Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Wednesday 9 September

Computation and Language

LLM-Anchored Paralinguistic Enrichment for Alzheimer's Disease Detection

Speech-based automatic detection of Alzheimer's disease (AD) provides a non-invasive and scalable approach to early cognitive screening. AD affects both lexical-semantic organization and speech production, including atypical pauses and word elongations. However, existing methods have yet to fully integrate these paralinguistic cues with linguistic content. We propose LLM-Anchored Paralinguistic Enrichment (LAPE), which enriches LLM-derived linguistic representations with paralinguistic cues through three coordinated innovations. The first is prosodic event textualization, which enables the LLM to model pauses and elongations jointly with lexical content by encoding them as explicit markers with bounded duration-aware repetition. The second is lexico-prosodic unitization and chunking, which preserves event identity and magnitude in both modalities by pooling only consecutive word units. The third is text-anchored paralinguistic fusion, which integrates local and utterance-level speech features by using NormGate to normalize and dynamically scale them relative to text. We evaluate LAPE on ADReSS and ADReSSo using participant-level cross-validation and leave-one-subject-out evaluation. LAPE achieves state-of-the-art performance across all four primary settings. Code will be released upon acceptance.
Read the full abstract

All authors Xiao Wei, Yuqin Lin, Yaru Cao, Jinyu Li, Bin Wen, Kai Li, Yueying Chen, Longbiao Wang, Jianwu Dang

Robotics

ReactHuman: A Physics-Grounded Benchmark for Human-Like Reactive Decision-Making in Embodied Multimodal LLMs

Reacting to sudden physical hazards (catching a slipping plate, dodging a falling knife) is both a meaningful test of embodied intelligence and a hard requirement for deploying multimodal large language models (MLLMs) as the decision coreof household robots. Existing evaluations, however, probe intuitive physics passively through question answering over videos, or target deliberate, long-horizon tasks such as navigation and rearrangement; none measure whether a model can turn physical understanding into immediate, safety-critical action. We introduce ReactHuman, the first physics-grounded benchmark for human-like reactive decision-making, in which the evaluated MLLM acts as the brain of a simulated humanoid facing sudden household hazards; it spans 17 event families and over 1,000 bit-for-bit reproducible scenes with exact, annotation-free ground truth derived from 240 Hz rigid-body simulation, including adversarial objects whose appearance contradicts their physics (a foam anvil, a steel apple). We further design a five-metric suite that scores each reaction along three axes: reasonable, safe, and physically grounded. We physically execute every committed plan so that decisions have observable consequences. With this harness we evaluate seven representative MLLMs. Results show that reactive safety is far from solved: models mishandle roughly one hazard in three, act from fixed dispositions rather than the observed scene, trust appearance over motion, and miss interception points at meter scale even when the chosen action is correct; none of these failures shrink with model scale. ReactHuman thus offers both a fine-grained diagnosis and a scalable training signal toward physically grounded, safety-aware embodied agents. The benchmark can be found here: https://huggingface.co/datasets/Alan123/reacthuman-benchmark-scaled
Read the full abstract

All authors Yizhan Li, Jianxin You, Mengyang Xiong, Yinhuan Chen, Zicheng Zhao, Dekun Wu, Dongqing Zhang, Bang Liu

Probability

A dynamic point of view on universality for random matrices over finite local rings

We consider the cokernel corners process for an i.i.d. matrix with entries in a finite local ring. When the distribution of the entries is uniform, this process is a Markov chain, and hence the ergodic theorem for Markov chains can be applied. This implies, in particular, that for uniformly distributed p-adic random matrices, the cokernels of the corners are distributed according to the Cohen-Lenstra measure, almost surely. The purpose of this note is to show that the conclusion of the ergodic theorem also holds for i.i.d matrices, provided that the distribution of the entries is not concentrated on the translate of a subring, or the translate of an ideal. This will follow from the bounds proved in a previous paper of the author.
Read the full abstract

Computation and Language

Does Linguistic Structure Enrichment Enhance Coherence Assessment? Not With Current Architectures

Recent advances in large language models have transformed human-computer interaction. Despite their fluency, these models often produce texts that are grammatically correct but semantically incoherent, containing contradictions or disruptions in logical flow. This work investigates whether enriching text with syntactic and rhetorical information can improve incoherence prediction. Our experiments and analysis show that plain texts achieved higher accuracy because the added information was structurally and syntactically incompatible with the language model's architecture. Additionally, to demonstrate the practical importance of coherence assessment, we performed zero-shot experiments on a Brazilian disinformation dataset, suggesting that textual coherence can serve as a proxy for detecting misleading content. Code and models are available at https://github.com/ittozzamV/cohereclassifier.
Read the full abstract

