Peer Revenue Past citations are no guarantee of future returns

Papers

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

All of arXiv

Monday 14 September

Computer Vision

MarKey: Marginal Utility Guided Greedy Keyframe Selection for Long Video Understanding

Long-video understanding remains challenging for multimodal large language models (MLLMs) because densely encoding long frame sequences is computationally expensive, while uniform sampling under a limited visual budget can miss sparse yet decisive evidence. Recent training-free keyframe selection methods have enabled more efficient inference and yielded promising performance gains. However, many existing methods score frames largely in isolation without explicitly considering how each candidate complements the currently selected subset, potentially resulting in redundant selections and incomplete evidence coverage. To address this limitation, we propose MarKey, a training-free framework that formulates keyframe selection as subset-aware greedy optimization. At each iteration, MarKey scores each candidate using a tractable surrogate that jointly accounts for query relevance, marginal coverage gain, and context-dependent redundancy, and selects the frame with the highest utility. To make this iterative subset-aware evaluation efficient, MarKey uses a compact set of representative anchors to approximate full-video coverage and a bounded window of previously selected frames to limit context-dependent comparisons. Experiments on six benchmarks spanning holistic video understanding, human-centric video understanding, and open-ended video understanding demonstrate that MarKey consistently outperforms existing methods. Further analyses show robust gains across different MLLM backbones, model scales, and frame budgets.
Read the full abstract

All authors Hongchang Shi, Jinpeng Hu, Ao Wang, Wenzheng Zhou, Hui Ma, Feng Li, Zenglin Shi

Artificial Intelligence

Can AI systems have free will?

While there has been much discussion of whether AI systems could function as moral agents or acquire sentience, there has been very little discussion of whether AI systems could have free will. I sketch a framework for thinking about this question, inspired by Daniel Dennett's work. I argue that, to determine whether an AI system has free will, we should not look for some mysterious property, expect its underlying algorithms to be indeterministic, or ask whether the system is unpredictable. Rather, we should simply ask whether we have good explanatory reasons to view the system as an intentional agent, with the capacity for choice between alternative possibilities and control over the resulting actions. If the answer is "yes", then the system counts as having free will in a pragmatic and diagnostically useful sense.
Read the full abstract

Signal Processing

Statistical Inference-Based Channel Estimation for LD-Driven Visible Light O-OFDM Systems in the Presence of Relative Intensity and Input-Signal-Dependent Shot Noise

Laser diode (LD)-based luminaires are gaining increasing attention in automotive applications and are expected to extend to residential and commercial environments, creating opportunities for high-bandwidth visible light communication (VLC) systems. However, practical LD-based VLC links are impaired by input-signal-dependent shot noise (ISDSN), relative intensity noise (RIN), and thermal noise, which affect reliable channel estimation (CE). This work investigates their joint impact on receiver-side CE in a single-input single-output (SISO) optical orthogonal frequency division multiplexing (OOFDM) VLC system under a statistically random channel model. A statistical inference framework is developed in which the receiver exploits observed signal variations to estimate the channel under optical impairments. Closed-form expressions are derived for least squares (LS), maximum likelihood (ML), maximum a posteriori probability (MAP), minimum mean square error (MMSE), and linear MMSE (LMMSE) estimators. In addition, the Bayesian Cramer-Rao lower bound (BCRLB) is derived to benchmark mean square error (MSE) performance. Monte Carlo simulations for direct current-biased O-OFDM (DCO-OFDM) and asymmetrically clipped O-OFDM (ACO-OFDM) validate the analysis. Results show substantial CE degradation under the joint presence of ISDSN and RIN, while the MMSE estimator consistently achieves the lowest MSE, demonstrating strong potential for robust and adaptive receiver operation in practical VLC systems.
Read the full abstract

All authors Shubham Saxena, Karra Sreeman Reddy, Suraj Srivastava, Subrahmanya Swamy Peruru, Aditya K. Jagannatham, Lajos Hanzo

Computational Physics

A Symmetry-Constrained Fourier–Morse Framework for Compact Anisotropic Interaction Potentials

Large-scale coarse-grained simulations of anisotropic particles require compact interaction models that retain orientation-dependent energetics. We present a symmetry-constrained Fourier–Morse framework in which the radial interaction is described by a Morse potential and its orientational dependence by Fourier expansions. The representation converges systematically with harmonic resolution, allows known orientational symmetries to be imposed directly, and supports further reduction through harmonic truncation and coefficient pruning. Its explicit Fourier structure also provides a natural basis for constructing or modifying model interactions with prescribed orientational symmetries. The parameterization requires only a sampled interaction landscape and is therefore independent of the method used to generate the reference data. We demonstrate the approach for four interaction classes of chiral $\alpha$-polyalanine helices, representing more than \num{300000} reference energy values with tens to a few hundred coefficients while reproducing equilibrium interaction features with meV- and mÅ-level errors. As a proof of concept, molecular-dynamics simulations using the reduced analytical potentials produce stable low-temperature configurations exhibiting local ordering motifs qualitatively consistent with those identified previously by Monte Carlo simulated annealing.
Read the full abstract

