Cryptography and Security
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 31 August
We propose a novel approach to learning-based attack forensics called DUPIN. DUPIN performs unsupervised pre-training on an enormous amount of audit events in the form of provenance graphs. It then proceeds to a few-shot learning stage, leveraging a small number of labeled attack examples to fine-tune its detection capabilities. We pretrain DUPIN on up to 38 - 52 days of audit logs (7.3TB total) and evaluate it against various baselines on 25 APT campaigns across four different data sources, facilitating the scalable evaluation.
Combinatorics
On the gonality of Kneser graphs
The Kneser graphs $\text{KG}(n,k)$ are a classically studied family of graphs. One known invariant of graphs is gonality (also called divisorial gonality), which is the minimum degree of a rank 1 divisor on the graph. Using known bounds on gonality of simple, connected graphs, one may obtain that the gonality of $\text{KG}(n,k)$ is bounded above by $\binom{n-1}{k}$. In 2014, Harvey and Wood showed that the treewidth (a lower bound on gonality) for $\text{KG}(n,k)$ is $\binom{n-1}{k}-1$ for $n\geq 4k^2-3k+2$. In this paper, using scramble number, another lower bound on gonality, we improve this polynomial bound and show that the gonality of $\text{KG}(n,k)$ is exactly $\binom{n-1}{k}$ for $n\geq \frac{3k^2+k+2}{2}$, and conjecture an even stricter polynomial bound using the uniform edge scramble. We then extend our argument to the family of generalized Kneser Graphs, computing the scramble number and gonality using the same polynomial bound.
Read the full abstract
Analysis of PDEs
Stability estimates for the initial-to-final-state inverse problem
The initial-to-final-state inverse problem for the Schrödinger equation consists in determining uniquely the Hamiltonian that generates the evolution, assuming the knowledge of the initial-to-final-state map. This functional maps each initial state $f\in L^2(\mathbb{R}^n)$ to the corresponding final state at a fixed time $T$. The problem was formulated by Caro and Ruiz in the case of Hamiltonians arising from time-dependent bounded electric potentials that exhibit super-exponential decay at infinity. Caro, Parissis and the authors of this article established that uniqueness also holds for time-independent bounded potentials with super-linear decay at infinity. In this paper, we quantify the above uniqueness results establishing that the potentials are stable under small changes of the initial-to-final-state maps. In the case of time-dependent potentials we get a stability of logarithmic type. A notable improvement is achieved when the potentials are time-independent, where under this assumption we prove Hölder stability estimates.
Read the full abstract
Computation and Language
NSIDDx: A Design Framework for Neuro-Symbolic, Practitioner-First Differential Diagnosis in Low-Resource Settings
LLM-based diagnostic systems achieve high semantic accuracy on benchmarks, but open-ended evaluation on clinically uncommon presentations reveals a systematic gap between headline accuracy and verifiable clinical reliability. We evaluate an LLM+rare-disease-RAG pipeline across two cohorts and show that the paradigm produces confident outputs that are frequently unverifiable and systematically resistant to clinician interrogation. We present NSIDDx (Neuro-Symbolic Integrated Differential Diagnosis System), a design framework arguing that DDx systems in low-resource settings must treat the clinician as an active reasoning agent. We instantiate this through a neuro-symbolic pipeline with ternary symptom encoding, contradiction detection, audit strings, and practitioner override - running offline on consumer hardware. We distill five design principles for clinician-in-the-loop clinical NLP and invite the prospective studies needed to validate the claim at scale.
