Multiagent Systems
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Thursday 10 September
Clinical education must prepare medical students to conduct safe and coherent patient interviews under conditions of uncertainty. Traditional standardized patient (SP) training is resource-intensive and difficult to scale. We developed a scaffolding-oriented multi-agent Large Language Model (LLM) AI Standardized Patient (AI-SP) training platform1. The system includes a patient agent for simulated dialog, a tutor agent providing Socratic prompts without disclosing diagnostic information, and a turn-level evaluator agent that monitors clinical progress without revealing summative scores. In a randomized controlled study (N = 100 medical students), participants were assigned to either a multi-agent (MA) scaffolding condition or a control condition. All students completed two learning sessions under their assigned condition followed by an examination conducted in a patient only environment. Performance was assessed using a standardized Objective Structured Clinical Examination (OSCE) based rubric. While no significant difference was observed in final diagnostic accuracy between groups, the multi-agent AI standardized patient system improved final examination scores compared to the control group utilizing structured progressive information disclosure; the most substantial and consistent improvements were observed in communication, the expression of empathy, and specific history-taking behaviors. These findings suggest that specialized LLM agents enhance the process quality of simulated clinical interviews without artificially inflating examination outcomes. To support future research, we release a multi-expert annotated dataset comprising transcripts, checklist annotations, turn-level evaluations, and OSCE-aligned scoring outcomes. This resource aims to facilitate the development of pedagogically grounded AI-SP systems and advance research on AI-supported clinical reasoning training.
Optimization and Control
Deterministic and Random Bipartite Matching on General Networks: Convex Flow Reformulation, Asymptotic Properties, and Fast Algorithms
Minimum-distance bipartite matching on general networks has numerous applications various fields. This paper first focuses on deterministic problems and presents an exact edgewise-separable convex-flow reformulation. By introducing a smooth monotone rearrangement approximation of the edge-wise imbalance profiles, the convex-flow reformulation's can be solved efficiently. If we further conduct a first-order resistance-based approximation of the convex program, a one-step Laplacian-based estimator can be analytically derived in closed forms. The paper also studies random problems where supply and demand points are randomly distributed. We show that the expected optimal matching distance scales with the square root of the number of points if the supply/demand point distributions are identical, or linearly otherwise. In the former case, the optimal flow is proven to be centered, symmetric, and sub-Gaussian. In the latter case, the limiting resistance network characterizes how supply-demand imbalance is redistributed and motivates a fast algorithm that approximate the optimal flow based on the limiting resistance. Numerical experiments show that the proposed estimators closely approximate the exact matching cost while substantially reducing computation time. The proven theoretical properties of the random matching solution are numerically verified by large-scale Monte Carlo simulations.
Read the full abstract
Mathematical Software
The Art of Closed-Formula Defaults: Search-Free Code Generation for Tensor Operators
Agentic search and automated optimization of GPU kernels are powerful tools for large language model inference. Their effectiveness, however, depends not on the sophistication of the search itself, but on the clarity of the optimization problem being solved. We provide an application-first approach that drives a hierarchical code generation tool from operator specifi cation down to GPU instructions, and show that a clearly defined computational model makes the optimization problem tractable.
Read the full abstract
Analysis of PDEs
Well-posedness of the two-dimensional unsteady Prandtl system in Sobolev space with degenerate critical points
This paper is devoted to the well-posedness of classical Prandtl equations in a finite order Sobolev space. For a initial data with degenerate critical points and general outflow, we obtain the local-in-time existence and uniqueness of the solution to the Prandtl equations in a Sobolev space, by introducing a new iteration scheme and linear cancelation. This result shows that Oleinik's monotonicity condition is not a necessary condition for the Prandtl equations to be well-posed in Sobolev spaces and provides evidence to demonstrate that zero shear stress does not necessarily lead to boundary layer separation in two-dimensional unsteady boundary layers.
Read the full abstract
Cryptography and Security
Empirical Evaluation of Membership Inference Attacks on NLP Text Classifiers: A Baseline Study on SST-2
Membership inference attacks (MIAs) try to determine whether a specific record was used to train a model, a privacy risk that matters in natural language processing (NLP), where training data can contain sensitive user text. This paper presents a controlled benchmark of membership inference vulnerability for text classification on the GLUE SST-2 sentiment dataset. A TF-IDF + Logistic Regression pipeline and a fine-tuned DistilBERT classifier are compared under a loss-threshold MIA, with utility measured by development accuracy and macro F1. DistilBERT reached 0.9466 accuracy and 0.9460 macro F1 against 0.8756 and 0.8727 for Logistic Regression, yet both models leaked membership signal (Attack AUC 0.5615 and 0.5800, respectively). Two mitigations were tested. Stronger regularization reduced leakage for Logistic Regression at a visible utility cost, whereas fine-tuning DistilBERT for 2 epochs instead of 3 reduced leakage with negligible accuracy loss. Lightweight training adjustments can improve the privacy-utility trade-off without complex defenses.