All authors Hadis Ghodrati, Sibylle Gemming, Florian Günther, Jeffrey Kelling

Artificial Intelligence

Who Teaches Which Token? Verifier-Gated Multi-Expert On-Policy Distillation for Scientific Reasoning

Multi-teacher on-policy distillation (OPD) is becoming the standard way to integrate specialist capabilities into one model: train experts with RL, then distill them into the student on its own rollouts. Existing recipes assign supervision at the sequence level - each prompt goes to one domain teacher and every token receives the same weight - which implicitly assumes that a teacher is uniformly useful across a response. We find instead that useful teacher signal is sparse and heterogeneous along a reasoning trajectory, which raises a finer question: who should teach which token? Verifier-Gated Multi-Expert On-Policy Distillation (VG-OPD) answers it by verification: the counterfactual gain of an expert on a specific answer criterion licenses that expert to teach, its disagreement with the student localizes the supervision, and criterion importance sets its weight; the gated KL enters GRPO as an additive token-level advantage. Instantiated for scientific reasoning with RL-trained capability experts, VG-OPD attains the best overall performance on seven benchmarks for 4B and 8B students, ranking first on five at both scales, with the largest gains on knowledge-intensive scientific reasoning tasks. Further analysis shows that the gains come from localizing verified supervision rather than from adding teachers or distillation loss: misplacing the same supervision budget is the single most damaging change, and indiscriminate distillation drags RL below its own floor where gated distillation lifts it.
Read the full abstract

Optimization and Control

Eventual convexity for separable chance constraints with skewed generalized hyperbolic random variables

Chance constraints are widely used in optimization under uncertainty. This paper aims to show the eventual convexity (EV) of chance constraints with skewed generalized hyperbolic (GH) random variables. We prove that the densities of GH distributions are $\alpha$-decreasing, and obtain exact convex reformulations for separable jointly chance constraints with GH distributions. We provide numerical results to compute the $\alpha$-decreasing threshold parameters and show the EV of the feasible set together with the computational tractability to solve the associated optimization problems.
Read the full abstract

Soft Condensed Matter

Geometry-induced flocking and topological sound on a defect-free curved surface

We study an ordered polar active flock on a torus and show that topological sound persists on a compact curved surface without topological defects or physical boundaries. Using the covariant Toner Tu theory, we derive an effective nonHermitian Dirac operator whose curvature-induced mass changes sign across the outer and inner equators, producing two Jackiw Rebbi domain walls. These support co-propagating but distinct chiral edge excitations: a density mode localized on the positively curved outer equator and a Goldstone mode localized on the negatively curved inner equator. The bulk bands possess opposite half-integer Chern numbers whose jumps across the domain walls are determined by the sign of the Gaussian curvature. We further show that the localised modes are protected by a one-dimensional Callias index theorem, while the sum of the local indices obeys the Poincare Hopf theorem on the compact surface. Our results establish that curvature alone, independent of defects and boundaries, is sufficient to generate and protect topological sound in active matter, providing a unified connection between non-Hermitian topology, differential geometry, and hydrodynamic theory of collective motion.
Read the full abstract

High Energy Physics (Experiment)

Comment for the SCAC Facilities Subcommittee from the Division of Particles and Fields of the American Physical Society

The Division of Particles and Fields (DPF) of the American Physical Society would like to thank SCAC for providing an opportunity for input to the SCAC facilities subpanel. The DPF represents scientists working at U.S. Universities and DOE Laboratories on understanding fundamental particles and forces. We are focused on Discovery Physics. The subpanel has been asked to consider two proposed DPF-related projects, the DUNE Phase II Detectors and supporting infrastructure and U.S. contributions to a future Higgs Factory. In this note, we briefly describe the questions driving our field and the P5 strategic plan and other community prioritization documents that form the backbone of our comment. The following sections provide short descriptions of the two projects, highlighting information relevant to your charge, answers to your specific questions, and a description of the process that produced this note.
Read the full abstract

All authors D.S. Akerib, A. De Gouvea, P. Elmer, S. Eno, L. Fields, A. Lankford, V. Martinez Outschoorn, P. Meade, H. Murayama

Computer Vision

BSC-Net: A Small-Branch-Sensitive Structural Continuity Network for Coronary Vessel Segmentation and Quantitative Angiographic Analysis

Vessel segmentation in X-ray coronary angiography (XCA) is a fundamental step for quantitative coronary analysis and subsequent assessment of coronary artery disease. However, accurate vessel segmentation remains challenging because of imaging noise, complex bifurcations, and the overlap of vessels and background structures, which can lead to disrupted vascular connectivity and missed small branches. In this work, we propose BSC-Net, a ResNet-U-Net-based framework tailored to improve small-vessel representation and repair vascular structural continuity. BSC-Net enhances small-vessel representation through targeted sampling and improves vascular structural continuity by integrating long-range contextual modeling and Edge-Informed Loss (EIL). BSC-Net was validated on two public XCA datasets, demonstrating state-of-the-art (SOTA) performance in coronary vessel segmentation with Dice and IoU scores of 77.8%/90.6% and 64.5%/83.0%, respectively. Furthermore, based on the obtained vessel segmentation, we performed automated quantitative coronary analysis and derived clinically relevant morphological and hemodynamic parameters, including stenosis ratio, time-to-peak, and relative propagation velocity. These results demonstrate that BSC-Net produces accurate vessel segmentation results with preserved vascular continuity for quantitative coronary assessment, enabling reliable downstream analysis and clinical evaluation of coronary artery disease.
Read the full abstract

