Peer Revenue Publish or profit

Papers

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

All of arXiv

Monday 7 September

Combinatorics

Global and local degree conditions for matchability

A corollary of Hall's marriage theorem is that a sufficient condition for a list $(V_1, \ldots ,V_m)$ of sets to have a system of distinct representatives is that $|V_i|\ge deg_{\{V_1, \ldots ,V_m\}}(v)$ for every $i\in [m]$ and $v \in \bigcup_{i\in [m]}V_i$. This we dub a global condition. A folklore result is that a local condition - that the inequality holds for pairs $i,v$ for which $v \in V_i$ - suffices. These are special cases of a general type of results - large sets, whose elements are sparse in some sense, have a system of representatives that is independent in a related graph. We study two such scenarios, in both of which each $V_i$ is replaced by a $k$-uniform hypergraph $H_i$, the representatives are hyperedges, and distinctness is replaced by disjointness. In one setting the sparsity is measured by the degrees of vertices in the hypergraphs, in the other by the degrees of vertices in the line graph. The proofs use the topological version of Hall's theorem. In particular, we shall use a lower bound on the topological connectivity of the independence complex of a graph, defined by vector representations. We also provide short proofs of the local infinite version, known as the “Milner-Shelah theorem”.
Read the full abstract

Computation and Language

In-Place Instruction Following in Diffusion Language Models

Diffusion Large Language Models (dLLMs) generate text via bidirectional iterative denoising, naturally supporting user-specified constraints anchored at arbitrary output positions, a paradigm known as In-place Prompting (IPP). We formalize this as the In-place Instruction Following (IIF) task and construct IIF-Bench, a hierarchical benchmark spanning literal, style, and discourse-function constraints, paired with a rubric-based local-global evaluation protocol. An inference-time attention-bias probe suggests that vanilla dLLMs often under-prioritize constraint spans during denoising. We then propose GRAFT, an IPP-oriented post-training framework combining constraint-aware SFT and preference optimization. On four representative dLLMs, GRAFT raises the average IIF score from 57.75 to 73.10 (+15.35 points), with absolute gains of 15.91 and 15.57 points on literal and discourse-function constraints, while preserving general generation ability.
Read the full abstract

All authors Zheng Nie, Zherui Li, Jiaming Zhang, Kun Wang, Zhenhong Zhou, Yufei Guo

Computer Vision

Unsupervised Domain Adaptation for Symbol Spotting in Historical Encrypted Manuscripts

The decipherment of historical encrypted manuscripts poses a fundamental challenge in Digital Humanities: before any transcription can begin, the symbol inventory of the underlying cipher alphabet must first be identified and characterized. We address this challenge through symbol spotting: given a candidate alphabet specified as a set of rendered font glyphs, the task is to determine whether and where its characters appear in an unseen handwritten document, without any labeled examples from the target script. The main difficulty lies in the domain gap between clean, digitally rendered font queries and degraded handwritten manuscript symbols. We propose a three-stage pipeline that bridges this gap without manual annotation, combining a joint SimCLR+DANN encoder for domain-invariant glyph representations with an embedding-space style-adaptation mechanism applied at retrieval time, requiring no re-training. Experiments on fourteen pages from seven encrypted manuscript collections show that our method outperforms zero-shot foundation models, including CLIP and DINOv2, by a large margin ($+0.194$ P@1 over CLIP ViT-L/14), and surpasses task-specific trained baselines by $+0.138$ P@1. We further demonstrate that the Raw-Cover metric, computed in a fully unsupervised setting, provides a meaningful script-family fingerprint that identifies the underlying alphabet of an unknown document. This capability is of direct practical relevance to palaeographers, historians, and other researchers working with undeciphered manuscripts.
Read the full abstract

All authors Giuseppe De Gregorio, Alicia Fornés, Lei Kang, Beáta Megyesi

Analysis of PDEs

The capillary $L_p$ dual Minkowski problem for $p>q$ in higher dimensions

We prove that the capillary $L_p$ dual Minkowski problem in $\mathbb{R}^{n+1}$, $n\geq3$, admits a unique positive smooth strictly convex solution for any prescribed positive smooth function $f$ when $p>q$ and the contact angle satisfies $\theta\in(0,\pi/2)$. Earlier results for $p>q$ either required the additional condition $q\leq1$ or were restricted to $n=2$. We reduce the problem to a Monge–Ampère type equation with a Robin boundary condition on a spherical cap. By constructing a new auxiliary function for the global $C^2$ estimate, we overcome these limitations and completely resolve the case of $p>q$ in higher dimensions, i.e., prove existence and uniqueness throughout the case of $p>q$, without any further restriction on $p$ or $q$.
Read the full abstract

Applied Physics

Physics-Constrained Inverse Estimation of Irradiation-Induced Strain in He-H Ion-Implanted 4H-SiC Using Nanoindentation and Finite Element Modeling

