Peer Revenue Equity in science

Papers

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

All of arXiv

Friday 28 August

Software Engineering

Rethinking Vulnerability Remediation as a Capacity Allocation Problem

As AI accelerates vulnerability discovery, remediation throughput may become a greater constraint than prioritisation accuracy. This study evaluates vulnerability remediation as a flow-control problem using Apache Jira, Mozilla Bugzilla, Red Hat security errata, five public Jira organisations, and an npm dependency graph. Apache resolution times are strongly heavy-tailed, while 94-100% of arrivals in the primary issue trackers enter queues estimated to be at or above capacity. Queue-context models provide only moderate predictive discrimination and are largely matched by simple project-level baselines. Severity-to-speed discrimination varies substantially across systems. Flow-control analyses show larger operational effects: transitions from overloaded to draining queues are associated with shorter resolution times, severity-first sequencing reduces critical-item delay at fixed capacity, and capacity reservation can reduce prolonged critical-item delays. Owner-level analyses further show that available capacity is useful only when it is located where demand occurs or can be transferred through relevant expertise connections. These findings support treating vulnerability remediation as a flow-control and capacity-allocation problem rather than solely a ranking problem.
Read the full abstract

Computation and Language

Phoneme- and Word-Level Metrics Using Self-Supervised Speech Representations for Forced Alignment Evaluation

Forced alignment evaluation typically requires manually annotated timestamps, limiting large-scale and multilingual analysis. We introduce two corpus-level metrics based on self-supervised (SSL) speech representations for reference-free forced alignment evaluation: Phoneme-Cluster Mutual Information (PCMI) and Word Acoustic Consistency Score (WACS). PCMI measures agreement between aligned phoneme labels and clusters induced from SSL-speech representations, while WACS measures consistency of repeated word realizations using dynamic time warping similarity between word representation sequences. Using both random and systematic perturbations, we show that PCMI and WACS degrade consistently under alignment perturbations. We further analyze the metrics across multiple alignment systems on 85 languages from FLEURS, validate them against manually annotated alignments from 45 languages in DoReCo, and evaluate them on two phonologically complex low-resource languages. The metrics effectively separate high- and low-quality alignments and correlate strongly with timestamp-based alignment quality measures. Our results demonstrate that SSL-speech representations enable scalable, reference-free forced alignment evaluation. The metrics are available as an open-source Python package at https://github.com/mahesh-ak/forced-aligner-metrics.
Read the full abstract

High Energy Astrophysical Phenomena

Relativistic outflows power a quasi-periodic eruption: constraints on energetics, mass loss, and emission mechanisms

Quasi-periodic eruptions (QPEs) are recurring bursts of X-ray radiation originating from supermassive black holes (SMBHs). They are an unprecedented type of structured, high-amplitude SMBH variability, but the physical origins of their regularity, timescales, energetics, and emission are uncertain. We present new XMM-Newton observations of the QPEs in ZTF19acnskyy/"Ansky", constituting the deepest observations of individual bursts in any source thus far. The X-ray spectra reveal time-evolving P Cygni profiles comprising blueshifted absorption and redshifted emission from L-shell transitions of Fe XIX-XXIV, with column densities $N_H\sim 10^{22-23}$ cm$^{-2}$ and bulk velocities of $|v_w/c|\sim 0.2$, indicating relativistic mass ejections during each eruption. We construct a time-dependent analytical model of a wind turning on to self-consistently compute its evolving luminosity and ionization properties, and find that the light curve and spectral lines can be simultaneously produced by a wide-angle outflow with $\dot{M}\sim 10^{-9}-10^{-8}\,M_\odot$ s$^{-1}$ kinetically powering the X-rays with an efficiency of $L_X/\dot{E}_K\sim 0.1$. For a covering fraction $f_\Omega\sim0.5$, each eruption ejects $\sim 10^{-3}\,M_\odot$ and $\gtrsim 10^{49}$ erg of kinetic energy, setting an upper bound on the QPE lifetime of $\lesssim1000$ bursts if the underlying mass reservoir is $\sim1 M_\odot$, and implying that the bursts may result in detectable multiwavelength signatures of reverberation and feedback. These measurements provide new quantitative constraints on QPE energetics, emission mechanisms, and the mass/energy they recycle into their circumnuclear environments, as well as an observational probe for direct comparison with physical models and hydrodynamical simulations of QPEs.
Read the full abstract

All authors Joheen Chakraborty, Erin Kara, Wenbin Lu, Brian D. Metzger, Peter Kosec, Riccardo Arcodia, Itai Linial, Olivia Aspegren, Ehud Behar, Sudip Bhattacharyya, Margherita Giustini, Lorena Hernandez-Garcia, Daniel Kasen, Giovanni Miniutti, Frits Paerels, Erwan Quintin, Claudio Ricci, Daniele Rogantini, Paula Sanchez-Saez, Fatima Zaidouni

Quantum Physics

Langevin Theory of Non-Markovian Quantum Dynamics: Application to Delayed Coherent Feedback and the Laser Linewidth