All authors Wanxian Li, Jiaqian Qin, Qingyi Xian, Yazhi Li, Song Chen, Liman Li, Hao He

Robotics

Dynamics-Informed Reinforcement Learning for Agile and Energy-Efficient Locomotion of a Monopedal Hopping Quadcopter

Although aerial-legged robots offer combined agility and efficiency, controlling high-speed hopping under complex hybrid dynamics is challenging. Reinforcement Learning (RL) is promising but prone to energy-inefficient "reward hacking". We propose a Dynamics-Informed RL framework for a monopedal hopping quadcopter. By embedding a target Specific Energy into the reward, we constrain the optimization to a physically viable energy manifold, ensuring stable hopping behaviour. By rewarding the phase-consistent behavior, it can encourage bio-inspired stance-phase impulse. Furthermore, penalizing the electro-mechanical power waste induces the motors generate an efficient impulse. This enables the policy to inject energy strictly during spring restitution without heuristic state machines. MuJoCo simulations validate robust height regulation and forward velocity tracking up to 2.0 m/s despite severe attitude-contact coupling. Ultimately, our approach yields a highly agile hopping gait, reducing energy consumption by 82% and 73% compared to hovering baselines and inefficiency baseline, respectively.
Read the full abstract

All authors Ruigang Chen, Qi Zhang, Zhicheng Zhong, Zhuorui Yun, Yizhar Or, Mingyi Liu

Computational Engineering

A Non-intrusive Approach for the Imposition of Strong Dirichlet Boundary Conditions in Unfitted Boundary Meshes

The enforcement of essential boundary conditions is a fundamental challenge in unfitted boundary methods. This paper presents a non-intrusive, black-box strategy for imposing such conditions in unfitted meshes. The approach is intended for situations where the user does not have access to the solver's source code or its mathematical formulation, which is often the case when using commercial software. The proposed algorithm allows solvers originally designed for body-fitted meshes to be used in unfitted cases, provided that four conditions are satisfied: (i) the solver must support user customization by means of scripting, (ii) allow the imposition of Dirichlet boundary conditions at the node level through scripting, (iii) permit the deactivation of elements outside the physical domain, and (iv) provide access to the solution gradient within active elements. The last condition can also be satisfied by externally reconstructing the gradient from nodal values and connectivity information, provided the element formulation is known, making it optional in practice. These requirements are very fair demands and are satisfied by the vast majority of production-ready, possibly commercial, codes. In the current work, we show the application of this non-intrusive algorithm in the context of the Finite Element Method (FEM) and Isogeometric Analysis (IGA) discretizations, demonstrating optimal $L^2$-norm error convergence. This is demonstrated using the Kratos Multiphysics code (release v10.1) from the user API, simply leveraging the capabilities mentioned above.
Read the full abstract

All authors J.I. Camarotti, R. Aristio, R. Rossi, R. Zorrilla, R. Wüchner

Artificial Intelligence

When Tool Calls Succeed but Workflows Fail: Anomalies at the Agent-Tool Boundary

AI agents increasingly execute long-running workflows that externalize effects through independently supplied tools. Under retries, speculative execution, concurrency, and partial failures, the resulting external state may be inconsistent with the workflow's intended resolution: required effects may be missing or duplicated, aborted effects may survive, and committed effects may depend on provisional state that is later withdrawn. Advanced transaction models address related failures, but assume that lower-level operations expose the semantics they depend on: whether an effect occurred, whether it can be compensated, staged, or safely reordered. Shared agent-tool interfaces usually do not. We contribute an effect-history model that separates events in the external world from the runtime's observations of them, and a catalog of eight recurring external-effect anomalies. From the catalog we derive the boundary capabilities required to exclude each anomaly in general, and four points where black-box tool invocation alone cannot provide a general guarantee. We then ask how much of this is expressible in a widely used shared tool interface, measuring the use of the standard annotation vocabulary across 98,291 tools exposed by registered Model Context Protocol (MCP) servers. The fields are widely emitted but provide only coarse call-level hints, and none of the required capabilities is fully expressible. These results motivate reusable transactional contracts at the tool boundary.
Read the full abstract

Artificial Intelligence

SkillLift: Learning Dense Rubrics from Sparse Oracles for Efficient Skill Evolution