Nanoindentation is widely used to evaluate the mechanical properties of irradiated materials however its potential for quantifying irradiation induced subsurface strain remains underexplored. In this work, an integrated experimental numerical framework based on a physics constrained inverse modeling approach is employed to estimate the magnitude of a depth dependent irradiation induced strain distribution in single crystal 4H SiC following sequential He and H ion implantation. The approach combines depth sensing nanoindentation, finite element modeling FEM, and a simplex based inverse optimization routine to calibrate a physically motivated eigenstrain profile derived from ion damage simulations. The strain field is assumed to follow a lognormal distribution consistent with independently determined damage profiles SRIM, and is implemented in the FEM model through a depth dependent thermal expansion formulation. By minimizing the squared error between simulated and experimental force displacement curves, the peak tensile strain is estimated to be 0.91, accompanied by an effective Young s modulus of 310 GPa and a yield strength of 16.4 GPa. Independent validation using nano beam precession electron diffraction N PED confirms good agreement between the reconstructed and experimentally measured out of plane strain profiles in both magnitude and spatial distribution. The results demonstrate that nanoindentation, when combined with physics based inverse modeling, can provide a practical tool for quantifying irradiation-induced strain and residual stress in nuclear ceramics. This methodology offers a complementary approach to diffraction based techniques for assessing subsurface damage in ion irradiated materials relevant to advanced nuclear systems.
Read the full abstract

All authors M. Bensalem, N. Daghbouj, J.Duchon, B.S. Li, A.T. AlMotasem, S. Magalhaes, A. Yie, F. Munnik, Xin Ou, W.J. Weber, T. Polcar

Cryptography and Security

Frequency-Domain Mixing Data Augmentation for Malicious Traffic Detection

The strong dynamics of network traffic often force malicious traffic detection models to handle out-of-distribution data. Typically, deep learning-based malicious traffic detection models require a large amount of high-quality training data. However, owing to challenges such as high labeling difficulty and resource consumption, existing datasets often suffer from insufficient diversity and fail to capture evolving traffic patterns, leading to poor out-of-distribution generalization ability of the trained models. Data augmentation has been widely adopted to improve data diversity and model generalization. Recently, frequency-domain mixing augmentation has shown promising performance because it effectively perturbs data while preserving key structural information. This approach shows potential for enhancing malicious traffic detection models. However, existing studies lack theoretical interpretation of the mixing mechanism, and do not adapt to the characteristics of network traffic. In this paper, we first conduct a theoretical analysis of the current frequency-domain mixing method, revealing its underlying principles and limitations. We further propose an improved frequency-domain mixing-based data augmentation method for network traffic data, which enhances the diversity of sequence features in network traffic and improves the out-of-distribution generalization of malicious traffic detection models. Extensive experiments on multiple artificial and real-world datasets demonstrate that our method substantially improves detection performance across diverse network environments and outperforms other data augmentation approaches.
Read the full abstract

Computer Vision

Mind the Approximation: Fisher-Weighted SVD Compression for ViTs

Model compression is key to mitigate deployment challenges of ever growing machine learning models. In this area of research, singular value decomposition (SVD)-based compression offers a compelling trade-off between computational efficiency and model accuracy. Fisher-weighted SVD in particular provides principled, loss-aware compression. However, we find that improving the fidelity of Fisher approximation used in the compression is poorly predictive of post-compression accuracy for Vision Transformers (ViTs). Motivated by this observation, we propose FACTS, a structured Fisher Approximation tailored to Compressing ViTs with Fisher-weighted SVD, which enforces token-local aggregation while preserving within-token activation-gradient dependence. Additionally, we introduce a fast Constrained Rank Search (CoRS), that optimizes layer-wise rank allocation while adhering to a fixed floating point operation (FLOP) constraint. Extensive experiments across ViTs and hybrid architectures demonstrate that FACTS consistently improves accuracy-efficiency trade-offs without requiring finetuning. Notably, it outperforms the strongest SVD baseline by up to +5.8 percentage points (p.p.) Top-1 on Swin-B, with further gains driven by our search method. Code is available at https://github.com/MoritzTho/FACTS.
Read the full abstract

All authors Moritz Thoma, Maximilian Groezinger, Maximilian Forstenhäusler, Emad Aghajanzadeh, Ryan Pegoud, Manoj Rohit Vemparala, Pierpaolo Mori, Alexander Frickenstein, Daniel Mueller-Gritschneder, Ulf Schlichtmann

Computer Vision

Ambient @ EgoLongQA 2026: Distilling Long-Video perception into a Sub-2B Model

We describe our entry to the EgoLongQA track of the Wearable-AI Challenge in ECCV 2026, which placed first in the <=2B parameter division with 0.8279 on the held-out test set. Our system is a single 2B vision-language model that answers multiple-choice questions about ten-minute egocentric videos in one greedy forward pass; It is obtained by distilling the junior perception module of a tool-using agentic pipeline, not the agent itself into a small student, using teacher traces filtered to those that answered correctly. it reaches 89% of the accuracy of the large agentic pipeline using 1.1% of its parameters. This raises a 27.1% base model to 81.4% on our held-out questions. The 2B backbone has 2.2132B parameters and therefore over the divisional limit, to make the entry admissable we prune the multilingual embedding table from 248,320 to 143,469 rows, reaching 1.9985B with provably identical logits on retained rows.
Read the full abstract

