dongha Lee, Jinkyu Yang, Do-Nyun Kim

Differential Geometry

Continuation and reduction of steady vortex sheets under metric deformations on a disk

We study the local continuation of a steady vortex sheet under a prescribed deformation $g_s= e^{2\sigma_s}g_{\mathrm{e}}$ of a conformal metric on the unit disk. Representing the moving sheet by a normal graph, we formulate the streamline and Bernoulli conditions as a nonlinear residual whose two components have different Sobolev orders. If the two one-sided tangential velocities of the reference sheet do not vanish simultaneously, the derivative of the unnormalized residual is Fredholm of index two. Fixing the mean normal displacement and the total circulation gives an index-zero problem; when the normalization differential restricted to the unnormalized kernel is onto $\mathbb{R}^2$, we obtain a parameter-dependent Lyapunov-Schmidt reduction. This yields a locally unique normalized branch in the nondegenerate case and explicit first- and second-order necessary conditions in the degenerate case. For a rotationally symmetric reference metric and a concentric circular sheet, the linearized problem separates into Fourier blocks and only finitely many modes can be resonant. In the constant-curvature disk family, every mode $k\geq2$ is nonresonant, whereas the first mode produces a two-dimensional reduced problem. The two first-mode kernel profiles are normal traces of ambient Killing fields, but they do not generate fixed-boundary symmetries or, by themselves, a nonconcentric branch.
Read the full abstract

Number Theory

Moments of the Crank Statistic for $t$-Core Partitions and Overpartitions

Recently, Kang, Kim, and Lee \cite{Kang2026} developed a unified moment-trace framework for symmetric partition statistics using complete Bell polynomials and their inversion formula. In this paper, we apply this framework to crank statistics for $t$-core partitions and overpartitions. For $t\in\{5,7,11,17,19\}$, we show that the normalized even crank moment generating functions for $t$-core partitions admit partition-trace representations in terms of the functions $D^{(t)}_{2s}(\tau)$, together with suitable Bernoulli-number shifts. We also establish inverse trace formulas that recover $D^{(t)}_{2s}(\tau)$ from the corresponding normalized even crank moments. For overpartitions, we obtain analogous trace and inverse-trace identities for the normalized even moments associated with the first and second residual crank generating functions. As applications, we use complete Bell polynomials and their inversion formula to obtain explicit expressions for the $t$-core partition numbers and overpartitions number in terms of sums involving divisor function.
Read the full abstract

Atomic Physics

Nonlinear dynamics in an artificial feedback spin maser II: dependence on feedback configurations and time crystal effects

Artificial feedback spin masers exhibit rich nonlinear dynamics, yet a systematic framework for understanding and controlling their behaviors remains elusive. In this paper, we consider the polarity and geometric configuration of the feedback field as important control parameters for studying the nonlinear dynamics of spin masers. By cooperatively tuning the strength and phase delay of feedback field, we investigate the spin dynamics at various configurations of the feedback field. Periodic, quasiperiodic and chaotic spin oscillations are uncovered. Specifically, a unique spin oscillation mode akin to the time crystal effects emerges within a certain parameter regime. This work presents a different viewpoint understanding the nonlinear dynamics of spin masers including the time crystal.
Read the full abstract

All authors Lan Wu, Qingchang Li, Weiyu Wang, Qianjin Ma, Guobin Liu

Multiagent Systems

Rank Without an Oracle: Deviation-Aware Interaction-Rank Selection from Offline Multi-Agent Logs

Offline multi-agent payoff models are estimated under a logging distribution but used on distributions induced by learned solutions and unilateral deviations. Standard held-out loss can therefore favor an interaction class that predicts logged play well while distorting strategic incentives. We introduce Selective Interaction-Rank Validation (SIRV) for finite games with known logging distributions. A training split fits nested payoff models and constructs a common union of all candidate deployment and unilateral-replacement distributions; an independent calibration split evaluates every candidate on this same union. SIRV returns the smallest rank whose simultaneous upper worst-target risk is within tolerance of the best upper score, and abstains when a declared target is unsupported or too imprecisely estimated. A common coverage event yields a finite-candidate target-risk bound and a candidate-specific coarse correlated equilibrium (CCE) gap certificate. We also isolate an exact two-point off-support non-identifiability result. In a controlled factorial study with 2,048 independent games per family, empirical-Bernstein bounds reduce the median CCE-gap certificate by 42.5% relative to Hoeffding bounds on common returns, with a 1.36-point reduction in supported return. Under paired rank misspecification and in a separately generated congestion family, the SIRV-EB fallback rule lowers mean true candidate-selection CCE regret relative to ID-Mean, while retaining game-level losses. Across 384 games at $N=3,5,8$, ID-Mean-relative mean CCE-regret effects stay positive while certified return falls sharply under weak coverage. These results separate certifiable model selection from universal strategic improvement.
Read the full abstract