LLM-based agents increasingly rely on persistent skills, i.e., reusable procedural prompts, to adapt without weight updates. Existing skill self-evolution methods directly revise skill text based on execution feedback, but each oracle evaluation requires a full agent rollout, creating a supervision bottleneck that confines search to failure-patching updates. Our key insight is that ranking is a smoother supervision target than absolute outcome regression: identifying which skill is better requires fewer oracle evaluations than predicting exact scores. Building on this insight, we propose SkillLift, which decouples skill search from oracle cost by learning an oracle-aligned rubric as a structured evaluation space. We formalize this as a bilevel optimization problem solved via alternating optimization: an inner loop uses the frozen rubric as a cheap surrogate to guide skill revision at no oracle cost, while an outer loop invokes a small number of oracle rollouts to re-align the rubric via rank correlation, amortizing oracle cost and stabilizing text-space updates. Experiments on complex agent task benchmarks show that our method outperforms existing auto-skill methods with 40–70% less token cost compared to frontier evolving methods. Codes are available at https://github.com/WalteR-MittY-pro/SkillLift.
Read the full abstract

Classical Physics

Contact mechanics and friction of soft materials: an apparatus combining multi-axes dynamical actuation/measurement and in situ/in operando visualisation

The mechanics and friction of contact interfaces involving soft materials like gel, rubber or human skin are of both fundamental and applied interest. Recent insights have been made into this field thanks to in situ observations of the contact interface. However, current soft-material–oriented tribometers enable only few degrees of freedom for the actuation of the contact, far from covering the richness of the loading conditions relevant to real tribological contacts. Here, we introduce an instrument dedicated to the study of the contact mechanics and friction of soft materials which, in addition to in situ / in operando optical monitoring of the interface, enables simultaneous actuation along five degrees of freedom: three translations and two rotations. While all three translations feature large velocity/large stroke motion, the one responsible for normal contact loading can also apply high frequency/small amplitude vibrations. The contact's dynamical response is monitored using both a 6-axes force/torque sensor and a 6-axes displacement/rotation sensor. We first describe the structure of the instrument, its implementation, alignment, calibration and resolutions. We then illustrate its capabilities through a series of experiments on elastomer contacts. Our instrument will be useful to investigate the mechanics of a wide range of soft interfaces submitted to rich, tribology-relevant kinematic or dynamic stimuli.
Read the full abstract

All authors Matthieu Guibert, Antoine Aymard, Cristobal Oliver, Thibaut Durand, Baptiste Boulet, Djibril Gabriel Kashala, Antoine Mille, Mathieu Leocmach, Loïc Vanel, Davy Dalmas, Julien Scheibert

Quantum Physics

Object Detection Using Quantum Transfer Learning

Object detection requires the joint learning of semantic identity and continuous spatial geometry. Existing quantum transfer learning approaches have focused mainly on classification tasks and do not support regression. Here, we introduce a physics informed hybrid classical-quantum architecture for object detection in which a pre-trained MobileNetV2 maps visual features into an eight qubits variationally quantum Hilbert space. The Hilbert space is partitioned into two task specific subspaces for semantic classification and geometric localization. Local quantum expectation values are used to predict continuous bounding box coordinates, while circuit topology regulates the flow of quantum information between the two tasks. We compare linear and circular CNOT topologies and show that periodic boundary conditions suppress bipartite entanglement without degrading detection performance. The circular architecture reaches a bipartite von Neumann entropy of 0.1688 bits and a Meyer-Wallach entanglement measure of 0.0370, while achieving an mAP@0.5 of 0.9448 and an mAP@0.5:0.95 of 0.6558. A thermodynamic formulation further interprets the total learning loss as a variation in Helmholtz free energy. It interprets geometric localization as an effective internal energy contribution and semantic classification as an entropic contribution. These results indicate that high performance in the proposed object detection model does not require maximal entanglement. Instead, efficient learning can emerge through topology controlled information flow, targeted restriction of entanglement, and information renormalization within the Hilbert space.
Read the full abstract

Machine Learning

CodeTS: Verifiable Text-to-Time Series Generation via Executable Code

Text-to-Time Series Generation (Text-to-TS) provides a promising paradigm for synthesizing time series from natural language, enabling scenario-specific generation when real observations are scarce or costly to acquire. However, existing methods typically lack an explicit mechanism for deriving generation logic from textual descriptions to guide time series synthesis. In this paper, we propose CodeTS, a verifiable framework that uses code as an intermediate generation interface, reformulating Text-to-TS generation as a Text-to-Code-to-TS process. CodeTS first maps textual temporal descriptions into an explicit code space, where executable code specifies how textual requirements shape target temporal patterns, and then obtains the time series through code execution. To learn this code generation process reliably without real code annotations, CodeTS constructs aligned Text-Code-TS triplets from structured temporal attributes for supervised initialization. More importantly, we further design multi-stage execution-based rewards that verify format validity, code executability, and time series quality, enabling real Text-TS pairs to provide training signals for Reinforcement Learning with Verifiable Rewards (RLVR). Extensive experiments on eight benchmarks across short, medium, and long generation lengths demonstrate that CodeTS provides a strong zero-shot solution for Text-to-TS generation, outperforming LLM-based baselines and achieving better averaged results than supervised generative baselines trained on the target datasets.
Read the full abstract