Computation and Language

FreqBLiMP: Frequency-Controlled Minimal Pairs Reveal Robustness and Fragility of LLMs Under Lexical Rarity

Minimal-pair benchmarks such as BLiMP evaluate linguistic knowledge by testing whether language models (LMs) prefer acceptable sentences over minimally different unacceptable ones. However, these benchmarks largely ignore lexical frequency variation, despite lexical frequency being a pervasive and highly skewed property of natural language use. Consequently, existing evaluations do not test whether grammatical preferences remain stable when contrasts involve rare lexical items. We introduce FreqBLiMP, a frequency-controlled extension of BLiMP that regenerates all 67 paradigms under explicit Zipf-frequency regimes while preserving each minimal-pair's grammatical contrast. Evaluating multiple open-weight LLM families across scales, we find that decreasing lexical frequency produces a consistent, monotonic decrease in sentence likelihood, but only a modest reduction in overall contrastive acceptability accuracy. However, this aggregate stability masks substantial variation across linguistic phenomena, with LLMs remaining robust on overt morphosyntactic generalization while degrading on phenomena that require lemma-specific information.
Read the full abstract

Artificial Intelligence

An Auditable Symbolic-RAG-Generative AI Architecture for Goal-Oriented Conversation Orchestration

Goal-oriented conversational systems must answer factual questions, understand visitor-provided information, and advance business objectives without becoming rigid questionnaires. This paper proposes a Symbolic-RAG-Generative architecture centered on the Goal-oriented Retrieval-Augmented Conversation Engine (GRACE). An instruction-constrained Business Goal Compiler transforms business intent into an immutable objective set, normalized priority vector, canonical questions, and initial state vector. At runtime, GRACE receives the complete conversation history, latest visitor message, current state, and grounded answer generated by a separate RAG component. It updates completion only from visitor-authored evidence and selects one contextually modulated follow-up. The core policy maximizes expected business progress subject to a minimum visitor-utility constraint. We formalize the state, monotonic transitions, source separation, question modulation, and constrained policy; present the reference architecture; and define an evaluation comprising 24 English real-estate and 10 Spanish professional-cleaning conversations, totaling 119 protocol-defined visitor turns. Across both domains, GRACE achieves 84.9% exact state-transition accuracy, 91.6% evidence precision, 89.6% evidence recall, 100% monotonicity, and 94.1% terminal-state accuracy. The evaluation establishes compelling symbolic-state performance across standard, multi-goal, RAG-detour, validation, refusal, and robustness scenarios.
Read the full abstract

Computation and Language

Vishing-Tactics-Bench: Forecasting Exploitation Trajectories in Voice Phishing Calls

Voice phishing (vishing) unfolds in real time; by the time a call has ended and post-hoc classification is possible, the harm has already been done. The more actionable question is which concrete harm (Information Gathering or Financial Exploitation) an ongoing call is tactically progressing toward. We present Vishing-Tactics-Bench, a benchmark grounded in Endsley's situation-awareness (SA) framework that recasts vishing defense from after-the-fact fraud classification to harm projection: predicting at each turn whether the call will reach either terminal harm. We adapt MITRE ATT&CK to vishing as a 6-tactic taxonomy (Vishing-Tactics) and label 35,340 scammer utterances across 5,645 synthetic Chinese calls. We define Exploitation Trajectory Forecasting, a survival-style protocol over the two terminal harms with three metrics: AP@k, C-index, and divergence error. Baselines ranging from a Markov heuristic to fine-tuned LLMs show that the tactical trajectory serves as an interpretable representation of the call's tactical state, supporting harm-specific forecasting, which can then be used for the downstream application of intervention selection; a stratified lead-time analysis at a tight false-alarm budget further identifies at what point in a call the trajectory signal yields early warning.
Read the full abstract

All authors Jeongmin Lee, Dongmyung Sul, Seung Yun, Jinxia Huang

Information Theory

Conditional Fisher-Information Central Limit Theorems under Log-Concavity with Information-Theoretic Consequences

We establish conditional central limit theorems in Fisher information under log-concavity in every fixed dimension. For conditionally centered normalized sums, after whitening by the averaged conditional covariance, the averaged conditional Fisher information converges to the dimension if and only if it is finite at one convolution level. The scalar criterion follows as the one-dimensional case; we also provide an independent scalar proof based on a second-order continuity theorem for Fisher production on Gaussian-smoothed, tail-controlled classes. For the original sums, the averaged Fisher information matrix converges in operator norm to the inverse averaged conditional covariance. Consequently, the conditional relative Fisher information with respect to the limiting Gaussian law vanishes, and the Gaussian logarithmic Sobolev inequality yields convergence in conditional relative entropy and conditional entropy. We give two operational consequences. For any fixed finite-constellation low-power input, the first-order conditional mutual-information slope converges to the Gaussian-noise benchmark. For Gaussian signaling at any fixed signal covariance, the mutual-information gap from that benchmark is bounded by the conditional relative Fisher deficit and hence vanishes asymptotically.
Read the full abstract