Read the full abstract
Computation and Language
Using Semantic Uncertainty to Estimate Transition Relevance in Turn-taking
Turn-taking is a fundamental mechanism that governs when interlocutors speak and listen. Although Spoken Dialogue Systems (SDS) exploit a range of linguistic, acoustic, and non-verbal cues, they produce ill-timed responses in unscripted interaction. A central challenge is anticipating Transition Relevance Places (TRPs), or opportunities, not obligations, for a listener to take the floor. Human listeners do not wait for turn endings; as an utterance unfolds, they use expectations about its developing meaning to anticipate TRPs and decide whether to take the floor. We examine whether these evolving expectations can be modeled through semantic uncertainty – an LLM-derived measure of how strongly a turn so far constrains what may plausibly come next. To do so, we sample possible continuations of an ongoing turn and use changes in semantic dispersion to identify TRPs within turns. We evaluate this account on a dataset with TRP labels derived from real-time listener responses, rather than retrospective annotation. Our approach substantially outperforms prompt-based and fine-tuned text-only baselines, providing empirical support for the view that evolving semantic constraints inform perceived turn-taking opportunities in unscripted interaction.
Read the full abstract
Statistical Methodology
Generalized Ridge Refitting for the Lasso and Prediction Improvement Bounds
We study a class of Lasso based estimators obtained by applying a quadratic correction on the Lasso equicorrelation set. The penalty matrix determines both the magnitude and geometry of the correction and contains, among other cases, the isotropic Lasso–Ridge correction, least squares refitting, Gram proportional interpolation between the Lasso and least squares, and coordinate specific penalties. We first derive a closed form representation and isolate the positive gain component of the resulting prediction improvement. We then control the remaining stochastic linear term in expectation by localizing the random signed equicorrelation model around a deterministic reference support. This yields a finite sample expectation bound that explicitly accounts for the randomness induced by Lasso model selection. The resulting decomposition provides a unified framework for understanding when Lasso based quadratic corrections can improve prediction.
Read the full abstract
Quantum Physics
Distributed variational quantum computing with deterministic entanglement tuning
Distributed quantum computing offers a scalable route to quantum information processing by entangling spatially separated processors. Although gate teleportation enables universal computation across distributed nodes, it requires repeated consumption of high-fidelity Bell pairs, ancillary qubits, and real-time feedforward, which imposes significant overhead and reduces fidelity. However, many variational quantum algorithms do not demand full universality; rather, they rely on sufficient expressibility to explore solution spaces effectively. Building on this, we propose a distributed variational quantum computing protocol based on deterministic entanglement tuning. In contrast to probabilistic filtering, our approach deterministically modulates pre-shared entanglement using only local operations and classical communication. We validate the protocol through a proof-of-principle experiment by estimating ground-state energies of the He-H$^+$ molecule and the Schwinger model, showing that diverse entanglement levels can be engineered to match problem-specific requirements. Our results suggest a practical alternative for near-term distributed quantum applications.
Read the full abstract
High Energy Astrophysical Phenomena
IceCube neutrino point-source searches in the direction of the KM3NeT ultra-high-energy event
While still under construction, the KM3NeT Astroparticle Research with Cosmics in the Abyss (ARCA) detector recorded a $\sim$200 PeV neutrino on February 13th, 2023. This event is the highest-energy neutrino reported. IceCube, a cubic kilometer neutrino detector located at the geographic South Pole, has previously detected neutrinos up to approximately 10 PeV. We search for high-energy neutrinos from the location of the KM3NeT event using 15 years of IceCube data and considering three temporal hypotheses: steady or flaring in time coincidence, or at an arbitrary time. We find no evidence for neutrino emission for any of the studies performed. Correspondingly, we set upper limits on the neutrino flux from a point source in the direction of KM3-230213A. We compare these limits to KM3NeT's estimated flux and show that an astrophysical explanation of this event is strongly constrained for a variety of spectral assumptions for a steady or transient point source with the flux inferred from the single KM3NeT ultra-high-energy event assuming a spectral index of 2.0.