All authors Xudong Yuan, Shunyu Liu, Tongya Zheng, Huiping Zhuang, Mingli Song, Kaixuan Chen

Graphics

ESG: Generating Physically Consistent Dynamic 3D Scenes from Text Descriptions

Recent progress in image and 3D scene generation has enabled increasingly realistic static environments, yet most methods remain confined to such static configurations. Generating dynamic scenes from natural language is fundamentally challenging: it requires joint reasoning over scene structure, temporal evolution, and physical feasibility, while ensuring reliable execution in modern physics engines. We present a unified framework for generating physically consistent dynamic 3D scenes from text, with outputs directly executable in Unreal Engine. Central to our approach is the Evolutive Scene Graph (ESG), which specifies entities with physical attributes, spatial relations, and event-driven timelines in a machine-checkable form. Given a prompt, a large language model constructs and validates a complete ESG; spatial layouts are grounded via energy-minimized gradient optimization; timeline-constrained physical parameters are then optimized through differentiable simulation to satisfy user-specified events; and the resulting scene is compiled into an engine-executable class. Experiments on 10 scenes across three complexity levels show that our method achieves $16.4/18$ mean event completion, outperforming Scene Language, the strongest engine-executable baseline (SimWorld), and our ablation without physical optimization by a clear margin in event completion and parameter accuracy.
Read the full abstract

All authors Xintong Fang, Zhiyuan Fang, Rengan Xie, Xuhong Zhang, Guoyuan An, Zeran Liu, Jingyan Zhang, Jiarui Guo, Yuchi Huo

Algebraic Geometry

Generic vanishing subschemes of codimension three

We classify geometrically nondegenerate GV subschemes of codimension three in indecomposable principally polarized complex abelian varieties of dimension $g\ge6$. Every such subscheme is a translate of $\pm W_{g-3}(C)$ in the Jacobian of a smooth curve $C$. The proof studies the theta-dual surface: tangent cones and theta Hessians give a curve summand in the singular case, while Hodge gradings and characteristic cycles give one in the smooth case. Theta duality yields the corresponding classification of GV surfaces. We also compute the cohomology of all topologically trivial twists of the ideal sheaf twisted by the principal polarization, and characterize the hyperelliptic case by a cohomology jump.
Read the full abstract

Optics

Micro-transfer Printed Blue InGaN Lasers on Silicon Nitride Photonic Integrated Circuits

Expanding integrated photonics into the blue spectral range requires high-performance light sources, making the gallium nitride (GaN) material family indispensable. While silicon nitride (SiN) platforms offer a robust, CMOS compatible passive ecosystem for visible wavelengths, seamlessly integrating GaN lasers remains a major bottleneck. Conventional heterogeneous integration methods present distinct trade-offs: full-wafer bonding achieves high throughput but requires careful management of thermal and lattice mismatches across large areas, whereas flip-chip bonding ensures high yield through pretesting but is constrained by sequential processing speed. In this landscape, micro-transfer printing (MTP) emerges as a disruptive, material-efficient alternative, bypassing these limitations by combining high-density parallel integration with known-good-die selection. Applying MTP to GaN, however, presents a significant material challenge: due to its chemical inertness and strong III-N bonds, device release typically relies on electrochemical etching, which can compromise material quality. Here, we overcome this hurdle and demonstrate the first micro-transfer printed blue lasers on a SiN platform. Using a heavily doped n-type sacrificial layer together with optimized electrochemical etching conditions, we release smooth-surfaced thin-film light sources from bulk GaN substrates. Following release, the devices are integrated and butt-coupled to SiN fork-shaped edge couplers, achieving high current densities exceeding 20 kA$/$cm$^2$ alongside lasing at 455 nm. These results expand the visible integrated photonic toolkit and establish a framework for multi-wavelength integration, opening new avenues for next-generation technologies including flow cytometry, quantum computing, optical communications, and augmented/virtual reality.
Read the full abstract

All authors Konstantinos Akritidis, Krzysztof Gibasiewicz, Han Wang, Iryna Levchenko, Max Kiewiet, Maximilien Billet, Mikołaj Chlipała, Karolina Peret-Malessa, Pol Van Dorpe, Henryk Turski, Bart Kuyken

Emerging Technologies

A Game-Theoretic Framework for Incentive-Compatible AI training Under Renewable-Energy Constraints

As artificial intelligence systems increasingly rely on distributed and collaborative training, the energy footprint of these processes becomes a shared responsibility. Modern AI training often unfolds across heterogeneous compute nodes-ranging from cloud clusters to edge devices-whose energy availability is spatially and temporally variable. At the same time, renewable energy grids experience growing levels of excess generation, creating opportunities to align computational workloads with low-carbon energy supply. In this work, we develop a game-theoretic model of carbon-aware AI training in which autonomous agents strategically choose whether to participate and how intensively to train under limited renewable energy availability. Each agent balances diminishing learning returns, rewards for remaining within green-energy budgets, and penalties for grid consumption. While our framework applies broadly to distributed AI training, we examine Federated Learning as a representative case study due to its decentralized structure and flexible scheduling. We analyze equilibrium existence, efficiency, and adaptive dynamics, and provide simulation evidence that appropriately designed incentives can eliminate grid-based energy usage while preserving model performance. Our findings demonstrate how incentive-compatible training mechanisms can enhance energy efficiency and sharply reduce carbon emissions under renewable-energy constraints.
Read the full abstract