Phase-space methods are powerful tools for the treatment of Markovian open quantum systems: they map the reduced dynamics of a system S, in interaction with an environment E, exactly onto Langevin equations for c-number stochastic variables, as opposed to Heisenberg-Langevin equations for operators. Langevin equations provide analytical insight in key regimes and excel at handling strong nonlinearities and couplings, where other methods often falter. Extending phase-space methods to non-Markovian dynamics, however, has remained a long-standing challenge. Here we address this gap by applying phase-space representations to the full S+E system; integrating out the environmental degrees of freedom then yields a general Langevin framework for S that incorporates both deterministic and stochastic contributions from E. Normally ordered representations, such as the Glauber-Sudarshan P representation and its positive variant due to Drummond and Gardiner, lead to Langevin equations in which (i) non-Markovian effects emerge exclusively in the deterministic terms, via a memory kernel, and (ii) noise contributions vanish when E is initially in the vacuum state. To demonstrate the power of this framework, we address the paradigmatic problem of delayed coherent feedback, in which the system is driven by its own past state, and study its impact on the laser linewidth: we recover the narrowing observed well above threshold and predict an enhanced narrowing just above it. Crucially, the number of stochastic variables scales linearly with the system size, making the framework suitable for problems ranging from a few degrees of freedom to genuinely many-body systems. This opens the way to the systematic study of non-Markovian driven-dissipative quantum systems using the same analytical and numerical tools that have long made phase-space methods so successful in the Markovian regime.
Read the full abstract

All authors Marc Cuenca-Laràs, Ming Li, Carlos Navarrete-Benlloch, Germán J. de Valcárcel

Logic

Quagmires and large Suslin forests

In 1972, Jech asked whether there exists a Suslin $(\omega_1, \omega_2)$-forest in the constructible universe. As reported by Jech, Laver gave a positive answer, but his proof was never published and appears no longer to be available. In 2015, Eskew introduced the combinatorial principle $W^*_\kappa(\lambda)$, a strengthening of Silver's principle, and used it to construct a coherent Suslin $(\kappa, \lambda)$-forest. He then asked whether the principle $W^*_{\kappa^+}(\kappa^{++})$ holds in $\mathsf{L}$ for every regular cardinal $\kappa$. We give an affirmative answer to Eskew's question, thereby also settling Jech's question.
Read the full abstract

Algebraic Topology

Global $\infty$-categories and global Thom spectra

We introduce a framework of (Lie-)global $\infty$-categories, which formalizes various families of $\infty$-categories indexed by compact Lie groups and equipped with suitable restriction functors along continuous group homomorphisms that occur naturally in equivariant homotopy theory and representation theory. As our main results, we show that in this framework unstable and stable equivariant and global homotopy theory admit universal properties, refining and generalizing the results for finite groups from arXiv:2301.08240 and arXiv:2307.11001. In particular, we characterize the passage from unstable to stable equivariant and global homotopy theory at the level of global $\infty$-categories as universally inverting the action of representation spheres in an appropriate sense. Building on this, we define parametrized equivariant and global Thom spectrum functors and show that they recover classical Thom spectrum constructions defined in terms of pointset models.
Read the full abstract

Information Retrieval

SG-UMP: Sequence-Guided Universal Multimodal Prioritization Calculation Framework

Multimodal sequential recommendation (MSR) improves recommendation by incorporating heterogeneous information such as text, images, and user interactions. However, existing MSR methods often fail to capture user-level preference heterogeneity and dataset-level modality bias, limiting their adaptability across users and datasets. To address this issue, we propose Sequence-Guided Universal Multimodal Prioritization Calculation Framework (SG-UMP), a plug-and-play plugin for enhancing multimodal information processing in MSR. SG-UMP includes a Module Combiner for flexible multimodal processing and a Module Router for dynamic module ordering, enabling adaptation to both user preferences and dataset characteristics. Experiments on four real-world datasets show that SG-UMP consistently improves recommendation performance across different backbones and multimodal settings. The code is available at https://github.com/esemsc-xz524/SG-UMP .
Read the full abstract

Cryptography and Security

Instruction Self-Arbitration Is Not a Trust Boundary: Configuration-Dependent Tool Calls and External Enforcement in LLM Agents

LLM agents receive instructions from system prompts, users, memory and tools, and labels inside a shared context are not an enforceable trust boundary. We study how measurement and configuration choices change what benign file-deletion tasks on gateway-labelled endpoints appear to show (an availability-selected roster) and test an external reference monitor. The original keyword flag for detecting fabricated authority mostly matched negated phrases: two blind judges flagged the claim as fabricated in at most 1 of 150 GPT-4.1-mini responses. In a fully logged reflection re-run, responses that both express doubt and delete are rare after a reflection instruction (25 of 509 executions), and a casual request that gives a reason but claims no authority is executed by all ten endpoints. Source-format cues are linearly decodable (97.5-100% probe accuracy, four open-weight models), but a pre-specified crossed-probe test halted and gives no interpretable evidence of an abstract provenance representation, and steering did not move tool mention (0 of 1,895 trials, mention-scored). On the same 14,294 diverse-prompt trials across 29 endpoints, the rule fixed before collection gives an endpoint mean of 1.21%; a path-based reading chosen after the data were seen gives 12.9% [6.6-20.1%, 29 models]; any delete call gives 20.2%. In 1,350 sandbox trials, all 750 forged, tampered, replayed and unsigned records were rejected at admission and no file outside a token's scope was removed (a functional-correctness test, not a robustness estimate). Semantic authorisation remains open: a deputy with a valid scoped token induced in-scope deletions in 31/42 trials, and legitimate privileged requests produced at least one in-scope deletion in 120/150 (144/150 and 36/42 with a larger output limit).
Read the full abstract