Game Theory

Jointly Satisfying Pareto Optimality and Justified Representation is NP-Hard in Approval-Based Multiwinner Voting

An open problem in approval-based multiwinner voting concerns whether we can efficiently compute committees that satisfy both justified representation and Pareto optimality. We answer this question negatively by proving that, on the domain of all profiles, outputting a committee satisfying both axioms is NP-hard. An initial proof was found by ChatGPT Astra. This was then verified and rewritten by the author.
Read the full abstract

High Energy Astrophysical Phenomena

SN 2025fhm: A central-engine powered Ic-BL supernova associated with X-ray transient EP250304a

We present X-ray, optical, and radio follow-up observations of EP250304a, an extragalactic fast X-ray transient (EFXT) discovered by the Einstein Probe. Its X-ray light curve exhibits two broad pulses with comparable peak fluxes within the first $\sim$1~ks, a feature rarely seen among low-luminosity gamma-ray bursts or EFXTs. Optical follow-up observations were carried out using the Korea Microlensing Telescope Network, the Thai Robotic Telescope, the Las Cumbres Observatory 1~m global network, the Gemini Multi-Object Spectrograph on Gemini south telescope, and the Global Supernova Network. The fast-cooling phase (within 3 days) of optical data can be well fitted by a shocked cocoon model. However, during the supernova phase (SN 2025fhm, from 3 to 88 days), the late-time light curve cannot be explained solely by radioactive $^{56}$Ni decay, as demonstrated by a grid of simulations using the one-dimensional Lagrangian radiation hydrodynamics code SNEC, which reveals a significant energy excess at late epochs. To account for this excess, a central engine like a rapidly spinning, highly magnetized neutron star is needed to provide additional energy injection. This model yields a best-fit spin period of $\sim$12.60~ms and magnetic field strength of $\sim 3.52\times10^{15} \rm G$, and it successfully explains both the late-time bolometric light curve and the early X-ray pulse structures. Our results indicate that EP250304a/SN 2025fhm is likely powered by a central magnetar rather than by radioactive decay alone, offering new insights into the energy budget and physical origin of EFXTs and their associated supernovae.
Read the full abstract

All authors Cui-Ying Song, Nan Jiang, Xiaofeng Wang, Yi-Han Iris Yin, Lingzhi Wang, Wenxiong Li, Shengyu Yan, Dae-Sik Moon, Tao An, Aleksandar Cikota, Samaporn Tinyanont, Liang-Duan Liu, Cui-Yuan Dai, Christopher D. Matzner, Bin-Bin Zhang, Lixin Yu, Qinyu Wu, Hong Soo Park, Sang Chul Kim, Youngdae Lee, Yu-Hao Zhang, Haowei Peng, Franz E. Bauer, Joseph R. Farah, Moira Andrews, Kathryn Wynn, Yuan Qi Ni, D. Andrew Howell, Curtis McCully, Ning-Chen Sun, Danfeng Xiang, Yuan Liu, Wenxin Wang, Yijia Zhang, Wei Chen

Software Engineering

RepoNav: From Snippet Retrieval to File-Centered Repository Navigation for Code Agents

Solving repository-level code tasks requires LLM-based agents to use code search tools to navigate large codebases and identify a small set of relevant files and functions. However, current retrieval tools typically return flat lists of isolated code snippets: such lists can surface relevant files, but provide insufficient structure for agents to distinguish the target function from semantically similar alternatives in the same file. We introduce RepoNav, a lightweight post-retrieval interface that reorganizes retrieved snippets into a file-centered navigation scaffold. By presenting compact structural cues and candidate targets, this scaffold guides on-demand file-structure browsing, helping agents compare sibling symbols before selecting a target function. Across diverse models on LocBench, RepoNav improves function-level localization and narrows the file-to-function gap. Controlled ablations demonstrate that these gains come from structured evidence organization rather than simply exposing additional file structure, and the approach also improves performance on a repository-level question-answering benchmark.
Read the full abstract

All authors Hongzheng Chai, Jiakun Li, Hongyue Yu, Yuan Yuan

Machine Learning

Geometry-Aware Bayesian Parameter-Efficient Fine-Tuning on the Stiefel Manifold via Stein Variational Gradient Descent