Read the full abstract
Materials Science
Hyperspin Altermagnets
The behavior of spin quantum in k-space is key to identifying altermagnets (AMs) as the third kind of fundamental collinear magnetism. In contrast, non-collinear magnets,though abundant in nature,lack well-defined spin quantum numbers, and the resulting spin textures are often highly complex, which limits their potential for next-generation spintronic applications. Here we propose hyperspin, which lives in a higher-dimensional space, to address these drawbacks. Through analyzing the commutation relations between spin and Hamiltonian for a class of non-collinear magnets, we reveal it is a hyperspin, rather than the usual spin, that commutes with Hamiltonian. Unexpectedly, these non-collinear magnets should also show collinear spin-split bands in k-space like collinear AMs. We therefore classify such non-collinear magnets as hyperspin altermagnets (HAMs), as opposed to the usual collinear AMs. Our theory elucidates the fundamental physics of AMs and HAMs and provides a framework for exploring the wide range of non-collinear magnets that may possess other kinds of conserved quantities.
Read the full abstract
Quantum Physics
Probing the Error-Mitigation Threshold with Matrix Product States
Quantum error mitigation relies on accurate noise characterization, but mismatches between the actual and characterized noise can be amplified and drive a sharp threshold between successful and failed mitigation. In random circuits, this threshold maps onto a random-field Ising transition, but previous exact numerics were limited to small one-dimensional and all-to-all systems, leaving explicit two-dimensional architectures unresolved. We develop a fixed-bond-dimension matrix-product-state method for the replicated transfer dynamics that extends threshold calculations beyond exact propagation while retaining the finite-size signatures of the transition. At system sizes beyond previous exact studies, we recover the predicted absence of a threshold for quenched disorder in 1D, obtain a sharper annealed all-to-all critical point, and resolve architecture-dependent finite-depth thresholds in 2D square and heavy-hex circuits. These results establish replicated tensor-network dynamics as a practical tool for probing error-mitigation thresholds in large and higher-dimensional noisy circuits.
Read the full abstract
Machine Learning
AUC Maximization from Biased Positive-unlabeled Data with Confidence
Maximizing the area under the receiver operating characteristic curve (AUC) is a standard approach to imbalanced binary classification. Although positive and negative data are required for maximizing the AUC, negative data are often difficult to collect in some real-world applications due to privacy concerns or the need for specialized expertise to annotate them. Thus, AUC maximization from positive and unlabeled (PU) data has been attracting attention. Existing methods assume that labeled positive data are unbiased samples from the true positive distribution. However, this ideal assumption is often violated in practice. In this paper, we propose a method to maximize the AUC from biased PU data. To address the bias, our key idea is to exploit confidence, i.e., the probability that an instance is positive, associated with the small number of labeled positive data. We derive an estimator of the AUC risk using biased PU data with confidence, enabling AUC maximization under such bias. We further show that the rewritten AUC risk induces a Bayes-optimal AUC ranking even when the available confidence is any strictly increasing transformation of the true posterior probability. We experimentally show the effectiveness of our method on eight real-world datasets.
Read the full abstract
Statistical Methodology
Surprise Reduction and Nullification in Bayesian and Inverse Bayesian Inference under Ambiguous Prediction-Error Attribution
In non-stationary environments, prediction errors may signal environmental change or transient outliers, and adaptive systems must track such changes without overreacting to outliers. We distinguish surprise reduction, which updates beliefs to fit observations, from surprise nullification, which weakens constraints imposed by the predictive structure, and formalize both within Bayesian and inverse Bayesian (BIB) inference. Belief and likelihood updates are derived from variational objectives sharing a nullification strength, determined endogenously by minimizing surprise under the candidate post-update predictive distribution. In the Gaussian case, nullification expands belief and likelihood variances by a common factor relative to standard Bayesian updating, leaving the ratio unchanged. BIB thus defers attribution of the prediction error, committing to neither latent-state change nor observation-process uncertainty. The nullification strength is carried over as a candidate and is maintained or released according to the predictive surprise of the next observation. In a mean estimation task with outliers and changepoints, no scanned parameter setting of a Sage-Husa-type adaptive Kalman filter, fixed-strength BIB variant, or belief-forgetting-only variant outperforms BIB in both changepoint tracking and post-outlier stability. An oracle-informed reduced Bayesian model tracks changepoints better but is less stable after outliers. Although BIB maintains no explicit hypotheses about changepoints or outliers, it generates event-dependent dynamics. The learning rate increases after changepoints, whereas after outliers, nullification is released, and this increase is suppressed. Deferring attribution and letting subsequent observations differentiate the responses may constitute a principle of adaptive inference in non-stationary environments.