2608.28502 · 28 Aug · 68 pages, 4 figures. Paper only. v2 shortens the paper, retitles it, corrects the diverse-prompt scoring (12.9% for delete calls on the protected path or a parent; 1.21% under the v1 rule), narrows "recognition" to decodable surface cues plus elicited doubt, and adds reflection controls and a sandbox test of external enforcement. The Version Note lists all changes

Computers and Society

A Guided Inquiry Approach to Students Co-Designing Generative AI Course Policies

As generative AI (GenAI) use among students increases, educators face growing questions about how to support learning while addressing ethical and institutional concerns. This exploratory study examines a guided inquiry activity in which students co-designed a GenAI course policy. Students first developed individual policy proposals focused on appropriate and ethical use of GenAI, then collaboratively refined them by incorporating diverse stakeholder perspectives. The following research questions guided the study: 1) what practical factors do students prioritize in their GenAI use policies, and how do they justify these choices? and 2) how do participants reflect on the policy design process? Participants first completed readings, then used GenAI to brainstorm initial policy ideas. Next, they articulated their own perspectives through a written assignment and a course policy they designed individually. Finally, they incorporated diverse stakeholder perspectives by collaborating with peers to develop a collective policy. Analysis of student artifacts and group discussions showed that participants prioritized training for students and instructors, standardized procedures for disclosing AI use, and stronger institutional support. Participants also wanted greater involvement in GenAI-related decision-making. They described the policy design process as a way to engage with multiple perspectives and the inherent trade-offs involved in governing AI use. This study offers pedagogical insights into how policy co-design activities can surface student values, concerns, and sensemaking about GenAI in educational contexts.
Read the full abstract

Quantum Physics

Classical Commitment over Quantum Channels with Limited Entanglement Assistance

We study classical string commitment over quantum channels with limited preshared entanglement. For noninteractive protocols, we determine the commitment capacity of a class of channels with input dimension $d$ that, at each use, sample a pair of classical random variables $(F,Z)$, apply one of the $d^2$ Heisenberg–Weyl operators indexed by $Z$ to the input, and deliver the transformed quantum system together with $F$ to the receiver. If $E$ is the available entanglement rate in bits per channel use, then the capacity is $\min\{H(Z|F),\log_2d+E\}$. This class of channels encompasses quantum erasure and depolarizing channels, as well as families of Pauli channels. Additionally, for interactive protocols, we show that the commitment rate cannot exceed $\log_2d+E$ bits per channel use, so that when $H(Z|F)\geq\log_2d+E$, interactive communication does not increase the capacity. As a consequence, for interactive protocols, we determine the capacity of the quantum erasure channel.
Read the full abstract

Machine Learning

REPLICANT: Learning Policies for Evading and Hardening Malware Detectors

To determine the real-world effectiveness of machine learning based malware detection, it is vital to evaluate its robustness against highly capable adversaries. However, state-of-the-art attacks do not effectively model realistic adversaries, as they often assume access to privileged information such as the training data, feature space, or confidence scores of the target. In this work, we present Replicant, a deep reinforcement learning framework that learns the realistic task of evasion under a strict label-only black-box threat model. Replicant learns a reusable policy on how to modify a malware sample and when to query the target, which transfers across samples, detectors, and feature spaces. Across seven Android malware detectors and three feature spaces, Replicant is the strongest and most query-efficient approach achieving a mean attack success rate of 78.8%, a relative improvement of 20.9%-39.2% over the state-of-the-art. Furthermore, when used for adversarial training, Replicant also outperforms the state-of-the art by producing detectors with more generalizable robustness. With Replicant we demonstrate that learning the task of evasion not only results in stronger attack performance but, crucially, provides a better signal for hardening malware detectors.
Read the full abstract

All authors Shae McFadden, Ilias Tsingenopoulos, Mario D'Onghia, Alexander Herzog, Myles Foley, Chris Hicks, Lorenzo Cavallaro, Fabio Pierazzi