Optimization and Control

Mapping locally nondominated curves in multiobjective topology optimization of compliance and volume

This paper studies the nondominated set for topology optimization of compliance and volume. In the multiobjective topology optimization literature, scalarization is a popular method to approximate this set, which is better known as the Pareto frontier. However, previous work indicates that the pointwise approximation obtained by scalarization obfuscates its underlying mathematical structure: under an imposed length scale, it consists of locally nondominated curve segments. To investigate this structure, the paper uses a simple methodology to (i) generate local optima with a tightly controlled topology and (ii) use continuation to extend these points to locally nondominated curves. The paper then performs two rounds of numerical experiments. The first round applies the methodology to several cantilever examples. This leads to novel insights regarding topological complexity and symmetry and confirms prior conclusions regarding continuity, smoothness and convexity. The second round investigates two numerical instabilities: (i) a lack of uniqueness and (ii) discontinuities of the locally nondominated curves. The former is linked to an observed flatness of the optimization landscape, whereas the latter is attributed to branch splitting and linked to bifurcation theory. Namely, we classify these discontinuities as higher-dimensional analogues of the subcritical, symmetry-breaking pitchfork bifurcation. The paper concludes with suggestions for future research to address these features of the multiobjective optimization landscape.
Read the full abstract

Machine Learning

Stable-MM-R1: Anchoring Multimodal Reasoning Dynamics via Entropy-Guided Stratification

While Reinforcement Learning (RL) effectively incentivizes reasoning in Large Language Models, current pipelines are hindered by training instability and rapid entropy collapse. These limitations often stem from "Rollout Silencing" and low-quality gradient signals in standard sampling procedures. In this work, we propose a robust, data-centric framework to stabilize RL training. We first introduce Potential-Aware Query Mining (PAQM), which filters data dynamically to focus on the "Distillation Zone"—samples with high potential for capability elicitation. Furthermore, we present Hybrid Stratified Replay (HSR), a novel mechanism that restructures batches by stratifying rollouts based on Path Entropy, a rollout-level confidence proxy, and outcome reward. Within each optimization step, HSR reuses current-policy "Stability Anchors" and "Hard Negatives" to construct high-contrast optimization groups, then clears its buffers before the next step. This approach mitigates entropy collapse while improving the utilization of learning signals under limited compute. Our method outperforms strong baselines on complex reasoning tasks, offering a principled solution for stable and efficient RL fine-tuning.
Read the full abstract

All authors Yimeng Ye, Shuang Chen, Wenxuan Huang, Manyuan Zhang, Kaituo Feng, Zhangquan Chen, Jiayu Chen, Yucheng Zhou, Yicheng Xiao, Zhiyuan Feng, Tianyu Shi

Machine Learning

Fine-grained Distributed Backdoor Attacks in Federated Learning

Federated learning, as a privacy-preserving distributed machine learning paradigm, faces significant threats from backdoor attacks. Compared to centralized attacks, distributed backdoor attacks are more harmful but require more poisoned samples to compensate for the loss of trigger strength due to decomposition. Fixed trigger patterns are also easily detected by robust aggregation algorithms, increasing the risk of attack exposure. To address these challenges, we propose a fine-grained distributed backdoor attack framework (FDBA). This framework uses dynamic trigger generation and embedding vector optimization to perform attacks with fewer poisoned samples. First, we design a dynamic trigger generation method based on image edge structures using the Canny algorithm to extract edge features, which are then injected with Laplacian noise. RGB channel decomposition is applied for covert adaptation of the distributed trigger, reducing detection chances. Second, we introduce an embedding vector contrastive learning strategy that forces poisoned samples to approach the target class center in the feature space, enhancing attack effectiveness. On CIFAR-10, piecewise-linear estimates for target ASRs between 70% and 90% show that FDBA reduces the required poisoning ratio by 37.4%–48.4% compared with DBA. In non-independent and identically distributed (Non-IID) scenarios, FDBA retains 84.7% of its IID attack performance under extreme heterogeneity, whereas DBA drops to 73.5%, and the framework successfully bypasses mainstream defense mechanisms. This study offers new insights into federated learning security and emphasizes the potential threats and defense challenges posed by fine-grained distributed attacks.
Read the full abstract

All authors Jian Wang, Hong Shen, Wei Ke, Xue Hua Liu

Quantum Gases

Nonequilibrium full counting statistics of multicomponent strongly-interacting quantum gases with defects