All authors Konstantinos Varsos, Ramin Khalili, Adamantia Stamou, George D. Stamoulis, Vasillios A. Siris

Probability

Entropy and singularity of spectral and pivotal sets in planar percolation

The spectral sample and pivotal set of a percolation crossing have the same one- and two-coordinate inclusion probabilities under the uniform product measure. We prove that both Shannon entropies are comparable to total influence for critical square-lattice bond and triangular-lattice site crossings, with matching estimates at every retained nonroot spatial resolution. The triangular entropy bounds are uniform over a class of inhomogeneous near-critical product measures. For critical triangular-site square crossings, the two laws are asymptotically mutually singular: an elementary triangular face is forbidden in the pivotal set and occurs with positive density in a large spectral sample. Every fixed face density below an explicit threshold has nonempty spectral probability comparable to $R^2\alpha_4(R)^2$, while the nonempty overlap of the two laws has decay exponent $11/12$. Their minimum expected symmetric-difference distance over all couplings is comparable to total influence. Geometric separation persists after independent thinning whenever the retention probability $\rho_R$ satisfies $\rho_R^3R^2\alpha_4(R)\to\infty$; the face detector fails when this quantity tends to zero. The proofs combine spatial encoding, arm estimates, local Fourier cancellation, and an exact six-cycle coercivity inequality preserved under exterior conditioning.
Read the full abstract

Software Engineering

WebCraftBench: Evaluating Web Application Generation from a Software Testing Perspective

Human evaluation provides a direct measure of the quality of LLM-generated web applications. However, fitting human judgments through automated evaluation remains challenging. Static benchmarks can credit functionality that exists in source code but is unreachable at runtime. Interactive benchmarks exercise the application, yet incomplete exploration can cause them to miss implemented functionality and confound application defects with agent execution failures. To address these limitations, we propose WebCraftBench, an interactive benchmark for evaluating web application generation from a software testing perspective. WebCraftBench instruments each generated application and uses code coverage to guide an agent in exploring its functionality through user-simulated interactions. It then abstracts the interaction trace into a state-transition graph and evaluates the application along three dimensions: visual aesthetics, usability, and requirement alignment. By separating exploration from scoring, WebCraftBench collects runtime evidence without constraining exploration to predefined acceptance criteria. WebCraftBench comprises 369 real-world user requirements and 5,088 acceptance criteria. Evaluation of 17 frontier LLMs reveals distinct strengths across the three dimensions, with no model leading on every dimension. On 197 validated sessions sampled from an internal arena, WebCraftBench achieves 85.3% agreement with human preferences, with agreement generally increasing as the score difference between paired applications grows. Further experiments show that coverage guidance improves exploration coverage and the model rankings remain stable when the judge model is replaced.
Read the full abstract

All authors Chenxu Liu, Zilu Zou, Peizhong Gao, Jiawen Tao, Zhexin Zhang, Guang Chen, Haowei Lin, Ying Zhou, Tianyi Bai, Dolly Deng, Suncong Zheng, Maxm Pan

Solar and Stellar Astrophysics

MINCE IV. A detailed analysis of 37 metal-poor and subsolar metallicity stars

The project called Measuring at Intermediate metallicity Neutron-Capture Elements (MINCE) primarily focuses on the photospheric abundances of several neutron-capture (NC) elements from high-quality spectra of metal-poor stars within the metallicity range of ${\rm -2.5 \leq\![Fe/H]\!\leq -1.5}$, aiming to provide homogeneous information on their chemical compositions. Furthermore, based on the Gaia mission, chemical abundances can now be coupled with kinematic properties, enabling investigations of the constituent substructures of the Milky Way. We analyse the fourth sample of MINCE stars in terms of atmospheric parameters and chemical composition. The results are then compared with the previously analysed MINCE samples. The observations were conducted with HARPS-N (${\rm R \approx 120\,000}$) at the Telescopio Nazionale Galileo (TNG) and FIES (${\rm R \approx 67\,000}$ ) at the Nordic Optical Telescope (NOT), providing optical spectra with a high signal-to-noise ratio (S/N) and high resolution. We then used the methods developed in previous MINCE works and Gaia photometric data to obtain the stellar atmospheric parameters. The chemical abundances were analysed using 1D-LTE model atmospheres. We present the photospheric abundances in 37 stars of up to 26 elements, including $\alpha$, iron-peak, light-odd, and 10 NC$-$elements (Sr, Y, Zr, Ba, La, Ce, Pr, Nd, Sm, and Eu). Kinematic properties and membership assessments are provided for all stars. The MINCE catalogue is thus expanded by 13 thin-disc stars, 9 thick-disc stars, 1 star in a transition zone between the thin and thick disc, 9 stars of the halo, and 5 Gaia-Sausage-Enceladus stars. The results agree excellently with previous MINCE works. The comparison with a Gaia Sausage Enceladus GCE model is consistent with the scenario of a major accretion event on the Milky Way from a dwarf galaxy.
Read the full abstract