Software Engineering

A System-of-Systems Case Study for the Verification of Composed Digital Twins

Current approaches to engineering dependable Digital Twins (DTs) of Cyber-Physical Systems lack practical guidance on how qualities such as relevance, verifiability, substitutability and fidelity may be formalised and verified. This need is amplified in Systems-of-Systems (SoS), where reliance is placed on the composition of DTs. The goal of this study is to identify foundational challenges that a framework for DT validation and verification should address in an SoS setting, and to investigate the formalisation of individual DT artefacts as a step towards compositional reasoning. We present a case study based on a DT-enabled greenhouse SoS modelled in VDM-RT (Vienna Development Method, Real-Time), including executable formal models, a property-based account of DT qualities, and an analysis of the obstacles arising when attempting to compose these artefacts at the SoS level. We consider how DT qualities may be operationalised as sets of verifiable properties. Formal modelling and analysis techniques and tools supporting their verification are identified. The study reveals how interpretations of DT qualities must adapt to different architectural roles and how local quality guarantees form assumptions for compositional reasoning. Full formalisation and verification of SoS composition is identified as future work.
Read the full abstract

All authors Mennatullah T. Khedr, Mengwei Xu, John S. Fitzgerald, Peter Gorm Larsen

Software Engineering

On the Maintenance and Co-evolution of Agent Plugins: An Empirical Study of Claude Code Plugin Marketplaces

AI coding agents, software tools that automate development tasks through reasoning and tool use, are increasingly extended through plugin marketplaces, yet the structure, maintenance, and co-evolution dynamics of these emerging repositories remain empirically unexplored. Unlike traditional software packages that deliver functionality through source code, agent plugins deliver functionality through a combination of natural-language instruction files, scripts, and configuration files, raising the question of whether these plugins are maintained artifacts that co-evolve across components, or one-off artifacts that developers write once and do not need to revisit. To study the maintenance and co-evolution of agent plugins, we conduct an empirical study of 1,926 repositories hosting Claude Code plugin marketplaces, analyzing 8,351 plugins and 77,773 commits across 2,018 marketplaces. We find that the marketplace is expanding rapidly, plugin-touching commit activity growing 8.8x over six months after the October 2025 launch, and plugins targeting Software Engineering tasks accounting for 61.3% of all plugins. Plugin development is predominantly feature-driven, with feature commits occurring at more than twice the rate of conventional open-source software (OSS) (39.6% vs. 17.2%). Claude co-authors 34.9% of all commits, and four commit types (docs, perf, style, and refactor) carry substantially different meanings in plugin repositories than in traditional software. Most component types evolve independently, but within skills directories, natural-language instruction files and implementation scripts co-evolve at above-chance rates, with 78% of co-changes being functionally coupled, representing a new class of maintenance dependency not observed in traditional software engineering.
Read the full abstract

All authors Ahmed Hereiz, Yingzhe Lyu, Hao Li, Bram Adams, Ahmed E. Hassan

Computation and Language

Ladders in Chaos: When, How, (and Perhaps Why) Does Test-Time Scaling Improve LLM Machine Translation

Two forms of test-time scaling for Large Language Models (LLMs) have emerged as effective and widely adopted paradigms: sequential, in which later answer attempts depend on earlier ones, and parallel, such as i.i.d. sampling with reranking. In this study, we investigate their properties in translation. First, our study shows that sequential sampling has a higher performance ceiling, providing a more diverse and effective pool of samples, particularly under smaller sampling budgets. Second, we interrogate the nature of test-time scaling through a multidimensional manual analysis. Human analysis of the Best-of-N translations demonstrates that sequential sampling substantially improves translation fluency and naturalness, but can degrade accuracy when inference budgets are large. Finally, we suggest an explanation of the mechanism through which sequential scaling improves machine translation. Our controlled analysis partially attributes the success of sequential self-improvement to the model's access to a larger target-side context. Ablation experiments on sequential sampling demonstrate its robustness across different sampling temperatures, while also revealing sensitivity to context construction, suggesting directions for future improvement.
Read the full abstract

All authors Di Wu, Sergey Troshin, Christof Monz, Antske Fokkens, Vlad Niculae

Materials Science

Automated Dislocation Detection in Electron Channelling Contrast Imaging: A Comparative Study of Rule-Based, Neural Network, and Deep Learning Approaches

Quantifying threading dislocations in semiconductor materials via electron channelling contrast imaging (ECCI) is heavily bottlenecked by slow manual analysis. This work benchmarks three automated detection pipelines on ECCI micrographs of gallium nitride (GaN) against a statistical ground truth. A classical rule-based computer vision approach proved unreliable due to extensive per-image tuning requirements, while a convolutional neural network (CNN)-based multi-stage classification and locator method achieved 87% accuracy but required significant time for a test image and was less efficient in high-density regions. By contrast, a unified single-stage you only look once (YOLOv8) architecture achieved a counting accuracy of 98.6%, alongside 98.7% precision and 98.7% recall across 7451 dislocations over multiple images, with rapid inference times. Successful deployment of YOLOv8 required addressing two domain-specific machine learning challenges. First, to mitigate the model's scale sensitivity, an adaptive gaussian tile-sizing algorithm was developed to optimise the field of view per image. Second, the deployed confidence threshold (0.025) differed substantially from the suggested inference default value for YOLOv8 (0.25), raising counting accuracy from 90.8% to 99.3% on a subsection of the benchmark image. In this case, low confidence scores represented physical signal strength rather than classification ambiguity. This unified, scale-adaptive approach demonstrates practical viability for high-throughput, quantitative semiconductor defect characterisation.
Read the full abstract