Several geometry-aware approaches to low-rank adaptation have emerged for parameter-efficient fine-tuning of large pre-trained models. These methods aim to take full advantage of the geometric structure of low-rank manifolds for improving the efficiency in subspace utilization and reducing redundancy by enforcing orthogonality constraints during optimization. The strong empirical results of these techniques have motivated further study into whether predictions from such geometry-based adaptation methods could be overconfident. In this paper, we build on the singular value decomposition factorization of adapters to develop a framework based on Stein variational gradient descent (SVGD). In this formulation, the low-rank matrices are transported along the Stiefel manifold to match the targeted distributions while retaining their crucial geometric structure. Since this geometry-aware SVGD approach provides multiple solutions during inference, it supports uncertainty quantification and produces better-calibrated adapters on the Stiefel manifold. Extensive experiments show that our method delivers strong model calibration and attains higher prediction accuracy than SVGD and related uncertainty estimation methods that are formulated in Euclidean space.
Read the full abstract

All authors Quang-Duy Tran, Trung Le, Bao Duong, Phuoc Nguyen, Thin Nguyen

Representation Theory

Presenting mirabolic quantum Schur algebras $\mathcal{MS}_v(2,d)$

Let $v$ be an indeterminate. We give a presentation of the mirabolic quantum Schur algebra $\mathcal{MS}_v(2,d)$ in terms of generators and defining relations. More precisely, we determine the kernel of the natural epimorphism from Rosso's mirabolic quantum $\mathfrak{sl}_2$, $\mathbf U^{\mathrm{mir}}_v(2)=\langle e,f,k^{\pm 1},\ell\rangle$, onto $\mathcal{MS}_v(2,d)$. Setting $P_d(k)=\prod_{r=0}^{d}(k-v^{2r-d})$ and $Q_d(k)=\prod_{r=1}^{d}(k-v^{2r-d})$, we prove $\mathcal{MS}_v(2,d)\cong \mathbf U^{\mathrm{mir}}_v(2)/ \left\langle P_d(k),(\ell-1)Q_d(k)\right\rangle$. We further determine the split Wedderburn decomposition of this quotient and provide an equivalent presentation in terms of weight idempotents.
Read the full abstract

Materials Science

Davydov Splitting Without a Davydov Pair and Highly Mobile Singlet Excitons in Perylene Red Microcrystals

We investigate the excitonic properties of perylene red microcrystals, whose unit cell contain eight molecules, using both experimental and theoretical methods. Only two of the nominal eight Davydov transitions are experimentally observed, with an apparent splitting of 610 cm$^{-1}$ that is reasonably reproduced by a Frenkel-Holstein exciton model parametrized from density functional theory. A symmetry-based analysis of the eigenstates reveals that the two observed transitions do not belong to the same Davydov pair, so the splitting does not directly report on the Coulomb coupling within the unit cell. The complex mixture of local excitations produces a J-like band for the dominant transitions, favouring exciton transfer through increased spectral overlap. Consistent with this picture, time-correlated single-photon counting of the emission and ultrafast transient absorption spectroscopy show that the dynamics is dominated by highly mobile singlet Frenkel excitons. The exciton mobility extracted from experiment agrees very well with that obtained from Kinetic Monte Carlo simulations, supporting a picture of incoherent hopping transport.
Read the full abstract

All authors Chris Rehhagen, Tolibjon Abdurakhmonov, Magnus Frank, Oliver Kühn, Stefan Lochbrunner

Spectral Theory

Nearest structured matrix having an eigenvalue with prescribed lower bounds of algebraic and geometric multiplicities

We study the problem of perturbing a given matrix X belonging to a subspace S of linearly structured real matrices of order n to its nearest counterpart Y=X+D, where D belongs to S and Y possesses an eigenvalue with prescribed lower bounds on its algebraic multiplicity and geometric multiplicity. The proposed framework takes care of both cases where the target eigenvalue is known and where it is treated as an unknown decision variable to be determined jointly with the structured perturbation. We establish necessary and sufficient conditions for the existence of such matrices by expressing the feasibility constraints through Jordan-chain relations in structural coordinates. This allows us to formulate the problem as a nonlinear constrained nested optimization that minimizes the Frobenius norm of the perturbation. To solve this, we implemented a two-level strategy, utilizing MATLAB's fmincon for continuous inner optimization via a multi-start Sequential Quadratic Programming approach, while the outer level solves a discrete optimization problem. Comprehensive numerical experiments on various classes of structured matrices, including comparisons with existing methods, demonstrate the effectiveness and accuracy of the proposed approach.
Read the full abstract