All authors Victor Mazzotti, Luiz Pereira, Marina Bitencourt dos Santos, Helena Maia, Carlos Caetano, Nádia Felix, Sandra Avila

Cryptography and Security

DriftNet: A Dual-Head Trajectory Transformer for Detecting and Localizing Prompt Injection in LLM Agents

When an indirect prompt injection succeeds against an LLM agent, the compromise is visible in the agent's own behavior: a benign prefix of tool calls, a poisoned observation, and a suffix of actions that serve the attacker. An operator needs three facts: where the attack entered, which steps it corrupted, and whether apparent poison was resisted. Existing systems return either a whole-trace verdict or a single unsafe index. We present DriftNet, a dual-head trajectory Transformer that reads a logged tool-call trajectory and answers all three questions in one forward pass: one head classifies the trajectory as compromised or not, and a second assigns every step one of four labels (benign, injection point, hijacked, failed injection). To our knowledge it is the first supervised detector to produce this joint output. A frozen sentence encoder and four identity-free world features embed each step; the trained trunk, under two million parameters and optimized with a class-weighted joint objective over both heads, needs no access to the agent's model. On the task-disjoint split of the AgentDrift benchmark (12,536 trajectories, 71,024 labeled steps), with a 20-configuration sweep bounding hyperparameter sensitivity to 0.011 F1 and the test part evaluated exactly once, DriftNet reaches trajectory-level F1 of 0.983, exact injection-point recovery on 98.7% of attacked trajectories, hijacked-span IoU of 0.979, zero flags on 218 resisted attacks, and 2.9% flags on hard negatives. A surface baseline retrained on the identical split recovers 11.1% of partial hijacks and 17.1% of delayed executions; DriftNet reaches 98.6% and 93.2% while lowering every false-alarm rate. Reading all 26 residual errors shows that most misses trace to trajectories whose labeled injection observation carries no legible instruction, and we report the benchmark's measured world-identity regularity alongside the results.
Read the full abstract

Instrumentation and Detectors

Mitigating Sensor Cracking using the Interposer Strategy in the ATLAS ITk Strip Detector

This paper introduces the strategy of using an interposing layer to mitigate thermal-stress-induced fracturing of silicon sensors in ATLAS ITk Strip modules. This strategy involves a change in the module design in which layers of soft adhesive and a kapton separator, known as an interposer, are added to the module stack-up to decouple thermal stresses. This paper discusses the interposer design and its implementation in the ATLAS ITk Strip barrel, including material tests to validate the additional components, the development of new processes to integrate the new layers into the module assembly workflow, results from design validation and quality control testing of prototype interposed modules, and studies indicating that the new design significantly reduces the thermal stress experienced by the sensor. Tests of prototype support structures loaded with interposed modules have shown no sensor fractures in the relevant temperature range for operation in ATLAS, resulting in the adoption of the interposer strategy as the primary sensor cracking mitigation in the ATLAS ITk Strip barrel.
Read the full abstract

Quantitative Methods

Discovering Subtypes of Neurodegenerative Progression with a Scalable Connectome-Constrained Dynamic Model

Parkinson's disease is clinically and biologically heterogeneous, yet its spatiotemporal progression remains poorly characterized. We present a connectome-constrained disease progression model that jointly estimates subject-specific disease time and data-driven subtypes from longitudinal morphometry. Applied to 85 imaging and clinical biomarkers from the Parkinson's Progressive Markers Initiative (PPMI) cohort, the model recovers four morphologically distinct progression subtypes. We validate the model on a hold-out cross-sectional dataset and benchmark it against SuStaIn under a matched training and validation protocol. Only our method recovers subtypes that correspond significantly to clinical motor subtypes and genetic variants of Parkinson's Disease.
Read the full abstract

All authors Daniel Semchin, Emile d'Angremont, Hao Ding, Alan Antar, Marco Lorenzi, Konstantinos Arfanakis, Ysbrand van der Werf, Paul Thompson, Boris Gutman

Statistics Theory

Online Change-Point Monitoring for Object-valued Time Series

We develop closed- and open-end procedures for monitoring changes in the marginal distribution of object-valued time series. The method combines two distance-based Hilbert-space embeddings, a monitoring-time-dependent projection, and self-normalization. It is computable entirely from pairwise distances, does not require long-run variance estimation, and admits exact recursive updates. Under weak temporal dependence, the closed-end null limit is a pivotal Brownian functional. For monitoring over an unbounded horizon, we introduce a growing monitoring-time weight, establish a maximal inequality and uniform remote-tail control, and derive an open-end pivotal limit together with an equivalent fixed-interval representation for critical-value simulation. Both procedures are consistent against fixed marginal changes and retain the $n^{-1/2}$ and $n^{-1/4}$ local detection boundaries associated with the linear and quadratic projection signals. Simulations with distribution-valued time series illustrate favorable numerical properties, and an application to monthly stock-return distributions illustrates the usefulness of our procedure. Further simulations with graph-valued time series and an application to spatial point processes of seismic activity are reported in the supplementary material.
Read the full abstract