We investigate the full counting statistics of a one-dimensional multicomponent impenetrable gas released from a bipartite state in the presence of a local defect. Spin–charge separation reduces the multivariate generating function to an exact Fredholm determinant whose kernel is built from time-evolved single-particle orbitals. Expressing these orbitals in terms of scattering data gives access to both transient dynamics and the nonequilibrium steady state. We derive the leading long-time asymptotics and identify persistent oscillations in densities and currents caused by multiple bound states at the junction. We argue that interactions between components lead to a linear low-temperature correction to the current, in contrast to the quadratic correction in the single-component case. At equilibrium, we obtain the M-Wright distribution of spin transfer with $t^{1/4}$ scaling.
Read the full abstract

Robotics

EquiGQNet: Fast Grasp Quality Evaluation via Shared Equivariant Point Cloud Encoding

Planning six-degree-of-freedom (6-DoF) grasps for unseen objects in cluttered tabletop scenes from a single-view depth image requires accurate and efficient evaluation of diverse grasp candidates. Existing early-fusion methods capture local object geometry relative to each grasp candidate but repeatedly encode the scene, whereas late-fusion methods reuse a shared scene representation but may lose this grasp-relative local geometry. We propose EquiGQNet, an efficient 6-DoF grasp quality evaluator that combines the strengths of both approaches. For grasp orientation, EquiGQNet replaces the early-fusion operation of rotating and re-encoding the point cloud for each grasp candidate with an SO(3)-equivariant encode-once-then-rotate scheme, yielding grasp-aligned geometric features from a shared scene encoding. For grasp translation, Mid-level Action Fusion (MAF) injects the grasp position into intermediate features before global aggregation, retaining local geometry relative to each candidate. We evaluate EquiGQNet in two grasp planning pipelines: Cross-Entropy Method (CEM)-based continuous grasp search and candidate ranking with a pretrained generative planner. In simulation, EquiGQNet achieves grasping performance comparable to the early-fusion baseline and substantially outperforms late fusion on objects with complex geometry and limited graspable regions, while reducing CEM planning time from 3.31s to 0.48s, a 6.9x speedup over early fusion. In real-world household-object decluttering, EquiGQNet achieves a 95.2% grasp success rate and 230 picks per hour, versus 153 and 170 for early- and late-fusion baselines. Code is available at https://equigqnet.github.io/.
Read the full abstract

All authors Sungwon Seo, Jaeseog Won, Jiyou Shin, Youngjin Seo, Hyunjun Kim, Seokmin Yoon, Tuan Luong, Hyungpil Moon

Biological Physics

Self-Organized Dynamic Reactors: Product-Condensate Feedback Accelerates Reactions

Condensates can accelerate reactions by concentrating reactants, yet product accumulation can eliminate this advantage. We show that product-induced suppression of condensation generates rotating, deforming, and propagating condensate domains through spontaneous symmetry breaking. Their reorganization shifts reaction sites away from accumulated products while retaining reactant enrichment, yielding rates above those of static condensates and homogeneous states. Reaction flows can thus self-organize condensates into dynamic reactors, with implications for biological and synthetic systems.
Read the full abstract

Information Retrieval

EAGER: Enrich-and-Align Generative Query Recommendation from Clicked Items in E-commerce Search

E-commerce platforms increasingly display clickable query suggestions alongside items in the user feed, enabling users to refine or expand their intent without manually reformulating queries. Existing approaches either mine suggestions from historical logs – limited to past behavior and blind to long-tail, personalized intents – or rely on off-the-shelf LLMs whose lack of platform-specific knowledge yields fluent but generic queries disconnected from real click behavior. We propose EAGER (Enrich-and-AliGn gEnerative Query Recommendation), a two-stage framework for generating query suggestions from clicked items. In the enrichment stage, supervised fine-tuning (SFT) follows a four-stage curriculum that scales information richness (from item-only to user-conditioned) and reasoning depth (from direct to chain-of-thought). Each stage incorporates rationale augmentation, diversity regularization, and self-distillation. In the alignment stage, we post-train via GRPO with a hybrid reward of multiple rule-based business signals and a preference-aware click reward. Extensive offline experiments and online A/B test demonstrate the effectiveness of EAGER, which has been deployed in production at a major e-commerce platform.
Read the full abstract

All authors Shuwei Yuan, Mingqian Ding, Luxin Liu, Rong Xiao, Xiaoyi Zeng

Optics

Controllable ultrabroadband supercontinuum during twocolor femtosecond laser filamentation in high pressure gases