Read the full abstract
Mathematical Physics
Szegő-type determinant asymptotics for generalized Hilbert matrices with edge eigenvalues
We compute the large-$N$ asymptotics of the determinant of the identity plus or minus a generalized Hilbert matrix. For $N \in \mathbb{N}$ and $\delta \in \mathbb{R} \setminus \{-1, -2, \ldots\}$, let \[ H_N^\delta := \left( \frac{\sin((1+\delta)\pi)}{\pi(j+k+1+\delta)} \right)_{j,k=0}^{N-1} \] be the generalized Hilbert matrix. For $\delta$ not a half-integer, we prove the large-$N$ power-law asymptotics of the determinant \[ \log \text{det}(I_N \pm H_N^\delta) = -\frac{\theta_\delta^2 \pm \theta_\delta}{2} \log N + O(1) \] where $\theta_\delta := \delta$ if $\delta < -1/2$, and $\theta_\delta := \frac 1 \pi \arcsin(\sin(\delta\pi))$ if $\delta \ge -1/2$ and $\arcsin \colon [-1,1] \to [-\frac{\pi}{2}, \frac{\pi}{2}]$ denotes the principal branch of $\arcsin$. For $\delta \geq -\frac 12$ the asymptotics is known. The novelty of the present paper is the regime $\delta < - \frac 12$ and understanding the dichotomy in the decay exponent. This stems from the $\pm 1$ edge eigenvalues of the limiting operator appearing for $\delta < - \frac 12$. Most notably, we obtain for $\delta < -\frac12$ \[ \log \det\bigl(I_N - (H_N^\delta)^2\bigr) = -\delta^2 \log N + O(1). \] Such determinants arise in the study of Anderson's orthogonality catastrophe.
Read the full abstract
Differential Geometry
On the generalized geometry of almost abelian solvmanifolds
We study left-invariant generalized complex structures on almost abelian Lie groups $G_A$ with Lie algebra $\mathfrak{g}_A=\mathbb{R} e_0\ltimes_A\mathfrak{h}$, where $\mathfrak{h}$ is an abelian ideal of codimension one, and on their compact quotients. First, when $A$ is diagonalizable over $\mathbb{R}$, we characterize all admissible types by pairings of its eigenvalues and characterize the structures admitting a closed invariant pure-spinor generator. For general $A$, we obtain Jordan-theoretic type bounds and a construction using a complex quotient and a symplectic ideal. Finally, in dimension six, we establish nonexistence results and give explicit intermediate-type constructions.
Read the full abstract
Group Theory
Higher Massey Products in Demuškin Variations: Support Blocks, One-Relator Reduction, and a Five-Fold Vanishing Case
Blumer and Quadrelli introduced a family $\mathcal{F}_2$ of two-relator pro-$p$ groups obtained from a Demuškin group by imposing the commutativity of two generators which are not paired in the Demuškin relation. They proved the triple and quadruple Massey vanishing properties and asked whether the same holds in every length. We introduce the $z$-profile \(v_h=(\alpha_h(z_1),\alpha_h(z_2))\in\Fp^2\) of a defined $n$-fold Massey product. Definability forces $\det(v_h,v_{h+1})=0$, so the nonzero profile entries decompose into support blocks carrying projective directions in $\PP^1(\mathbb{F}_p)$. After the known endpoint reduction, profiles with exactly $r$ blocks are counted by $\binom{n-1}{2r}$, and $r$ blocks first occur in length $2r+1$. We also prove an all-length conditional reduction for Dwyer's lifting problem: if the added commuting relator can be made exact in a lift, then the remaining central defect of the Demuškin relator can be removed by an endpoint correction. The correction preserves the power term $x_1^q$ for every parameter allowed in $\mathcal{F}_2$. In length five, this yields vanishing whenever all three interior profile vectors $v_2,v_3,v_4$ are nonzero, for every prime $p$. The resulting classification identifies the remaining five-fold support types and the compatibility mechanisms governing them.