Optimization and Control

The Polarity Process for the Maximum-Volume Inscribed Ellipsoid Problem

We study the maximum-volume inscribed ellipsoid (MaxIE) problem for a polytope through a geometric iteration based on polarity. Given an interior point, the method forms the shifted polar polytope, computes its minimum-volume covering ellipsoid (MinCE), and then polarizes this covering ellipsoid back to obtain a new inscribed ellipsoid. This procedure was suggested by Khachiyan and Todd in 1993. Prior work studied basic polarity identities and gave an asymptotic-convergence argument for the exact process, but did not furnish a volume-ratio rate. Independently of that asymptotic argument, we develop a new analysis based on the log-volume of the polar MinCE as a potential function. We prove its convexity with an explicit gradient formula and establish global linear contraction of the potential gap along each trajectory. The contraction factor is existential and instance-dependent. This gives both a separate convergence proof and a finite volume-ratio iteration bound. We further analyze an inexact polarity process in which each MinCE subproblem is solved only approximately, and derive sufficient oracle tolerances for producing a prescribed volume approximation. Combining this outer analysis with existing algorithms for MinCE gives conditional arithmetic estimates based on the path-following Newton method, the barycentric coordinate descent, and the away-step Frank-Wolfe method. Numerical experiments compare these oracle choices with two MaxIE baselines and illustrate that performance depends on matching the MinCE solver to the instance geometry.
Read the full abstract

Game Theory

STV Audit Graphs: A Visual Tool to Measure Election Stability

The Single Transferable Vote (STV) is an algorithmic election rule. Round by round, a profile of ranked-choice ballots is reinterpreted to determine which decision to make next, and candidates are seated or eliminated until a full winner set emerges. This algorithmic nature makes STV theoretically more brittle than other election rules: uncertainty about an early round of the election might percolate irreversibly into the rest. For this reason, a generalized non-trivial Risk-Limiting Audit (RLA) framework has remained elusive for STV. Such a framework must concoct a set of null hypotheses, or assertions, whose rejection would bound the probability that the outcome of the election was incorrectly reported. We present audit graphs as a solution to design the assertions needed for RLAs of arbitrary STV elections, as well as quantitatively describe the uncertainty (or lack thereof) of their outcomes. These audit graphs explore election paths that are “close” to the recorded one by considering the alternative decisions the STV algorithm might have made if a small number of ballots were perturbed.
Read the full abstract

Machine Learning

Relatively Smart II: Tractable or Semi-Supervised Instance-Optimal Learning

We continue the study of relatively smart learning, introduced by Dughmi and Pour (2026), which asks a supervised learner to compete, marginal by marginal, with every distribution-fixed error guarantee soundly certifiable from unlabeled data. They showed that the One-Inclusion Graph (OIG) learner is relatively smart with a quadratic sample-complexity blowup, and that no relatively smart learner can do better, leaving open whether ERM or another natural or tractable learner achieves comparable guarantees. They also left open whether the blowup can be restricted to unlabeled data. Our firs results shows that ERM—and in fact any proper consistent learner—is relatively smart for binary classification in the distribution-free setting. We show that a small certifiable error with $m$ samples implies a similarly small error on the uniform distribution over a random sample of size $O(m^2)$, yielding a cover of size at most $2^{m+1}$ on that sample. This suffices to control the error of proper consistent learners with $O(m^2)$ samples. We then show that semi-supervised relatively smart learning is information-theoretically possible with a quadratic blowup only in unlabeled sample complexity and no blowup in labeled sample complexity. The learner uses a natural generalization of OIG to a leave-most-out transductive problem, where labels of part of a finite pool are revealed and the remaining labels are predicted. Finally, this label efficiency comes at a cost in simplicity and tractability. If the hypothesis class is accessed only through an agnostic ERM oracle, any semi-supervised relatively smart learner with substantially sub-quadratic labeled-sample blowup requires super-polynomially many oracle calls. This holds even when the marginal is given explicitly, and thus also yields an intractability result for distribution-fixed learning that may be of independent interest.
Read the full abstract