Twocolor laser filamentation is considered a promising strategy for deep transformations of the spectrum and generating ultrashort light pulses, which is critically important for developing effective methods to control broadband coherent radiation in attosecond physics and ultrafast spectroscopy. We present the results of our experiments on the study of the spectral dynamics of supercontinuum generated during collinear twocolor filamentation of femtosecond pulses at the fundamental (800 nm) and second harmonic (400 nm) frequencies in Ar, N2, and CO2 gases at pressures up to 11 atm. We demonstrate that the transition to multiple filamentation in high pressure gases qualitatively changes the interaction dynamics between the color components. We demonstrate that instead of a simple merging of spectra broadened during filamentation, a complex, nonmonotonic dependence of the shape of the twocolor supercontinuum on the time delay between pulses is observed, including selective suppression of radiation in the 400 nm region. For the first time, significant differences are identified between the dynamics of twocolor filamentation in atomic and molecular gases. Particularly, in CO2, spectral transformation is preserved at picosecond interpulse delays, which is explained by the strong contribution of molecular rotational wave packets to the effective nonlinear polarizability. Our results demonstrate the possibility of controlled redistribution of spectral energy and dynamic switching off of individual regions of supercontinuum in atomic and molecular gases as their pressure changes, which fills the gap in understanding the spatiotemporal evolution of multiple two color filaments and paves the way for the controlled formation of ultrashort laser pulses.
Read the full abstract

Social and Information Networks

How Well Do LLMs Simulate Survey Responses Following a Breast Cancer Screening Intervention?

Collecting survey data is laborious and limited by privacy constraints. Large language models (LLMs) have shown promise as predictive social simulations. It is unclear whether they can replicate population-level response distributions before and after a healthcare intervention. Using information derived from 4125 women aged 35-59 years, we evaluate whether agents informed solely by pre-intervention profile information can reproduce post-intervention response distributions. Groups of LLM agents (n=50) were created with Gemma 4 E4B and Qwen3.5 9B; conditions ranged from zero-shot prompting to agent profiles enriched with aggregate or individual-level demographic characteristics and pre-intervention questionnaire responses. We compared predicted and observed response distributions with Total Variation Distance (TVD) and Normalized Wasserstein Distance (NWD). Across both LLMs, profile-based agents improved distributional accuracy relative to zero-shot and random baselines. Nevertheless, direct sampling of 50 real participants remained more accurate. Prediction errors were also higher among participants aged 55-59 years and those living in private property. Errors also varied by question theme and LLM model, with the highest errors observed for cancer fatalism and post intervention attitudes toward genetics. Sensitivity analyses showed that performance was influenced by prompt template changes and temperature hyperparameter. Our results show the potential of LLM-based agents to model behavioral responses to interventions in silico. However, profiles containing additional information beyond demographics did not consistently outperform simpler ones. Certain cultural constructs and population groups also remain inadequately represented by the LLM models evaluated. Future work may include building behaviorally grounded and locally validated virtual populations.
Read the full abstract

All authors Kenneth Koh, Ryan Jak Yang Lim, Alessandro Sparacio, Peh Joo Ho, Mile Sikic, Borame L Dickens, Mikael Hartman, Jingmei Li

Statistical Mechanics

Probing magnetic correlations in space and time within predefined topological sectors of a macroscopic spin liquid

The triangular Ising antiferromagnet, with its residual entropy density and critical correlations at absolute zero, is the archetype of a two-dimensional spin liquid. Its ground state is partitioned into topological sectors connected by global spin flip events that wrap the lattice boundaries, and become statistically irrelevant in the thermodynamic limit where magnetic fluctuations are restricted to the dominant sector. Thus far, these properties have been mainly investigated from a theoretical perspective, and one might wonder to what extent they can be transposed to real materials. Here, we present experimental observations obtained in an artificial, macroscopic realisation of the seminal triangular Ising antiferromagnet that consists of a lattice of millimeter-sized NdFeB cylinders put into motion by a mechanical shaker. Specifically, we demonstrate that the very-low-energy physics and the true ground state of this model can be reached experimentally. Besides, we are able to probe, in space and time, the magnetic properties within manually preselected pockets of the ground-state manifold that emulate, to a good approximation, the behaviour in distinct topological sectors. Our approach opens new avenues for naked-eye visualisation and hand manipulation of many-body phenomena associated to frustrated magnetism and models of statistical physics.
Read the full abstract

All authors Rémy Dangoisse, Jeanne Colbois, Laurent Del Rey, Nicolas Rougemaille, Johann Coraux

Artificial Intelligence

Encoded Early, Used Late: Where Transformers Begin to Act on an Inferred Partner's Expertise

A transformer can make an attribute linearly decodable in its residual stream at a depth where that attribute does not yet influence the output. This gap between where information is readable and where it is used has been shown for attributes stated directly in the input. We ask whether it also holds for an attribute the model must infer gradually over a conversation, namely how expert its dialogue partner is. Using ExpertCollab, a corpus of multi-turn research-planning dialogues between model-played personas at four expertise levels, we find that partner expertise is most decodable in the early layers and falls to near chance before the midpoint of the network. Counterfactual patching shows that injecting the expertise difference at the layer of peak decodability barely changes a fixed late-layer readout, whereas the same difference injected past the midpoint propagates almost completely, a separation of more than an order of magnitude. A content-matched random control and a probe-free diagnostic place the transition at the same early layer, and a statically specified control attribute stays decodable throughout. An inferred relational attribute is therefore represented well before it becomes causally active, which bounds where any attempt to read out or steer partner-conditioned behavior must intervene. We use one model on a synthetic corpus as an initial demonstration.
Read the full abstract