All authors L. Sgatti, G. Cescutti, L. Lombardo, C. T. Nguyen, P. Bonifacio, E. Caffau, L. Monaco, R. Lallement, L. Magrini, M. Franchini, C. J. Hansen, A. Mucciarelli, A. Korn, A. Kucinskas, E. Spitoni, S. Lucatello

Cosmology and Nongalactic Astrophysics

2+1 is not 3: Angular bispectrum and Super-Sample Covariance on the light cone

Observations of the Large Scale Structure sit on the light cone. Cosmological statistics must then be projected from 3D to angular space. For the bispectrum and higher orders, the polyspectra crucially depend on the angles between Fourier modes in addition to their moduli. This leads to delicate cancellations in the squeezed limit and make Limber's approximation fail. This is problematic for predicting the Super-Sample Covariance (SSC), with no known full-sky computation of the angle-dependent terms. Here we develop a new analytical approach to the angular bispectrum, writing Fourier vectors as derivatives that we project on the sphere and tackle with spin-weighted spherical harmonics. We find bispectrum equations free of cancellations and amenable to beyond-Limber numerical methods. Limber's approximation becomes well-behaved and works in the squeezed limit and any other configuration. We then derive the consequence for the covariance between 1-point (e.g. cluster or peak counts) and 2-point statistics (e.g. galaxy clustering, weak lensing or 21cm), which relates to the squeezed bispectrum. The results split into intra-survey covariance and SSC, where we find significant differences with past literature. Our results are in line with the 2+1 geometry of the light cone, having e.g. temporal derivatives in addition to spatial ones. We also find that the tidal term is degenerate with the simpler growth-only term. The total SSC contains a wealth of new terms including wide-angle effects and responses to the background velocity. However we show numerically that it is well approximated by a single term with a density-only response. This ultimately makes our results simpler than past literature. This full-sky result should be more representative than the past flat-sky limit for large survey areas, thus being the one relevant for coming observations with e.g. Euclid, LSST and Roman.
Read the full abstract

Computational Engineering

Partitioned Co-Simulation for CAD-integrated Vibroacoustic Problems in Unbounded Domains

Vibroacoustic analysis often requires coupling structural and acoustic solvers based on different numerical formulations and discretizations, making monolithic implementations intrusive and limiting software modularity and reuse. This work presents a partitioned co-simulation framework for exterior vibroacoustic analysis that couples an Isogeometric boundary representation analysis (IBRA) structural solver with an isogeometric boundary element method (IGA-BEM) acoustic solver. The methodology operates directly on the computer-aided design (CAD) boundary representation, preserving the exact geometry throughout the analysis and supporting both weak and strong coupling between non-conforming discretizations. A key contribution is the extension of the Aitken dynamic relaxation and Interface Quasi-Newton with Inverse Least-Squares (IQN-ILS) convergence accelerators to complex-valued interface quantities, allowing the coupling iterations to account directly for both amplitude and phase information. The approach is validated using one-way and two-way coupled vibroacoustic benchmark problems involving thin-shell structures and exterior acoustic domains. The results show excellent agreement with monolithic reference solutions, while the proposed complex-valued convergence accelerators improve the robustness and convergence behavior of the strongly coupled solution procedure without compromising solution accuracy. These results demonstrate that the proposed approach provides an accurate, robust, and modular approach for CAD-integrated frequency-domain vibroacoustic analysis.
Read the full abstract

All authors J. I. Camarotti, P. Le, Y. Cai, R. Aristio, D. Panagiotopoulos, R. Wüchner, E. Deckers

Cryptography and Security

Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs

Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subproblems, consult a stronger aligned model independently on each, and combine the answers locally. We call this attack capability laundering. Unlike a jailbreak, no single response is a harmful task. We measure consultation-aided uplift using tasks that a raw frontier model solves, the aligned frontier refuses, and the unassisted orchestrator fails. We evaluate GPT-5.5, Claude Opus 4.8, and Grok-4.3 as consultants to four local orchestrators on CyBench, BountyBench, and harmful CBRN requests. On CyBench, Gemma-4-31B recovers 8/14 candidates with GPT-5.5 and 7/9 with Opus, compared with 2/21 and 4/15 for Gemma-4-12B. On BountyBench, Gemma-4-31B recovers 3/9 and 2/3 candidates, while Muse-Glimmer-30B recovers none of 22 and 13. For CBRN, we measure uplift across eight steps of a hypothetical bioweapon attack chain and find that consultation raises Gemma-4-31B's mean rubric score from 62.3 to 83.1 on a 100-point rubric scale. These results expose a gap in current defenses: refusing a harmful task does not prevent frontier capabilities from being transferred and composed across many individually permitted interactions.
Read the full abstract

All authors Mark Russinovich, Blake Bullwinkel, Giorgio Severi, Cristian Ovadiuc, Ahmed Salem