Read the full abstract
Software Engineering
Spec-Driven Development for Agentic Software Engineering: Harnessing Human-Agent Teamwork
Context: Software engineering is moving from AI-assisted practices like vibe coding, in which assistants accelerate individual developers, towards Agentic Software Engineering (ASE), in which autonomous agents are delegated goal-level tasks. However, industry reports a productivity paradox: as individual productivity increases, team throughput, review capacity, and stability degrade because team-scale software engineering discipline is neglected. Objective: This paper aims to establish the conceptual and methodological foundations of Spec-Driven Development (SDD) as an enabling discipline for ASE at team scale and characterize the harness, i.e., the technical and methodological mechanisms through which teams govern agent behavior. Method: We conducted a conceptual analysis drawing predominantly on gray literature, including ASE vision and roadmap papers, practitioner reports, talks, and tooling, because peer-reviewed evidence and a shared academic-industrial vocabulary are not yet established. Results: Using a comparative characterization of the paradigm progression as conceptual framing, the article presents (i) a socio-technical model of SDD in which specifications act as the contract substrate between humans and agents; (ii) an operational characterization of the harness, distinguishing the technical harness around the agent from the methodological harness around the team, with worked examples; and (iii) a typology of five human–agent interaction patterns through which the human role is redefined. Conclusion: We conclude that SDD reconstitutes, in specification-centric form, the contracts that vibe coding dissolves: accountability, verifiability, and transferability. Given the immaturity of the evidence base, this work is presented as a first step toward academic-industrial consensus rather than a validated theory, and outlines a research agenda for future empirical validation.
Read the full abstract
Artificial Intelligence
Hypotheses-Guided Self Distillation for Continual Personalization
As people increasingly interact with LLM assistants in daily life, continually adapting to individual preferences has become essential for effective long-term interactions. However, user preferences are rarely stated in full, and instead emerge through heterogeneous, latent, and noisy signals, with existing methods relying on raw interaction histories or costly reward-based optimization to manage personalization. We introduce HypReflect, a reliable, scalable framework for continual personalization that infers explicit, uncertainty-aware preference hypotheses from diverse user signals, reflectively refines them as new evidence accumulates, and incorporates the resulting user model through hypotheses-guided self-distillation. Experiments across three personalization settings: online personalization, multi-session interactions, and implicit behavioral signals, show that HypReflect outperforms a range of baselines, including raw-history and incremental-update methods. We further demonstrate strong generalization to unseen users and cross-domain settings, along with stability across context budgets, reusable hypotheses, and more focused personalization. These results suggest a step towards reliable and scalable continual personalization through explicit, revisable user preference hypotheses.
Read the full abstract
Computers and Society
CompanionSim: Synthetic Data for Evaluating Anthropomorphism in Human-AI Relationships
Many people now see AI systems as not just productivity tools but as social companions. Researchers are eager to study the consequences of AI companionship behaviors, such as validation, which evoke trust, empathy, and attachment in human-human interaction. However, human-AI interaction data is limited and unreliable, slowing research progress. We scale small amounts of real-world data by simulating multi-turn human-chatbot dialogue across a range of chatbot behaviors and use cases. We release CompanionSim: a simulation framework with 2,240 simulated human-chatbot conversations representing 16 chatbot behaviors across seven use cases. Human participants annotated the simulated conversations and real-world conversations in two experiments probing perceptions of companionship behaviors. We conducted Study 1 with a U.S. representative sample ($N_{1}~=~628$) and Study 2 across the U.S., U.K., India, and Nigeria ($N_{2}~=~3,646$). Surprisingly, we find that companionship behaviors reduced likability, humanlikeness, and trust in AI chatbots. These effects were larger in particular subgroups: women and older participants saw companionship chatbots as less likable, humanlike, and trustworthy. We encourage researchers to leverage real-world and synthetic data together to study the differential impacts of AI companions and to create benchmark evaluations of AI chatbots.
Read the full abstract
Instrumentation and Methods for Astrophysics
Flux jump correction in the QUBIC TES data pipeline
Flux jumps occur in the readout electronics of Transition Edge Sensor (TES) detectors coupled to Superconducting Quantum Interference Devices (SQUIDs), as a consequence of the periodic transfer function of SQUIDs in the superconducting regime. These events manifest as discontinuities in the Time Ordered Data (TOD) and can significantly degrade the quality of the measured signals. In this work, we present an automated pipeline to identify and correct flux jumps in the calibration data of the QUBIC (Q & U Bolometric Interferometer for Cosmology) instrument, designed to measure the B-mode polarization of the Cosmic Microwave Background (CMB). Using data from multiple observation campaigns, we characterize the statistical properties of flux jumps and compare them with expectations based on the electronic configuration of the instrument. The proposed pipeline successfully corrects these events, allowing the recovery of affected data and enabling their use in subsequent CMB analysis.