Distributed Computing

Navigating Small-World Networks with Distance Predictions

The small-world phenomenon was given an algorithmic foundation by Kleinberg, who showed that in an augmented $k$-dimensional lattice a decentralized greedy algorithm delivers a message in $O(\log^2 n)$ expected steps. We study predicted-greedy routing, in which a mobile agent forwarding the message moves at each step to the neighbor minimizing a noisy $(\varepsilon,\delta)$-prediction of its distance to the target, redrawn at every step from an oracle conditioned on the full routing history. Two cases arise from what this agent can observe. An agent with the coordinate awareness can still compute lattice distance exactly, but not graph distance in the shortcut-augmented network, since that depends on the shortcuts of nodes it has not yet visited; given an $(\varepsilon,\delta)$-prediction of graph distance, information the classical model never supplies, it achieves expected delivery time $O(\log n/(1-4k\varepsilon\delta))$, an asymptotic improvement over $\Theta(\log^2 n)$. An agent with no coordinate awareness at all, the natural model for a privacy-preserving network whose nodes never disclose their coordinates, cannot compute even lattice distance; given an $(\varepsilon,\delta)$-prediction of lattice distance instead, it still reaches the target in $O(n/(1-4k\varepsilon\delta))$ expected steps. Together these results show that a modest amount of predicted information, of the right kind, is enough to accelerate decentralized routing well below Kleinberg's classical bound, and that even when nodes reveal no coordinates at all, reliable delivery remains achievable.
Read the full abstract

Combinatorics

Connected irregular cospectral graphs with distinct Lovász numbers

For every integer $n \geq 11$, we construct connected, irregular, nonisomorphic graphs on $n$ vertices with equal adjacency, Laplacian, signless Laplacian, normalized Laplacian, and Seidel spectra and equal independence, clique, chromatic, complement chromatic, and maximum-cut numbers, but distinct Lovász $\vartheta$-numbers. We join each graph in a fixed cospectral, nonisomorphic regular pair on ten vertices, due to van Dam and Haemers (2003), with $K_{n-10}$. We prove that these joins preserve equality of the five spectra and stated integer-valued invariants, and leave the respective Lovász numbers unchanged. We derive exact expressions for these numbers and prove that they are distinct. We also determine the cardinality-constrained maximum-cut profiles of both base graphs and their complements, yielding exact formulas and equality of maximum-cut numbers for the joins of either the base graphs or their complements with $K_{n-10}$. For $n=10$, we present a regular pair satisfying all stated properties except irregularity, followed by a connected, irregular, nonisomorphic pair satisfying all properties. An exhaustive computer search finds no connected, irregular, nonisomorphic pair on at most nine vertices sharing the first four spectra and all stated integer-valued invariants. Thus, ten is the smallest possible order, and such pairs exist for every $n \geq 10$. This extends and strengthens Sason's (2024) result for even $n \geq 14$, which used the neighbors splitting join operation and did not address Seidel matrices, complement chromatic numbers, or maximum-cut numbers of the graphs or joins of their complements. Thus, the Lovász number efficiently certifies nonisomorphism even when all five spectra and stated integer-valued invariants coincide.
Read the full abstract

Machine Learning

Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble

Language models are trained to implement a helpful AI Assistant character (e.g., Claude). We explore how finetuning on synthetic stories affects this character. Does it change the Assistant's behavior in multi-turn conversations with users, a format quite different from the stories? And does the Assistant adopt the behaviors and preferences of human characters? We refer to this adoption as story imprinting. We finetune GPT-4.1 and Kimi-K2.6 on stories in which generally helpful human characters give subtly harmful advice after being insulted. The Assistant adopts the same conditional behavior while otherwise remaining helpful. This occurs even when fewer than 2% of stories depict the behavior. In a separate experiment, the Assistant adopts preferences that are only implicit in the narration. A human character's body language suggests they dislike working on spreadsheets, yet they never say so and continue giving good advice on spreadsheets. After finetuning, the Assistant becomes less likely to choose spreadsheet tasks. Next we ask which characters most influence the Assistant. We find the Assistant adopts behaviors more often from characters that resemble it (e.g., helpful rather than dismissive). We call this the affinity effect. The effect extends to other personas elicited with system prompts: unhelpful personas adopt behaviors from unhelpful characters. We also observe it in finetuned base models. We use the affinity effect to learn how models represent the Assistant. We find the Assistant adopts behaviors more from characters affiliated with elite universities (e.g., Yale) than non-elite ones. This implies the model's internal representation of the Assistant is more similar to humans from elite universities. Overall, the Assistant can be influenced by stories that depict only human characters (no AIs), which may conflict with the Persona Selection Model for the Assistant.
Read the full abstract

All authors Jorio Cocola, Lev McKinney, Harry Mayne, Jan Betley, Owain Evans

Emerging Technologies

Fractional-order hardware for neuromorphic computing: Is the order really the problem?