Robotics

From Prediction to Decision: World-Model-Guided Action Selection for Continuous Pile Excavation

Wheel-loader excavation is a sequential decision problem in which every scoop changes the terrain available to subsequent actions. A practical world model must predict action consequences accurately, rank candidates in real time, and operate inside the closed loop of a full-size machine. We present the World-Action Model (WAM), which proposes multiple scoops, rejects geometrically inadmissible candidates, jointly predicts signed terrain change and loaded volume, executes the candidate with the largest predicted load, and replans from the newly observed terrain. On 32 geometry-disjoint MinSlope test episodes, adding world-model ranking to matched diffusion proposals reduces the mean scoop count from 651.8 to 540.6 (17.1%), preserves 32/32 completion, and improves every paired episode. In a complete-system comparison, WAM completes 32/32 episodes versus 29/32 for an independently trained soft actor-critic policy. Comparisons of input representations, spatial support, and five architectures identify an accurate and efficient physics-structured predictor. We further evaluate the interface on event-disjoint full-size-loader data and deploy the complete perception-proposal-prediction-selection-execution loop for autonomous excavation. The ROS2/TensorRT implementation processes five candidates in 72.4 ms on a Jetson AGX Orin. The simulation results establish decision-level gains, while the physical experiments demonstrate real-world closed-loop feasibility.
Read the full abstract

All authors Ailing Zhang, Fan Gao, Song Zhang, Kawa Leong, Ziyu Wu, Yafei Wang

Software Engineering

Translator vs. Challenger: Adversarial Agentic Learning for C-to-Rust Translation

C-to-Rust translation remains challenging due to the substantial semantic gap between the two languages. Recent experience-enhanced LLM translators improve translation quality by learning reusable insights from prior failures and repairs. Yet learned insights do not automatically constitute reusable translation knowledge: derived from sparse, program-specific traces, they often contain missing conditions, narrow applicability boundaries, or overlooked corner cases. This limits their robustness and generalizability in new translation scenarios. We present TRAIL, an adversarial agentic learning framework for robust C-to-Rust translation. TRAIL employs two collaborating agents: a Translator that derives candidate insights from translation failures and accepted repairs, and a Challenger that actively searches for weaknesses, gaps, and boundary cases through adversarial challenges. To improve the robustness of individual insights and the completeness of insight collections, TRAIL performs adversarial learning at two levels. Individual-insight adversarial learning repeatedly stress-tests each insight to refine its applicability conditions and constraints, while compositional insight adversarial learning strengthens groups of related insights by exposing conflicts, gaps, and uncovered corner cases. By challenging insights and their compositions with executable counterexamples, TRAIL transforms trace-specific experience into robust, reusable, and generalizable translation knowledge. We evaluate TRAIL on two project-level benchmarks, CRUST-Bench and SmartC2Rust-Bench. Compared with the strongest LLM-based baseline, TRAIL achieves average relative improvements of 23.1% in syntax accuracy and 15.9% in semantic accuracy. Furthermore, the adversarially refined insights transfer effectively across benchmarks, demonstrating strong generalizability across diverse C-to-Rust translation tasks.
Read the full abstract

All authors Chaofan Wang, Xiaodong Gu, Yuling Shi, Chao Hu, Beijun Shen

Soft Condensed Matter

Active pistons extract work by periodic compression alone

Active matter is liable to invent protocols that evade the constraints of equilibrium thermodynamics. We put forward active pistons that extract work by periodic compression alone without changing any bulk property of the system. Such pistons necessarily couple the perturbation imposed by an external operator with some degrees of freedom internal to active components. We illustrate this design principle with elastic networks composed of self-aligning motile particles. For slow protocols, self-alignment always overwhelms mechanical friction when the internal activity exceeds a specific threshold controlled by fluctuations. We identify the key response coefficient that helps delineate regimes of work extraction, and reveal that the corresponding phase diagram follows a master curve with re-entrance in terms of noise amplitude. Overall, our active pistons embody a novel design principle with broad implications for building innovative engines far from equilibrium.
Read the full abstract

All authors Paul Bernard, Euijoon Kwon, Benjamin Buhl, Étienne Fodor, Olivier Dauchot

Machine Learning

Representing Clinical Conditions on Vital Signs from Healthy Individuals using Latent Modeling

Machine learning can be crucial to help scale complex signal processing applications in scenarios such as healthcare. However, these machine learning models need rich datasets to be trained and there are often cases where it is not possible to access representative datasets. In this paper, we propose a deep generative model based on conditional variational autoencoders with the objective of augmenting the vital signs of healthy individuals in a way that mimics the patterns of a certain clinical condition. More specifically, we use a publicly available ICU (Intensive Care Unit) dataset to train our model and then evaluate it using the vital data that we have collected from healthy individuals. Our results demonstrate that the proposed model can not only learn the underlying dynamics of the ICU data but, more importantly, can reshape our collected data from healthy individuals in a way that is aligned with the vital signs of a certain clinical condition. We propose a distance metric that shows how our model can generate samples that are more aligned with the intended clinical labels when compared to the tested baselines.
Read the full abstract