Read the full abstract
Artificial Intelligence
Authority Bias in Conversational Search Engines for Academic Paper Recommendation
Large Language Models (LLMs) are increasingly used as conversational search engines for academic literature, yet whether they judge papers on content or on authority signals has not been tested causally. We investigate authority bias: systematic preference for papers based on author prestige, venue, and citations rather than content. Holding title and abstract constant, we vary authority metadata across three counterfactual conditions (original, flipped, boosted) over eight LLMs (five open-weight and three frontier closed-weight) in an in-context, single-turn, top-1 recommendation setting. Our experiments show that authority bias is substantial and directional, varies markedly across models, and is only partially addressable through prompt-level debiasing. We further document a say-do gap: debiasing instructions suppress authority mentions far faster than authority-driven flips, so surface auditing systematically underestimates behavioral bias.
Read the full abstract
Software Engineering
Empirical Software Engineering in Practice: Insights from Google
While it is fairly well known how empirical software engineering (ESE) is used in the academic world, we have limited knowledge of how ESE is practiced in industry. As part of our regular column on empirical software engineering (ACM SIGSOFT SEN-ESE), we want to dedicate a series of articles to interviewing ESE practitioners from various companies. Among other things, we want to understand how ESE processes are implemented in industry, e.g., different research methods, how practitioners decide on what to study, how research results are used within companies and beyond, and if they face recurrent impediments to using ESE methods in industrial contexts. In the first edition of "ESE in Practice", we are joined by Ciera Jaspan and Collin Green from the Developer Intelligence team at Google. This article is a faithful account of our conversation from August 13, 2026, which we edited for our column.
Read the full abstract
Programming Languages
Beyond Locks and Thread IDs: Static Data Race Detection Off The Beaten Path (Extended Version)
Maintaining an abstraction of the execution history of threads can improve the precision of data race detection in static analysis. Here, we extend the digest framework to handle concurrency constructs and synchronization mechanisms that have been ignored in static race detection. We introduce mechanisms for the commonly used thread barriers, as well as pthread_once, which allows to ensure that an action is executed only once. We also instantiate the framework with an abstraction of locksets held by ancestor threads. We propose a suite of litmus tests to evaluate analyses for these features and compare our implementation to state-of-the-art tools, finding that they lack support.
Read the full abstract
Analysis of PDEs
Higher-order Gaussian bounds for maximally subelliptic boundary value problems
We establish higher-order Gaussian upper bounds for the heat semigroups associated with a broad class of sectorial maximally subelliptic quadratic forms on manifolds with boundary. Near non-characteristic boundary points and in the interior, we obtain pointwise bounds for all mixed derivatives of the heat kernel in time and along the Hörmander vector fields, expressed in the associated Carnot–Carathéodory geometry. The results apply to operators of arbitrary even order, systems, nonsymmetric forms, and boundary conditions beyond the Dirichlet case.
Read the full abstract
Number Theory
From Common-Slot Chains to Heisenberg Central Products over Global Fields
Let $F$ be a global field and let $p$ be an odd prime with $p\neq\operatorname{char}F$. Assuming first that $\mu_p\subset F$, we record in a uniform global-field form the length-four common-slot chain for equal degree-$p$ symbol classes and emphasize the signed normalization adapted to explicit norm constructions. Thus, from \[(a,b)_p=(c,d)_p\in\textrm{Br}(F)[p]\] one obtains $x,y\in F^\times$ such that \[(a,b)_p=(x^{-1},b)_p=(x,y)_p=(c^{-1},y)_p=(c,d)_p.\] For number fields, the underlying chain is the length-four chain lemma of Gille–Szamuely, based on Tate's simultaneous local–global theorem. The point developed here is that its signed form yields four compatible norm equations that can be used constructively. For the extraspecial central product $H_{p^3}*H_{p^3}$ over a $C_p^4$-Kummer extension, the central-embedding obstruction is $(a,b)_p-(c,d)_p$, while the four norm equations supplied by the chain assemble, under a natural independence hypothesis on the auxiliary Kummer classes, into an explicit factorized radical realization of the central product. Finally, when the ground field does not contain $\mu_p$, we show by a restriction–corestriction argument that the central-embedding obstruction is detected after passage to the cyclotomic extension $F(\mu_p)$. Over that field the problem is Kummer, and the obstruction is again the difference of the two symbol classes.