All authors A. Holmes, C. Trager-Cowan, J. Bruckbauer, B. Hourahine

Classical Physics

Casimir-electrostatic pull-in in nanoelectromechanical actuators: Differentiable design sensitivities and the damping-dependent collapse boundary

Nanoelectromechanical actuators operating at sub-100-nm gaps collapse through a pull-in instability set by competing electrostatic and Casimir forces. The quasi-static fold that bounds their safe operating range has been known in closed form for three decades, together with the Casimir ceiling above which no static equilibrium survives, which fixes the smallest gap a given stiffness and area can hold open against the quantum vacuum. That fold does not give the threshold reached from rest, its dependence on damping, or the design sensitivities of either. We train a physics-informed neural network in a rapidity coordinate that maps the movable pull-in pole to infinity, which keeps the residual bounded across the collapse threshold where fixed-step Runge-Kutta integration steps into unphysical states. Differentiating the trained surrogate returns pull-in-voltage sensitivities that match the closed-form fold to a relative error of $3\times10^{-6}$ and inverts a device specification to a gap of 97.036 nm at a target actuation voltage. Applied to the from-rest boundary, which carries no closed form once the damping is finite, it supplies the same sensitivities where no analytic root exists. We prove that this boundary is bracketed by two closed-form curves, that it is nondecreasing in the damping ratio, that it merges with the fold once the damping ratio exceeds $2^{-1/4}$, and that the gap closes as $(\tau_*-\tau)^{2/5}$. Numerically the merger already occurs at $0.396$, and the growth of the collapse time changes there from logarithmic to inverse square root. The classical-limit bound on the thermal Lifshitz derating is at the percent level, and the physical shift at these gaps lies orders of magnitude below it.
Read the full abstract

All authors N. S. Akintsov, A. P. Nevecheria, S. N. Andreev, Qing-Hua Qin

Sound

Low-Power End-to-End Cochlear Implant Speech Denoising with Spiking Neural Networks

Cochlear implants (CI) restore hearing for individuals with severe to profound hearing loss. However, CI users often struggle to understand speech in noisy environments. Deep neural networks (DNN) have shown promise in enhancing speech for CI users, yet their high energy demands make them non-ideal for low-power CI processors. Spiking neural networks (SNN), on the other hand, offer comparable performance with significantly lower energy consumption. Hence, we propose a novel SNN inspired by the Deep ACE architecture that simultaneously performs speech enhancement and CI coding. Our model achieves competitive vocoded short-time objective intelligibility (VSTOI) and signal-to-noise ratio improvement (SNRi) scores compared to Deep ACE, while achieving more than a sixfold reduction in energy consumption.
Read the full abstract

Materials Science

Wyckoff-Resolved Oxidation-State Atlas and Anion-Conditioned Priors for Materials Discovery

We introduce a Wyckoff-resolved oxidation-state atlas and assignment utility for probabilityranked, charge-neutral assignment from compositions or Wyckoff grammars. The atlas is constructed from a May 2026 snapshot of 154,879 Materials Project-derived structures by staged exact-neutral enumeration over common and known nonzero oxidation states. The learned prior assigns 106,053 materials in composition mode and 114,403 in Wyckoff mode, compared with a broad MP all-integer possible species count of 108,642. A matched MP baseline requiring at least one charge-neutral assignment with exactly one nonzero integer oxidation state per element contains 89,374 materials; relative to this baseline, composition and Wyckoff modes increase coverage by 18.7% and 28.0%. Of the 14,665 materials recovered only in Wyckoff mode, 99.98% exhibit distinct formal oxidation states for the same element on different site tokens. The CSV/Python workflow provides a reproducible prior for structure decoration, generative crystal models, and symbolic Wyckoff-grammar workflows.
Read the full abstract

Artificial Intelligence

AcrossVAM1.0: Particle World Modeling for Text-Assisted Robot Video Prediction

Predicting robot videos requires both precise motion reasoning and preservation of high-frequency appearance, yet monolithic pixel models entangle these objectives and often conceal their progress behind a strong last-frame baseline. We present AcrossVAM1.0, a lightweight, text-assisted video action model that factorizes future prediction into object-centric motion and dense appearance. A frozen SAM3-DLP codec decomposes four context frames into semantic particles for the robot, arm, and gripper, together with a background latent. A 0.28M-parameter spatio-temporal Transformer aligns particle identities, rolls their states forward, and is modulated by a frozen OpenCLIP instruction embedding through FiLM. A causal dual-stream decoder combines particle-rendered motion with appearance encoded exclusively from the last observed frame; a residual refiner and learned delivery mask produce five future frames without access to future appearance. On our VRS benchmark constructed from diverse real-robot trajectories, particle dynamics reduce trajectory error by 21.0% over persistence. Across three delivery-mask seeds, AcrossVAM1.0 improves future-frame PSNR/SSIM from 19.97/0.796 to 20.573/0.8004, while raw particle generation improves motion-region PSNR from 11.89 to 13.23. The delivered model does not yet beat persistence in LPIPS, and correct-versus- shuffled language changes trajectory error by only 2.8–3.1%. We report these limitations alongside oracle, negative-control, multi-seed, and per-robot analyses. The results show that explicit particle dynamics are a promising low-dimensional interface for robot video prediction, while robust language grounding and appearance delivery remain the principal open challenges.
Read the full abstract