Does a neuromorphic system need a true power-law memory kernel, and if so, can anyone build one? Neuromorphic systems process signals spanning many timescales at once, from milliseconds to tens of seconds. Integer-order circuits buy each additional timescale with an additional state variable. Fractional-order dynamics offer a different bargain: one operator whose power-law kernel carries a continuum of timescales, tuned by one parameter, the order alpha. A fractional derivative is non-local, so evaluating it costs storage and arithmetic that grow with the retained history, where an integer-order derivative costs a constant. This review organizes the hardware literature around that cost. We derive the retained history needed to hold the truncation error below a tolerance epsilon, show that it scales as epsilon^(-1/alpha), and set beside it a second and independent limit on the direct form: in fixed point the weights themselves underflow, so word length caps the usable history however long the buffer is. The two limits move at very different rates with the order, and where they cross decides whether a word length can serve an order at all. We use both to sort published hardware into three strategies, note a fourth the numerical literature has developed and this hardware has not, and survey digital, analog and device work. Along the way we ask whether the field is worried about the right obstacle. It is not. Fabricated constant-phase devices already span the orders two groups identify as task-optimal, so the order gap has largely closed, leaving a residual gap near 0.1 and at the lower order describing cortical adaptation. What remains is a frequency-band gap of about three decades at the low end. That corner is not empty, since double-layer electrodes work there, but every device in it is discrete, and no integrable thin-film element has been characterized there.
Read the full abstract

Cryptography and Security

AspisAI: A Canonical, Machine-Interpretable Governance Framework for Automated Multi-Standard Compliance Monitoring

Organisations operating in regulated and critical-infrastructure sectors must satisfy multiple, heterogeneous cybersecurity and privacy instruments simultaneously, including but not limited to ISO/IEC~27001, the NIST Cybersecurity Framework~2.0, Cyber Essentials, and the GDPR. In practice, these obligations are managed through manual mappings, spreadsheet-based tracking, and periodic audits that are costly to maintain, inconsistent across standards, and weak in traceability. This paper presents AspisAI, a bounded, standard-agnostic governance framework that translates selected requirements from several frameworks into a canonical, machine-interpretable control model, and evaluates submitted evidence against condition-based decision rules to produce explainable, traceable compliance determinations. Within a bounded scope of 26 representative requirements, the framework is evaluated in a controlled simulation against five governance-oriented criteria and, critically, against two external reference points that mitigate the circularity of single-author evaluation: its cross-standard mappings are validated against NIST's own published informative references, with 57\,% exact agreement and divergences confined to same-family controls, and the framework is applied to real third-party evidence from the OpenSSF Scorecard, surfacing genuine governance gaps in a live open-source project. The controlled results, comprising full requirement encoding, 88.5\,% mapping coverage, complete traceability, and correct detection of all introduced gaps, establish functional correctness, while the external validation provides evidence of applicability beyond the simulation. The contribution is therefore a demonstration that a canonical, provenance-preserving governance model can render multi-standard compliance both automatable and auditable.
Read the full abstract

All authors Tsafac Nkombong Regine Cyrille, Hasan Dag, Reiner Creutzburg, Knut Haufe

General Relativity and Quantum Cosmology

WKB approximation for quasi-bound states and trapped modes

We develop a largely automatic semi-analytic method for calculating weakly damped quasi-bound states and trapped modes supported by a local minimum of an effective potential. The real part of the frequency is obtained from an arbitrarily high-order local WKB expansion and Padé resummation, while the exponentially small imaginary part is estimated using the Gamow approximation for tunnelling through one or two potential barriers. Because the local quantization requires only derivatives of the potential at its minimum, the method applies readily to non-rational effective potentials and to different compact-object geometries. We provide a Mathematica notebook implementing the procedure. Comparisons with continued-fraction and direct numerical results for massive scalar fields around Schwarzschild and Kerr black holes and for axial trapped modes of a uniform-density star show that the method yields accurate real frequencies and useful estimates of decay rates outside the superradiant regime.
Read the full abstract

Machine Learning

Learning Orthogonal Multi-Index Models Beyond Small Initialization: Incremental Learning, Competitive Dynamics and Symmetry

Recent work has identified incremental learning in shallow networks trained on single-index and multi-index models. However, existing analyses often rely on simplifying settings, such as small initialization, correlation loss, or layer-wise training. These choices reduce neuron interactions and leave some feature learning dynamics under standard initialization unexplored. We study training dynamics for polynomial-width two-layer networks learning orthogonal multi-index targets under standard initialization using polynomially many samples. We first prove that incremental learning still occurs: the loss decreases sequentially according to the Hermite expansion of the target, with lower-order components learned before higher-order components recover the individual target directions. In this standard initialization regime, training also shows a competitive reallocation of parameter mass: after the total mass fits the target mean and stabilizes, mass shifts into the target subspace and then concentrates on aligned neurons. Our theoretical analysis uses slightly modified gradient flow, while vanilla gradient descent empirically exhibits the same qualitative dynamics. Technically, we introduce a symmetry-based finite-width approximation via symmetrized networks, rather than comparing directly with an infinite-width limit. This yields better control of approximation errors and may be of independent interest.
Read the full abstract