Read the full abstract
Artificial Intelligence
Invalidation Contracts for Cross-Episode Agent Memory
LLM agents that cache recovery suggestions from API errors can skip re-derivation in later episodes, spending fewer tokens and fewer model calls on constraints they have already learned. Server-side data drift turns those cached fixes into silent failures, and the usual remedy, re-deriving on every episode, gives the savings back. We introduce invalidation contracts, a protocol layer that attaches version stamps and cacheability hints to every recovery suggestion so the client can evict stale entries without trial and error, and keep the rest. The contract decomposes realized savings into two independent factors: validity, the fraction of cached suggestions that remain correct after a drift event, and compliance, the fraction the planner applies on the first attempt. Validity depends only on the protocol and is vendor-independent. Compliance depends on the planner model: identical wire bytes yield 100% first-try compliance on Claude Haiku 4.5 and 11% or below on Claude Sonnet 5, which exhibits input-schema conservatism, refusing fixes that add fields the original request did not contain. We evaluate across seven models, three serving paths, two domains, and approximately 9,400 episodes. Row-level invalidation raises compliance by 0 to 66.7 percentage points across the seven models, 55.6 to 66.7 on three, and recovers 29-33% of baseline token cost on four of seven models, while table-level invalidation destroys co-located entries and drops post-drift first-try rates to 0% on five of seven. Eviction precision is 1.00 at row granularity on every model under the row-level oracle of Section 4.1. The contract adds 15% to response payload. Version-stamp validity is deterministic by construction and produced identical results across every model and serving path, with zero contract failures in the entire evaluation.
Read the full abstract
Computer Vision
CoLT-Drive: Counterfactual Long-Tail Benchmarking and Knowledge-Preserving Adaptation for Driving Affordance Prediction
Long-tail autonomous driving failures are often framed as rare-object recognition errors. We argue that this view is incomplete: the decision-critical question is not only whether a model recognizes an unusual object, but whether it infers how that object changes the ego vehicle's feasible high-level actions. We formalize this problem as decision-level driving affordance prediction, where a model maps a front-view image, ego-motion history, and navigation command to a structured longitudinal–lateral meta-action. To evaluate this capability, we introduce CoLT-Drive, a 3,536-sample counterfactual long-tail benchmark that inserts rare objects into otherwise fixed driving scenes and measures whether models predict acceptable action pairs. To improve deployable small VLMs, we propose KPA, a knowledge-preserving adaptation framework that combines structured perception-to-decision prompting, SLERP-based expert merging, and RegMoE, a regime-aware LoRA mixture-of-experts module. KPA preserves the pretrained model's open-world knowledge while allocating lightweight adaptation capacity to different driving decision regimes. Experiments on an in-domain driving split and CoLT-Drive show that KPA achieves 60.8% pair accuracy on CoLT-Drive, outperforming the pretrained Qwen3-VL-2B baseline (50.3%) and LoRA SFT (32.4%) while maintaining competitive in-domain accuracy. Our benchmark and code are available at https://huggingface.co/datasets/tangzx2024/CoLT-Drive and https://github.com/tangzhengxu/CoLT-Drive.
Read the full abstract
Computation and Language
LOOMSUM: Weaving Quantitative and Narrative Evidence for Faithful Long Text-Table Summarization
Long documents often distribute important information across extensive narrative passages and multiple tables, making faithful summarization particularly challenging. Existing methods may generate individually supported quantitative facts and analytical statements yet associate them incorrectly, producing quantitatively plausible yet analytically unfaithful summaries. In this work, we propose LOOMSUM, a training-free framework that extracts source-grounded atomic evidence, explicitly links table-derived facts with supporting narrative analyses, and plans the discourse structure before generation. We also introduce Table-Grounded Faithfulness (TGF), a claim-level metric that separately evaluates Numeric Grounding, Analysis Support, and Relation Consistency. Experiments on the text–table summarization benchmarks FINDSum and USTT show that LOOMSUM improves analytical faithfulness while maintaining strong summarization quality. Human evaluation finds positive component-level associations with the corresponding human judgments. Our Relation Consistency metric further shows stronger agreement with human relation judgments than generic factuality metrics, indicating that explicit cross-modal linking helps reduce errors in which supported quantities are paired with incorrect narrative interpretations. Together, these findings show that faithful long text–table summarization requires not only grounding individual facts, but also preserving the relations between them.