High Energy Physics (Phenomenology)

A Dark-Dimension Origin of Geometric Inelastic Dark Matter: The LUX-ZEPLIN High-Recoil Event and Multi-Target Tests

The extended nuclear-recoil analysis of LUX-ZEPLIN (LZ) has reported one event compatible with a recoil energy near $248$ keV, motivating endothermic dark-matter scenarios with support at large recoil energy. We study a pseudo-Dirac electroweak-doublet dark-matter model embedded in a five-dimensional Dark-Dimension framework. The Standard Model and the vectorlike electroweak doublets are localized on the visible brane, while dark-number violation is communicated through a neutral bulk state. The resulting mass splitting is exponentially suppressed, $\delta=\delta_{\rm UV}e^{-2\pi MR}$, providing a geometric origin for the few-hundred-keV scale relevant to inelastic scattering. The neutral weak current is dominantly off diagonal, fixing the neutron-scale normalization at $\sigma_{\chi n}^{Z}\simeq1.86\times10^{-39}\,\mathrm{cm^2}$. Using a two-bin Poisson diagnostic of the extended LZ recoil window, we find a shallow minimum near $(m_\chi,\delta)\simeq(1.35~\TeV,366~\keV)$, while the thermally motivated point $(1.1~\TeV,360~\keV)$ lies only $\Delta\chi^2\simeq0.21$ higher. The same benchmark is kinematically inaccessible to light targets but remains open for xenon and tungsten, providing a characteristic multi-target test of the model. For CaWO$_4$ we obtain an integrated tungsten rate of $3.15\times10^{-2}$ events kg$^{-1}$ yr$^{-1}$ in the $95$–$150$ keV interval. These results connect the geometric origin of the pseudo-Dirac splitting with high-recoil direct-detection phenomenology.
Read the full abstract

Computer Vision

Flow3D-OPD: Multi-Teacher On-Policy Distillation for 3D Geometry Generation with Flow-Matching Diffusion Transformer

Recent image-to-3D generation models built on flow-matching diffusion Transformers (DiT) can produce high-fidelity meshes, yet their post-training strategy remains largely unexplored. There exist several critical bottlenecks in reinforcement learning: the inherent difficulty of defining comprehensive rewards for 3D geometric quality, and the gradient interference that arises when jointly optimizing heterogeneous objectives. Inspired by the practicability of on-policy distillation (OPD) in large language models and image generation, we propose Flow3D-OPD, a two-stage post-training framework that introduces multi-teacher distillation into 3D geometry generation. In the first stage, we utilize the semi-policy to enhance the foundational capability of the pretrained model and then design an agentic verifier for 3D geometric quality evaluation. Based on the verifier, we could cultivate domain-specialized teacher models via direct preference optimization (DPO). In the second stage, we consolidate heterogeneous expertise into a unified student model through on-policy distillation with hard task-routing sampling and gradient accumulation, which could mitigate the gradient interference in joint optimization. Without relying on elaborate modifications, our straightforward yet effective design achieves consistent improvements across all geometric quality dimensions and surpasses all teacher models in the average metric. Extensive experiments demonstrate that our approach provides an effective paradigm for reinforcement learning in 3D generation.
Read the full abstract

All authors Zhiwei Ning, Zhen Zhou, Puhua Jiang, Xintong Han, Gengming Zhang, Jie Yang, Zhonglong Zheng, Yuanjie Zheng, Wei Liu, Chunchao Guo

Game Theory

The Fine-Grained Complexity of Approximate Nash Equilibrium and Free Games

We study the fine-grained complexity of computing approximate Nash equilibria and approximating the value of free games in the regime where the approximation error vanishes. Under the PCP for PPAD and ETH for PPAD conjectures, we show that computing $\varepsilon$-approximate Nash equilibria in 2-player $N$-action normal-form games requires time $N^{(\log(N)/\varepsilon^2)^{1-o(1)}}$, thus showing that the classical Lipton-Markakis-Mehta algorithm (2003) is optimal through all regimes of $\varepsilon = \omega(1/\sqrt{N})$. While such optimality was known in the constant-$\varepsilon$ regime (Rubinstein, 2016), previous work could only rule out significantly smaller running times of $N^{O(\log(N)/\varepsilon)}$ in the regime $\varepsilon = o(1)$. Using similar techniques, we then establish an analogous tight lower bound of $N^{(\log(N)/\varepsilon^2)^{1-o(1)}}$ under ETH for $\varepsilon$-additive value estimation in free games, when $\varepsilon \geq 2^{-o(\sqrt{\log N})}$, answering a question of Aaronson, Impagliazzo, and Moshkovitz (2014).
Read the full abstract