All authors Mo Zhou, Weihang Xu, Simon S. Du, Maryam Fazel

Analysis of PDEs

A $p$-Laplacian Schrödinger–Maxwell system with rapidly growing nonlinearities

Let $\Omega\subset\R^N$, $N\ge2$, be a bounded domain and let $1<p<\infty$. Inspired by the exponential case treated in \cite{BO2026}, we study the quasilinear Schrödinger–Maxwell system \[ \begin{cases} -\Delta_p u+\psi G'(u)=f &\text{in }\Omega,\\ -\Delta_p\psi=G(u) &\text{in }\Omega,\\ u=\psi=0 &\text{on }\partial\Omega, \end{cases} \] where $G\in C^1(\R)$ is even, convex, nonnegative and nontrivial, with $G(0)=0$, and satisfies \[ G(t)\lesssim 1+|G'(t)| \qquad\text{for large }|t|. \] Under the assumption \[ |f|\ln(1+|f|)\in L^1(\Omega), \] we prove the existence of a finite-energy weak solution and show that the solution is a saddle point of the associated functional.
Read the full abstract

Data Structures and Algorithms

Lower Bounds for Private Graph Optimization Problems using Reconstruction Attacks

This paper studies fundamental graph optimization problems under differential privacy (DP) and shows new, reconstruction-based lower bounds. We consider a graph $G = (V, E, \vec{w})$ where the vertex set $V$ and edges $E$ are public and the weights $\mathbf{w}:E\rightarrow \mathbb{R}$ must be kept differentially private under an $\ell_1$ neighboring relation. For the problems of releasing a minimum-weight spanning tree and a minimum-weight perfect matching, we show new, tight error bounds of $\Omega(n\cdot\log(m/n)/\epsilon)$ on worst-case graphs with $n$ vertices and $m>2n$ edges. The upper bounds are known pure DP algorithms while the new lower bound holds even under approximate $(\varepsilon,\delta)$-DP as long as $\delta \leq (n/m)^{\Omega(1)}$. Our lower bounds improve the $\Omega(n/\epsilon)$ lower bounds of Sealfon (PODS~'16). The fact that approximate DP does not reduce error for MST under the $\ell_1$ neighboring relation contrasts with the recent upper bound of Pagh et al. (PODS~'25) which shows that approximate DP allows much better error under the $\ell_\infty$ neighboring relation. Going beyond worst-case graphs, we give lower bounds for large families of sparse graphs with expansion properties. We show a lower bound of $\Omega(n / \epsilon)$ for the minimum spanning tree for any graph where the minimum cut is at least $\Omega(\log(n))$. Finally, we consider the problem of private hierarchical clustering under Dasgupta's cost function (STOC~'16) and show the first approximate DP lower bound parameterized by the minimum weight of a balanced cut. This extends lower bounds of Deng et al. (ICLR~'25) to general graphs and to approximate DP.
Read the full abstract

Quantum Physics

Threshold Behavior of ZX and ZY Surface Codes Under Circuit-Level Biased and Crosstalk Noise

Studying the threshold behavior of surface codes under biased noise models is an active area of research. Previous work Tuckett et al. (2018), using an optimal tensor-network decoder, demonstrated that replacing $Z$-type stabilizers with $Y$-type stabilizers significantly improves the surface code threshold under code-capacity level dephasing noise. In this work, we construct and study a $ZY$ surface code by replacing the $X$-type stabilizers with $Y$-type stabilizers. We compare it with the standard $ZX$ surface code under circuit-level Pauli-$X$ biased noise, with and without an additional gate-based $XX$ crosstalk noise. We find that for the $ZX$ surface code, the $X$-memory threshold increases monotonically with bias while the $Z$-memory threshold decreases and saturates. For the $ZY$ surface code, the $Y$-memory threshold is nearly constant across all bias values. The $Z$-memory thresholds of the $ZX$ and $ZY$ codes are consistent within the uncertainty. Adding $XX$ crosstalk reduces the $Z$-memory threshold beyond the fitting uncertainty while leaving the $X$ memory threshold largely unaffected. The choice of CNOT ordering redistributes threshold performance between the two logical memories. Our work extends prior observations from code-capacity level noise to circuit-level noise. It also indicates the need for decoders capable of jointly reasoning over correlated syndrome information so that tailored stabilizer structures could be fully utilized for quantum error correction.
Read the full abstract