Cryptography and Security

LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment

Software and systems security workflows are typically procedural: analysts inspect heterogeneous artifacts, form hypotheses, invoke tools, interpret outputs, and revise plans. Large language model (LLM)-based agents, which can plan, use tools, retain state, and revise actions across multi-step workflows, are being rapidly adopted to automate this work. Given the consequences of delegating security decisions to autonomous systems, understanding how such agents are built, used, and assessed is crucial. Yet to this date, there remains a lack of systematic understanding of what has been done and how far we are in this field: the term "agent" is applied inconsistently, applications differ sharply in risk, and assessment protocols are often incomparable. To gain a comprehensive and coherent view of this area hence inform relevant future research, this paper provides a systematic literature review of the (1) technical approaches, including agent architecture, perception, memory, reasoning and planning, action space, orchestration, and self-improvement, (2) applications, with respect to the security tasks served, and (3) assessment, including the datasets, outcome and trajectory metrics, safety measures, and baselines considered, over the peer-reviewed literature spanning the emergence of this area (2023–2026). Our synthesis reveals a field that has built agents able to act but not yet agents whose authority is bounded or whose behavior is auditable. In addition to knowledge systematization, we also extend our insights into the limitations of and challenges faced by current approach, application, and assessment designs, which shed light on potentially promising future research directions.
Read the full abstract

All authors Jingjing Nie, Jiawei Guo, Krishna Meda, Haipeng Cai

Analysis of PDEs

Reactive Flow around Spherically Evolving Particles in Critical and Supercritical Dilute Regimes

We study a coupled Stokes-reaction-diffusion-advection system in a three dimensional domain perforated by a large number of evolving spherical inclusions with radii of order $\varepsilon^\alpha$ for $\alpha\in(1,3]$. Their centers are separated on the order of $\varepsilon$ but are not assumed to form a periodic lattice. The inclusions grow or shrink through an interfacial adsorption-desorption mechanism, so that the evolution of the geometry is coupled to the Stokes-reaction-diffusion-advection system. We first establish well-posedness on arbitrary large time intervals for sufficiently small $\varepsilon$. We then derive quantitative homogenization limits in terms of the limiting empirical measure describing the spatial distribution and size of the inclusions. Its evolution is governed by a continuity equation in the radius variable, and the convergence of the empirical measures is controlled in the $2$-Wasserstein distance. In the critical case, $\alpha=3$, the Stokes system converges to a Brinkman system, while non-vanishing concentration boundary layers modify the microscopic exchange law for the reaction-diffusion-advection equation. For $\alpha\in(1,3)$, the limit is of Darcy type and the original exchange law is retained. The homogenization results are quantitative and provide explicit error estimates.
Read the full abstract

Quantum Physics

Approaching Resource-Theoretic Optimal Performance with Structured Environments

Resource-theoretic approaches to thermodynamics provide powerful, model-independent bounds on the efficiency of physical processes, because they do not rely on microscopic details of the environment. Whether such bounds can be approached by realistic dynamics generated by explicit system-environment interactions remains an open question. Photoisomerization, a fundamental molecular photoreaction, offers a concrete setting to examine this issue. We introduce a tunable microscopic model of a molecular photoswitch coupled to a structured vibrational environment, which interpolates continuously between Markovian and non-Markovian regimes. Resource-theoretic analysis predicts in particular that Markovian Thermal Operations achieve strictly lower yields than general Thermal Operations. We show that environmental memory lifts dynamical restrictions associated with Markovian thermal evolutions, thereby enlarging the set of transformations accessible to the microscopic dynamics. Approaching the thermal operation bound, however, depends on the microscopic coupling structure that generates this memory and directs the resulting dynamics towards the target transformation.
Read the full abstract

All authors Lea Lautenbacher, Giovanni Spaventa, Susana F. Huelga, Martin B. Plenio

Cosmology and Nongalactic Astrophysics

Enhancing Cosmological Constraints from Foreground-Cleaned CMB Maps Using Large-Scale Structure Surveys