Computer Vision

NutriBench-Kitchen: Benchmarking Embodied AI for Nutrition Management

An embodied kitchen assistant must do more than recognize food in isolated frames. It must track ingredient states over time and integrate visual observations with recipe and nutritional knowledge to support constraint-aware decision-making. We formalize this capability as Embodied Nutrition Management: perceiving nutrition-relevant events, maintaining a persistent food state, and using it for knowledge-grounded planning. Existing benchmarks evaluate static food understanding or embodied cooking actions, but do not measure whether an agent can continuously update and use nutrition-relevant states in dynamic kitchens. To fill this gap, we introduce NutriBench-Kitchen, a benchmark containing 1,500 manually verified question–answer pairs from 160 cooking videos. It covers five task families: Ingredient Entry, Memory Management, Recipe Query, Long-Term Planning, and Short-Term Planning, spanning food-state construction, maintenance, knowledge retrieval, and decision-making across different planning horizons. Evaluations of proprietary and open-source large vision-language models reveal a substantial gap from human performance, particularly in quantitative ingredient estimation, long-term state tracking, and reasoning under interacting constraints. We further introduce Nutri-Vgent, a diagnostic long-video agent with separate episodic, food-state, and recipe memories. Its consistent improvements demonstrate the value of explicit state representations and structured memory for nutrition management. Together, NutriBench-Kitchen and Nutri-Vgent provide a testbed for studying persistent state tracking and knowledge-grounded reasoning in dynamic kitchens. Code is available at https://github.com/V1ol1n/NutriBench-Kitchen.
Read the full abstract

All authors Yulin Wei, Xiangchen Wang, Jianhui Pan, Jinyu Xiao, Zheng Tan, Ruozai Tian, Guanhua Chen, Feng Zheng

Computation and Language

Retrieval-Augmented Multi-Prompt Ensemble for Minor-Grain Breeding Information Extraction

This paper presents our system for CCL2026-Eval Task 5: Minor-Grain Breeding Information Extraction (MGBIE), which jointly extracts 12 entity types and 6 relation types from minor-grain breeding literature. We propose RAME (Retrieval-Augmented Multi-Prompt Ensemble), a training-free framework that elicits multiple LLM outputs under controlled diversity and aggregates them by majority voting to obtain high-confidence predictions. RAME combines (i) retrieval-augmented few-shot selection via a hybrid BM25-embedding retriever, (ii) a three-prompt ensemble (Strict, Relaxed, Balanced) spanning the precision to recall spectrum, and (iii) large-scale repeated sampling with majority voting to filter noisy predictions. Built on DeepSeek-V4-Flash, RAME achieves a Total Score of 0.499 (NER 0.730, RE 0.346) on the leaderboard, ranking 1st and surpassing the official Track-A baseline powered by GPT-5.5 (0.448), representing an 11.4% relative improvement. Code is available at https://github.com/king-wang123/CCL26-RAME.
Read the full abstract

Number Theory

Counting Representations of Quadratic Forms

We answer a question of C. S. Herz about the number of integral $k\times m$ matrices $T$ such that $T^tT\leq R$ for a fixed positive definite $m\times m$ matrix $R$. We give an asymptotic formula for the count. This can be seen as a generalization of the Gauss circle problem, as counting representation of quadratic forms by the identity $I_k$, and as counting integral points bounded by the Stiefel manifold $T^tT=R$. The main tool is a bound for Bessel functions of matrix argument that was proved over Jordan algebras.
Read the full abstract

Functional Analysis

A Levin-Milman theorem for Young variation

In 1940, Levin\footnotemark[1] and Milman proved that a closed linear subspace of $C[0,1]$ whose elements all have bounded Jordan variation must be finite-dimensional. We prove its analogue for variation in the sense of Young and, more generally, for every finite-valued nondecreasing gauge $\varphi:[0,\infty)\to[0,\infty)$ with $\varphi(0)=0<\varphi(t)$ for $t>0$. If $E$ is a closed linear subspace of $C[0,1]$ and every $f\in E$ satisfies $\Var_\varphi(\lambda_f f)<\infty$ at some scale $\lambda_f>0$, then $E$ is finite-dimensional. No continuity, convexity, or doubling condition is needed. The proof combines a lower-semicontinuous regularization that preserves the scaled class, Baire uniformization, Helly selection, and a quantitative nested-peaks construction for nonhomogeneous gauges. Consequently, for every infinite-dimensional closed subspace $F\subset C[0,1]$, the set $F\cap\cV_\varphi([0,1])$ is a meagre $F_\sigma$ subset of $F$. For every Young function, both the scaled and raw finite-variation families are maximal dense-lineable but not spaceable. The scaled family is itself a dense vector subspace of Hamel dimension $\mathfrak c$, whereas without local doubling the raw family need not itself be linear.
Read the full abstract

All authors N. G. Albuquerque, D. Bugajewska, E. Demétrio Jr., J.B. Seoane-Sepúlveda