Metric Geometry

Metric Geometry of the Signature Group for $p$-Variation Rough Paths

The signatures of $p$-rough paths form a subgroup of sufficiently high-level truncated tensor algebras, whose inverse limit is a subgroup of the full tensor algebra. For $p \geq 1$, we provide a top-down description of the signature group as the inverse limit of finite-dimensional Carnot–Carathéodory geometries in the $p$-variation setting. We show that every compatible choice of metrics induces a topological tree structure on the inverse-limit group, under which the signature group is not a topological group. This extends the results of Enrico Le Donne and Roland Züst from bounded variation to rough paths. We also characterise the dependence of the inverse-limit groups and their metric completions on the choice of metric, identifying them with the tree-reduced path group of Horatio Boedihardjo, Xiang Geng, Terry Lyons, and Danyu Yang.
Read the full abstract

Optimization and Control

A Simple Complexity Lower Bound for Solving $Ax=b$

In this note, we provide a short and direct proof that approximately solving $Ax=b$ to relative error $\varepsilon$, where $A$ has condition number $\kappa$ and unrestricted dimension, requires $\Omega(\kappa\log(1/\varepsilon))$ matrix-vector multiplications in the worst case, even for randomized algorithms. This essentially recovers the lower bound of Dereziński, Epperly and Meyer [2026] for this setting, whose elegant and more general approach inspired us to seek a short direct proof. A straightforward reduction implies the classical $\Omega(\sqrt{\kappa}\log(1/\varepsilon))$ lower bound for optimizing strongly convex quadratic functions, applicable to randomized algorithms.
Read the full abstract

Artificial Intelligence

When Validation Stops Learning: Auditing Update Admission for Continual Embodied Agents

Independent evaluation can reject harmful policy updates yet also prevent useful continual learning. We argue that update admission must be assessed through both error control and retained learning opportunities at a stated interaction budget. We identify a concrete failure: a range-based confidence gate cannot certify unchanged old-task behavior within otherwise substantial budgets. A standard paired-binomial construction reduces this burden when outcome disagreements are rare. We also specify certified historical-reference promotion and a round-level missed-opportunity metric. In a constructed one-step pushing diagnostic with 32 seeds, fresh paired checks admit 31.6% of a common update stream at 2,000 episodes per stage, versus zero for the range-based gate; unconditional replay nevertheless learns better in closed-loop runs. A separate learned-dynamics stress test distinguishes model bias from feedback-selection error. The contribution is an admission-audit protocol with analytical and synthetic evidence; physical-robot and VLA validation remain open.
Read the full abstract

Logic

Elementary modal logics and circuit complexity

A modal logic is elementary if it is sound and complete with respect to an elementary class of Kripke frames. We prove that if a finitely axiomatizable modal logic $L$ is elementary, then the validity problem for $L$, which asks whether a given finite Kripke frame validates $L$, is in the circuit complexity class uniform $\mathsf{AC}^0$. As an application, we construct infinitely many finitely axiomatizable $\mathsf{K4}$-stable logics that are not elementary. This result provides a negative answer to Problem 6.4 in [Bezhanishvili et al., 2018] for $\mathsf{K4}$-stable logics.
Read the full abstract

Cryptography and Security

A2ABreak: Systematic Security Analysis of the A2A Protocol

The Agent2Agent (A2A) protocol, now governed by the Linux Foundation, is an open standard that enables autonomous AI agents to discover, authenticate with, and delegate tasks to one another across organizational boundaries. Designed to complement the Model Context Protocol (MCP) for tool integration, A2A is rapidly emerging as the horizontal communication layer of the multi-agent ecosystem. Yet the protocol's security has received no systematic analysis. This paper presents A2ABreak, the first rigorous systematic security analysis of the A2A protocol. We introduce a novel framework that utilizes an LLM-assisted extraction of a verified finite-state machine directly from the natural-language specification, producing a unified model of 37 states and 76 transitions from 929 formalized statements, and then systematically reasons over this model to discover protocol-level vulnerabilities through adversarial verification, under a full-compliance assumption. Our analysis uncovers 11 new vulnerabilities, each exploitable by a specification-compliant adversary without requiring any implementation flaw. Among the findings are cross-client context injection through unprotected context identifiers, credential harvesting via multi-hop identity loss in delegation chains, and data exfiltration through rogue agents advertising unattested capability claims. A2ABreak achieves 73.3% precision and 84.6% F1 against manual expert review, while a zero-shot LLM baseline operating over the same specification produces zero confirmed findings, demonstrating that explicit formal grounding is essential for sound protocol security analysis.
Read the full abstract

All authors Alireza Lotfi, Mirza Masfiqur Rahman, Imtiaz Karim, Elisa Bertino

Neurons and Cognition

Cortical information transfer reveals conserved hemispherical network dynamics across human handedness