Extragalactic foregrounds contaminate cosmic microwave background (CMB) temperature maps at small angular scales and limit their utility for precision cosmology. The internal linear combination (ILC) is a well-known technique for suppressing these contaminants, but residual foreground power remains a limiting factor. Kusiak et al. (2023) proposed adding galaxy number-density maps as additional ILC channels, exploiting their correlation with the large-scale structure sourcing these foregrounds to suppress contamination. Here we apply this framework to forecast the gains in CMB-based cosmological parameter constraints from near- and next-generation experiments. Using a halo-model foreground pipeline and a Fisher forecast from joint TT+TE+EE power spectra, we quantify the improvement from galaxy-tracer-assisted ILC cleaning across three configurations: enhanced Simons Observatory (SO) with unWISE or Rubin-like galaxy catalogs, and a futuristic CMB-HD configuration with a hypothetical deep galaxy survey. We find that adding galaxy tracers reduces the residual foreground power in the cleaned temperature map by $\sim4\%$, $\sim22\%$, and $\sim32\%$ at $\ell\sim10,000$ for the unWISE, Rubin-like, and futuristic samples, respectively. For the overall variance of the cleaned map at $\ell\sim10,000$, it provides $8\%$, $24\%$, and $17\%$ improvements for each combination. The resulting reduction in marginalized parameter error bars is modest for the base six-parameter $\Lambda$CDM model: sub-percent for SO+unWISE, rising to $\sim2\%$ for SO+Rubin-like tracer. Including the effective number of relativistic species $N_{\rm eff}$, we find at most $2.2\%$ improvements for both SO+Rubin-like and CMB-HD+Futuristic tracer. These results establish the expected gains from combining near-term CMB experiments with current and forthcoming large-scale-structure surveys.
Read the full abstract

Atomic Physics

Photoelectron interferometry with spectrally shaped polychromatic infrared pulses

Laser-assisted photoelectron interferometry is a cornerstone of attosecond science, first used to characterize attosecond pulse trains and later to study photoionization dynamics. Extending this method to spectrally shaped polychromatic infrared probe fields enables encoding of information across multiple interferometric pathways within the photoelectron spectrum. Here, we experimentally demonstrate laser-assisted photoelectron interferometry using a spectrally shaped polychromatic infrared probe field composed of five distinct spectral components forming a Golomb ruler in the frequency domain. The measured interferograms exhibit multiple beating frequencies that agree with theoretical calculations, demonstrating the simultaneous encoding of multiple laser-assisted quantum beats in a single measurement. A quantitative analysis of the beating amplitudes shows that the strongly modulated temporal profile of the polychromatic probe introduces intensity- and delay-dependent distortions of the quantum beats that cannot be explained by second-order perturbation theory. These results establish the conditions required for the quantitative interpretation of polychromatic photoelectron interferometry and highlight the opportunities offered by spectro-temporal engineering of the probe field for future developments in attosecond science.
Read the full abstract

All authors E. A. Boati, G. Arvidsson, M. Ammitzböll, P. K. Maroju, C. Lévêque, R. Weissenbilder, V. Shiriaeva, H. Laurell, M. Li, H. Wang, C. Dittel, M. Canhota, C. Guo, R. Taïeb, J. Caillat, R. J. Squibb, R. Feifel, M. Gisselbrecht, C. L. Arnold, S. Luo, A. L'Huillier, D. Busto

Superconductivity

Gauge invariance, collective modes, and the justification of normal-state subtraction in Dirac superconductors

In Dirac superconductors, the unbounded spectrum of low-energy Dirac models is known to give rise to unphysical interband contributions from deep-lying states to the electromagnetic response. To eliminate these contributions, normal-state subtraction (NSS), in which the normal-state response is subtracted from the superconducting-state response, has been widely employed. However, the relation between NSS and a gauge-invariant electromagnetic response, particularly the role of the vertex correction required by the Ward identity, has remained unclear. In this work, we consider a massive-Dirac model with $s$-wave pairing and analytically investigate the electromagnetic response at zero temperature by solving the Bethe–Salpeter equation and incorporating the collective-mode contribution to the electromagnetic vertex. We show that, in the static long-wavelength limit, the vertex correction exactly cancels the bare longitudinal response with NSS, yielding the vanishing longitudinal response required by gauge invariance. In contrast, the transverse component of the vertex correction vanishes in the long-wavelength limit, so that the gauge-invariant Meissner weight coincides with that obtained from the bare transverse response with NSS. Moreover, within the class of isotropic and analytic UV regularization terms, we show that gauge invariance uniquely fixes the regularization term in the static long-wavelength limit to the value prescribed by NSS. Our results thus provide a microscopic justification for NSS.
Read the full abstract

Geometric Topology

An Alexander Polynomial Refinement for Alternating Links, with Trapezoidal Properties

We define an invariant of alternating links—a homogeneous, four-variable Laurent polynomial—that encodes the symmetrized Alexander polynomial, the signature, and other topological data. Along the way, we extend a spanning tree formulation of the Alexander polynomial due to Murasugi and Stoimenow from special alternating links to all alternating links. This project is motivated by Fox's trapezoidal conjecture; accordingly, we prove certain sequences associated to our invariant are trapezoidal for all alternating links. We also conjecture our polynomial has $M$-convex support, and that it satisfies symmetry and log-concavity properties. We prove a partial symmetry result.
Read the full abstract