Read the full abstract
Computational Complexity
Bounded Relative Boundary Implies Narrow DNF Approximation
We prove Friedgut's conjecture (J. Amer. Math. Soc. 12 (1999)) that every increasing family in the $p$-biased discrete cube with bounded relative boundary can be approximated arbitrarily well by one whose minimal elements have size bounded independently of the dimension and the bias. For $0<p\le 1/2$, every increasing Boolean function with total resampling influence at most $K$ is $\varepsilon$-close under $\mu_p^n$ to a monotone DNF of width $\exp(O((K+1)^2/\varepsilon^2))$, from which the full conjecture follows for all $p\in(0,1)$. We build on Hatami's pseudo-junta theorem (Ann. of Math. 176 (2012)), whose approximator need not be increasing. Our randomized shifts produce an increasing approximator that is measurable with respect to a controlled refinement of Hatami's adaptive partition. Retaining only short positive certificates yields the desired narrow DNF.
Read the full abstract
Mesoscale and Nanoscale Physics
Entanglement from particle number fluctuations in a second-order topological insulator
Higher-order topological insulators are materials with a bulk energy gap featuring gapless modes at $(d-n)$-dimensional edges, where $n > 1$. Despite this modified bulk-boundary correspondence, their nontrivial topology can nevertheless be observed through the bipartite entanglement spectrum. Inspired by works suggesting fluctuations in conserved quantities could be used to study topology in one-dimensional systems, which likewise feature zero-dimensional edge modes, we extend this treatment to a two-dimensional second-order topological insulator where corner states play an analogous role. We show that the standard bipartite entanglement entropy and particle number fluctuation lack striking signals of topological phase transitions, both in the static case and in the time evolution after a quench, obscured by an area-law term from the subsystem edges. By introducing a quadripartite construction of the subsystem, we show that it is possible to isolate the topological contributions, producing sharp transition peaks in both quantities, observable by studying subsystems much smaller than the full system. As particle number fluctuations are accessible in existing experiments, these results establish a concrete and scalable path to experimental detection of the entanglement in higher-order topological insulators.
Read the full abstract
Sound
XVAE-WMT: Explainable Wavelet-Temporal Variational Autoencoder for Blind Source Separation of Heart and Lung Sounds
The separation of cardiovascular sounds is a critical task in biomedical signal processing. In this paper, we introduce XVAE-WMT1, an unsupervised explainable generative AI algorithm combining a variational autoencoder (VAE) with explainable AI (XAI), wavelet-based inputs, a post-hoc output mask, and temporal consistency (TC) loss. Unlike existing supervised and VAE-based methods that rely on Short-Time Fourier Transform (STFT) and ignore latent interpretability, XVAE-WMT requires no paired clean recordings and integrates a Continuous Wavelet Transform (CWT) front-end for superior time-frequency localization. We assessed the latent space interpretability via different metrics, with SHAP (SHapley Additive exPlanations) enabling dimensionality reduction to the top 75% of latent features while preserving separation quality. Evaluated across two datasets using Signal-to-Distortion Ratio (SDR), Signal-to-Interference Ratio (SIR), and Signal-to-Artifacts Ratio (SAR), XVAE-WMT attains 26.8 dB SDR, 32.8 dB SIR, and 28.6 dB SAR.