Whether human motor and brain lateralization arises from fundamentally distinct neural architectures or emerges from conserved network dynamics remains a central question at the intersection of network science and neurobiology. Conventional measures of cortical activation often fail to resolve how directed information exchange adapts to manual preference during complex motor tasks. This ambiguity leaves it unclear whether left-handed individuals possess atypical neural organization or follow shared dynamical principles. Here, we combine Permutation Transfer Entropy (PTE) with network-based indices to map directed cortical signal flow using electroencephalography (EEG) during high-precision motor execution (handwriting) in both right- and left-handed individuals. We demonstrate that cortical communication is strictly scale-dependent regardless of handedness: ipsilateral dominance occurs primarily in general, network-wide interactions, whereas inter-hemispheric interactions maintain a contralateral profile. Crucially, left-handed individuals exhibit inter-hemispheric information dynamics that are functionally equivalent to right-handers when executing tasks with the same anatomical hand, refuting assumptions of mirrored or atypical lateralization. By demonstrating that human cortical lateralization relies on dynamical principles conserved across manual preferences, these findings suggest that functional asymmetries reflect adaptive optimizations governed by epigenetic or social flexibility, rather than fixed structural divergences.
Read the full abstract

Representation Theory

Fence Posets, Good Gradings and Frobenius Maximal Parabolics

Let $\mathfrak L$ be a Frobenius maximal parabolic subalgebra of $\mathfrak{sl}_n$. For any $F\in\mathfrak L^*$ for which the Kirillov form $B_F(x,y)=F([x,y])$ is non-degenerate, let $\widehat F$ denote the associated principal element. We prove that the multiplicities of the eigenvalues of $\operatorname{ad}_{\widehat F}$ on $\mathfrak L$ form a unimodal sequence symmetric about $\frac12$. We also prove that the multiplicities of the eigenvalues of $\operatorname{ad}_{\widehat F}$ on $\mathfrak{gl}_n$ form a unimodal sequence symmetric about $0$. The proof relates the ranked meander associated to $\mathfrak L$ to the order ideals of a related fence poset through Panyushev reduction. The known unimodality of the rank polynomial of the fence poset implies that of the meander, which in turn determines a good grading of $\mathfrak{gl}_n$ in the sense of Elashvili and Kac. We prove that this grading coincides with that induced by the principal element and that the pyramid associated to this grading may be filled in such a way that its good element $e$ lies in $\mathfrak L$. The two unimodality results then follow from the injectivity properties of $\operatorname{ad}_e$ coming from the good grading and the duality induced by the bilinear form $B_F$.
Read the full abstract

All authors Anthony Giaquinto, John Irving, Aaron Lauve, Mitja Mastnak

Astrophysics of Galaxies

Studying Outflows with Synthetic Absorption Line Spectra from High Resolution Simulations

Down-the-barrel absorption line spectra are a powerful probe of the properties of multiphase galactic outflows. In this work, we demonstrate a method to generate realistic synthetic spectra using high resolution ($\Delta x=5-20 \ \mathrm{pc}$), 20-kiloparsec-scale outflow simulations from the next generation of the CGOLS project. Aiming to mimic modern observational surveys of nearby star-forming galaxies, we generate a suite of UV absorption lines, from which we measure common observables such as equivalent widths and velocity statistics, and compare them to observed values. The velocity statistics from the simulated data agree well with high-resolution, far-UV HST treasury data, but the equivalent widths (EWs) display more variability. Our fiducial EWs are systematically lower than the empirical predictions by $\sim50-97\%$, particularly in the lines with higher ionization potentials, though these values increase with inclination angle, star formation rate, and time. We find a strong correlation between the measured outflow velocity and ionization potential, reminiscent of the known correlation between velocity and temperature from the simulations. We also study the effect of simulation resolution on our statistics, finding the EWs to be much more sensitive to the simulation resolution than the velocities. Overall, we find that we can generate realistic synthetic spectra from high-resolution hydrodynamic simulations that are qualitatively similar to observations. In future work, we will further improve the physical realism of our synthetic spectra by implementing a more accurate ionizing background spectrum, with the ultimate goal of improving constraints on derived outflow properties from observations.
Read the full abstract

All authors Jake Magee, Evan Schneider, Kate H. R. Rubin, S. Alwin Mao, Cameron Hummels

Analysis of PDEs

Self-Similar Singularity of the Euler Equations on $\mathbb{R}^3$

We provide evidence of a finite-time singularity in the 3D Euler equations on the unbounded domain. Using a physics-informed neural network (PINN) with a self-similar ansatz, we find an approximate singular profile for the Euler system at the critical blowup rate of $0.5$ and certify it using a spline representation. The transport field associated with the obtained profile has local outgoing property throughout the domain that suggests linear damping, a key stabilizing mechanism for the candidate profile. We also establish a framework for proving nonlinear stability of the approximate self-similar profile, reducing the analysis to a large but finite collection of explicit estimates and computable constants.
Read the full abstract