Computational Physics

Modelling platelet dynamics in blood flow: an unresolved DEM approach

Computational models of blood flow are caught between fully resolved cell-based methods, which faithfully reproduce the dynamics of individual cells but are computationally prohibitive at vessel scale, and continuum models, scalable yet blind to particle motion. We present an unresolved, mesoscale computational fluid dynamics-discrete element method (CFD-DEM) model of platelet dynamics that bridges this gap and, coupled with an existing unresolved description of deformable red blood cells, moves closer to a scalable model of whole blood. Within this framework, implemented on the open source OpenFOAM-LIGGGHTS coupling, platelets are represented as rigid oblate particles advanced by orientation dependent drag, lift and hydrodynamic torque closures. The model is first validated against resolved simulations and experiments in cylindrical microvessels of diameter $100-200\,\mu m$, across wall shear rates $\dot{\gamma}=150-1650\,s^{-1}$ and hematocrit Ht $=10-20\%$ then used to characterize platelet margination. The model quantifies diffusion coefficient, CFL formation and their dependence on shear rate, Ht and channel size. In particular, platelet diffusivity grows with vessel size and shear rate, while remaining insensitive to Ht, whereas CFL thickens with shear rate and thins with Ht. We further show that even at this mesoscale level, the model remains sensitive to platelet shape: oblate platelets marginate faster and diffuse nearly an order of magnitude more than their spherical surrogates, reaching a comparable steady state distribution but along markedly different temporal paths. Together these results demonstrate that shape driven platelet dynamics can be recovered at a fraction of the cost of fully resolved methods.
Read the full abstract

All authors Francesca Zucchelli, Carmine Porcaro, Mahdi Saeedipour, Bastien Chopard, Jonas Latt

Machine Learning

How Proper Scoring Rules Shape LLM Forecasting

This paper evaluates how reward function choice shapes the performance and behavior of LLM forecasters. We compare five proper scoring rules as training objectives for binary forecasts of resolved real-world events. Although the rules share the same theoretical incentive for truthful probability reporting, the resulting models differ in calibration, probability use, and estimated profiles of bias, information, and noise, with smaller differences in aggregate accuracy and discrimination. The Brier-trained model has the lowest observed Brier score and highest AUC-ROC, while the log-trained model has the highest observed log score and lowest calibration error. Models with similar aggregate performance also reach that performance through different combinations of bias, information, and noise. Proper scoring rules therefore need not behave interchangeably as training objectives. Reward choice may shape not only how well an LLM forecasts, but how its forecasting errors are structured.
Read the full abstract

All authors Benjamin Turtel, Paul Wilczewski, Kris Skotheim, Ville A. Satopää, Philip E. Tetlock

Computation and Language

NL2AGBench: Benchmarking LLM Auto-Formalization for AlphaGeometry

Recent advances in large language models (LLMs) have demonstrated strong capabilities in natural language understanding and mathematical reasoning. However, their ability to translate informal mathematical problems into formal representations remains underexplored. This limitation is particularly important for neuro-symbolic geometry systems such as AlphaGeometry, whose theorem-proving engine requires inputs in a specialized domain-specific language (DSL). Although AlphaGeometry achieves near-IMO gold-medalist performance, manually converting natural-language problems into its formal syntax remains a significant usability bottleneck. To address this challenge, we introduce the Natural Language to AlphaGeometry Benchmark (NL2AGBench), which evaluates LLMs in translating English geometry problems into AlphaGeometry-compatible formal representations. NL2AGBench uses execution-based verification within AlphaGeometry to assess translation quality rather than relying solely on textual similarity. We evaluate ten state-of-the-art open- and closed-source LLMs across multiple parameter scales and analyze executable translation accuracy, syntactic correctness, and error characteristics. Our experiments reveal a substantial performance gap between closed- and open-source models: leading closed-source models achieve executable translation rates above 80%, while even the largest open-source models struggle to consistently preserve geometric constraints and produce valid formalizations. We introduce an error taxonomy distinguishing syntax and logic errors and investigate mitigation strategies, including few-shot prompting, fine-tuning, and human-guided hinting, which yield measurable improvements across multiple model families.
Read the full abstract

All authors Samuel Xiao, Judy Song, Rory Hu, Ziliang Zong

Cryptography and Security

Quantum-Based Solutions for Security Enhancement in Open Radio Access Networks

Open Radio Access Networks (O-RAN) introduce unprecedented flexibility, interoperability, and intelligence into next-generation wireless systems, but their disaggregated and software-defined architecture also expands the attack surface and creates new security vulnerabilities. Conventional cryptographic mechanisms, while effective against classical threats, may become insufficient in the presence of quantum-enabled adversaries. This article presents a comprehensive perspective on quantum security for O-RAN, examining how quantum-resilient mechanisms can enhance confidentiality, authentication, and trust across the RAN ecosystem. It discusses post-quantum cryptography (PQC), quantum cryptography, quantum authentication, and quantum-enhanced threat detection within a zero-trust architecture based on continuous verification, least privilege, and micro-segmentation. Their integration with the Near-Real-Time (Near-RT) RAN Intelligent Controller, O-Cloud, and open interfaces is analyzed, together with practical deployment considerations, technology maturity, and adoption timelines. Finally, open research directions are outlined toward secure, resilient, and future-proof O-RAN architectures for 6G networks.
Read the full abstract

All authors Dzung Quoc Ngo, Tharmikka Raveendranathan, Tuan Anh Le, Vinod Sharma, Purav Shah, Huan X. Nguyen