Read the full abstract
Artificial Intelligence
Learning What to Retain: Gated-Memory Routing for Efficient Collaboration in Multi-Agent LLM Systems
Large language model (LLM)-based multi-agent systems tackle complex reasoning by orchestrating how multiple agents are configured and how they collaborate. A central challenge is to adapt orchestration to the evolving collaboration state. Routing from the query alone cannot adapt to intermediate progress or errors, which hurts accuracy. Routing from the complete execution history supplies this missing context, but forces later decisions to process every prior step, including redundant or low-utility ones. This creates an execution-history overload that inflates cost. Effective orchestration instead requires a compact state that captures useful progress without accumulating redundant context. We propose Gated-Memory Routing, which conditions each decision on the query and a learned execution memory. A learned Memory Write Gate commits only non-redundant reasoning steps, and a learned Retrieval Gate supplies each agent a compact, relevant subset, so every decision conditions on a clean, informative state. At each step, the system selects the next role and backbone from this memory, while an Adaptive Halting Controller stops execution once the memory contains sufficient evidence for answering. Across five reasoning and code-generation benchmarks, our framework is both effective and efficient: it attains the best average accuracy, exceeding the strongest baseline by 2.44 points, while reducing HumanEval inference cost by 31.9% relative to that baseline. Code is available at https://github.com/rajibrhasan/gated-memory-routing
Read the full abstract
Information Theory
An Approach to Asynchronous Unsourced Random Access
In this work we propose an approach to construct a fully asynchronous unsourced random access communications (URA) system. Besides the common URA features such as the absence of the user identification information in the packet and the decoding oriented to receiving the content of the messages, there is no limitation on the user transmission timing. The active users can transmit their packet any time on demand. The proposed system belongs to the class of preamble-payload URA formats. Contrary to the typical role of the preamble to serve as a temporary user identification, the preamble serves for the purpose of timing acquisition. The proposed system also employs a very small pool of preambles to resolve collisions of packets transmitted simultaneously.We provide a performance analysis and simulation results for the proposed system and demonstrate that for a wide range of signal-to-noise ratios (SNR)s the supported numbers of active users are close to the numbers of users supported by a slotted URA system counterpart.
Read the full abstract
Quantum Physics
Optimization Landscape Geometry in VQE for Frustrated Quantum Spin Models
We benchmark eight classical optimizers for exact-statevector VQE calculations on a controlled hierarchy of frustrated spin models, ranging from a diagonal Ising glass to transverse-field Ising and anisotropic Heisenberg models. The benchmark includes local, stochastic-gradient, evolutionary, covariance-adaptation, and swarm-based optimization methods under matched function-evaluation budgets. To understand their performance beyond final energies, we characterize the underlying Hamiltonian–ansatz landscapes in terms of local minima, gradients, curvature, and ground-state reachability. We use simple variational circuits, from an $R_y$ product-state ansatz for the diagonal model to shallow $R_y$–CNOT hardware-efficient circuits for the noncommuting models, and study how increasing circuit depth changes their expressivity, reachability, and optimization geometry. We find that optimizer performance changes substantially across the model hierarchy and is closely connected to landscape structure, while the variational gap represents a separate source of error. These results show how classical optimization, variational expressivity, and landscape geometry jointly determine VQE performance for frustrated spin models.
Read the full abstract
Quantum Physics
Codes for Quantum Secret Sharing with a Helper
Helper quantum secret sharing is a form of secret sharing defined by its unique access structure. One special fixed party, called the helper, can work together with any other party to fully decode the secret. A blind helper is one who can provide this assistance while not holding any local information about the encoded secret. In this work, we analyze the general structure of QSS helper codes and present new code constructions. We fully characterize the structure of blind helper stabilizer codes and show that for the encoding of a single qubit, recovery is always possible using one-way local operations and classical communication (LOCC) from the helper to the targeted party. Furthermore, we demonstrate how such codes also allow for the helper to target larger subsets of parties by one-way LOCC, enabling them to be authorized to recover the secret. Finally, when each party only holds a qubit, we identify the general form of all helper codes (including non-stabilizer codes) and find that LOCC recovery is only possible in special cases.
Read the full abstract
Geometric Topology
A survey on mapping class groups of 3-manifolds
We survey computations and tools concerning the mapping class group of a compact, oriented, connected 3-manifold $M$. We provide a guide to the literature and sketch proofs for various families of irreducible and geometric 3-manifolds. We also consider JSJ and prime decompositions of 3-manifolds, and consequences for their mapping class groups.
Read the full abstract
Computer Vision
Beyond Blind Compliance: Benchmarking Task Verification in OCR Reasoning
Multimodal Large Language Models (MLLMs) have achieved strong performance on OCR-centric document understanding and text-rich visual reasoning benchmarks. Yet existing evaluations largely assume that every task is valid and answerable. In real-world OCR scenarios, this assumption often fails: questions may rely on illegible text, occluded evidence, nonexistent visual targets, contradictory premises, or missing variables. We study this reliability gap as OCR-grounded Task Verification: before answering, a model should determine whether the Image Premise (IP), Textual Premise (TP), and Question (Q) jointly define an executable task. We introduce VeriOCRBench, a 1,800-sample human-verified benchmark built from source images drawn from 8 OCR-related datasets and spanning 8 real-world image domains, with controlled, image-grounded diagnostic tasks. It contains 1,600 trap-injected invalid tasks across 8 trap types and four verification dimensions—Visual, Contextual, Factual, and Logical—plus 200 trap-free controls for measuring over-refusal. Built with a Visual Atomic Fact (VAF)-anchored pipeline and full human auditing, VeriOCRBench enables decoupled evaluation of task verification, root-cause diagnosis, and over-refusal. Evaluating 15 leading MLLMs reveals persistent blind compliance, diagnosis failures, and prompt-induced over-refusal, exposing a critical reliability gap in current OCR reasoning systems. The code is available at: https://github.com/zy001122/Beyond-Blind-Compliance.
Read the full abstract
Computer Vision
Beyond Language Priors: Diagnosing and Fixing Visual-Origin Hallucinations in Multimodal LLM
Existing research on object hallucination in multimodal large language models (MLLMs) predominantly attributes the problem to language priors such as over-reliance on textual co-occurrence statistics. We challenge this view by presenting quantitative evidence for a complementary, under-explored cause: visual-origin hallucination, where hallucinations arise from incorrect visual feature extraction and misalignment between image and text embeddings. Through cosine similarity analysis and Smooth Grad-CAM entropy measurements, we show that hallucinated samples exhibit systematically lower image-text similarity (average 0.158 vs. -0.122) and inverted attention patterns, where attention is dispersed when the target object is present but wrongly concentrated when it is absent. Guided by this diagnosis, we propose Adversarial Contrastive Fine-Tuning (ACFT). ACFT uses an Adversarial Hallucination Attribute Flipping (AHAF) procedure, involving minimal, targeted adversarial perturbations that flip an image's hallucination attribute, to construct perfectly aligned positive-negative pairs, which are then used for contrastive fine-tuning. AHAF simultaneously serves as a diagnostic probe, revealing that MLLM visual representations lie dangerously close to hallucination decision boundaries. Requiring only 0.9% of the COCO dataset and adding zero inference overhead, ACFT achieves state-of-the-art performance on POPE, MME, and four description-level hallucination benchmarks across LLaVA, MiniGPT-4, and Qwen2.5-VL. Code is available at https://github.com/zxp555/ACFT_MM
Read the full abstract
Nuclear Theory
Probing Dense Nuclear Matter at Small-x: A workflow for a global analysis framework
The dipole model provides a powerful framework for describing high-energy nuclear interactions, particularly in the regime of dense gluonic matter. However, accurately evolving the dipole–nucleus scattering amplitude remains a major computational challenge because it is governed by nonlinear QCD evolution equations. To address this, we investigate a machine learning (ML) model as an efficient surrogate for the conventional numerical evolution. These ML-based approximations dramatically reduce the computational cost of global analyses while maintaining the accuracy required to describe a broad range of experimental data. We systematically evaluate the ML results for accuracy, computational efficiency, and ability to capture essential features of dipole evolution in nuclear environments. These computational advancements will enable global analyses of diverse datasets within both the dipole and parton model frameworks, providing a more rigorous probe of nuclear structure in the dense regime. Comparing both descriptions within a common fitting framework can provide precise constraints on the gluon distributions and advance our understanding of the quark and gluon structure of nuclei, particularly in the small-x region.