Read the full abstract
Strongly Correlated Electrons
Competing Interlayer Loop Currents and Superconductivity in the Bilayer $t$-$J_\perp$-$V$ Model
The recent discovery of high-$T_c$ superconductivity in pressurized and thin-film bilayer nickelates, featuring a strong interlayer exchange coupling, and their potential similarities with cuprate superconductors, has made this a very active topic in condensed matter physics. In the present paper we study the strongly correlated one-orbital ($d_{x^2-y^2}$) bilayer $t$-$J_\perp$-$V$ model for nickelates, where $V$ denotes the Coulomb interactions, using a controlled large-$N$ expansion at and beyond the mean-field level. Focusing on the out-of-plane spin exchange interaction ($J_\perp$), we find that it triggers both out-of-plane $s$-wave superconductivity and an out-of-plane bond-order phase ($z$-BOP) instability. The $z$-BOP gives rise to a complex $z$-axis hopping dominated by its imaginary component, which drives out-of-plane currents and induces in-plane ones, spontaneously forming on the vertical plaquettes a loop-current state that breaks time-reversal symmetry. Competition between this loop-current phase and superconductivity yields a dome-shaped superconducting region, with optimal superconductivity occurring near the $z$-BOP quantum critical point. The resulting phase diagram features a pure loop-current region, a low-doping coexistence phase, a pure superconducting state at higher doping, and a correlated metallic state.
Read the full abstract
Econometrics
Identification in Linear Quantile Panel Models
This paper studies identification in linear quantile panel models with unrestricted individual heterogeneity when the number of time periods is fixed and small. We impose strict exogeneity, whereby the conditional quantile restriction holds given the individual's complete regressor history and latent individual effect, but otherwise allow the disturbances to be arbitrarily dependent over time.
Read the full abstract
Mathematical Physics
Why we should condition denoising diffusion generative models on windows of past observations
Data assimilation (DA) is, traditionally, a cycling process that relies on time-dependent priors to propagate information from past observations to future cycles. Using denoising diffusion generative modeling for DA is challenging because standard approaches use a fixed training data set, which in turn leads to a static prior that ignores information from past observations. Because past observations are ignored, DA systems with static priors lead to larger posterior errors than cycling DA systems. Incorporating time-dependent priors into generative models, however, requires expensive and frequent retraining. Motivated by linear systems theory - where the dependence of a prediction of a Kalman filter on past observations decays exponentially - we condition diffusion models on short windows of past observations. Specifically, we describe training procedures for two frameworks: a diffusion DA system predicting the current state given a set of past observations, and a diffusion “direct observation prediction” (DOP) system, predicting future observations given a set of past observations. Using a canonical linear system, we show that both systems can achieve the minimal posterior error characteristic of a fully-cycled DA/DOP system, without re-training, provided the time windows are long enough. The linear setup ensures analytical tractability, avoids confounding neural network training errors, and confirms that conditioning on windows of past observations is required for efficient and accurate diffusion-based DA or DOP.
Read the full abstract
Computation and Language
Structurally Speaking: Motif-Oriented Graph Captioning through Bidirectional Graph-Text Translation
Graph captions should help readers understand graph structure, rather than simply translate adjacency matrices into long textual edge lists. A useful graph caption abstracts connectivity into recognizable motifs, such as hubs, paths, cycles, cliques, and bridges, because these motifs provide compact structural units that are easier to read, compare, and recover. In this paper, we study motif-oriented graph captioning as a bidirectional graph-text translation task, where captions must both preserve enough topology for graph recovery and express the graph through concise motif-level descriptions. We show that direct prompting of GPT-5.1 often produces graph-recoverable captions by enumerating node-to-node connections, but these captions are verbose and can contain inconsistent motif interpretations. To address this gap, we introduce Structurally Speaking, a lightweight structured prompting protocol that guides translation between explicit connectivity and motif-level abstraction. Experiments on a synthetic motif-based dataset show that structured prompting produces shorter and more motif-consistent captions while maintaining comparable graph recovery. These results suggest that explicit topology-to-motif reasoning guidance can make LLM-generated graph captions more interpretable without model fine-tuning.
Read the full abstract
Computation and Language
Auto-RecSys: Harnessing Autonomous Research Agents for Industry-Scale Recommender System
Auto-research agents have shown the potential to automate hypothesis generation, experiment execution, and iterative refinement. However, scaling this paradigm to industry-scale recommendation models introduces two challenges: (1) long feedback loops, where model training can take days, making serial iteration prohibitively slow and requiring parallel exploration across multiple research directions; and (2) system complexity, where large configurations, fragile infrastructure dependencies, and multi-day GPU jobs require robust and recoverable execution. We present Auto-RecSys, an autonomous research system for long-horizon experimentation on industry-scale recommendation models. Auto-RecSys addresses these challenges through three harness designs: (1) distributed asynchronous execution for running multiple experiments in parallel across servers, (2) centralized cross-server memory for persistent and recoverable execution across sessions and failures, and (3) cognitive-procedural separation, where natural-language skill files guide LLM reasoning while deterministic scripts enforce operational correctness. Auto-RecSys further employs a dual-loop self-evolving architecture: an Execution Evolution Loop in which model-specific playbooks accumulate operational knowledge by recording failed attempts and crystallizing successful pipelines, and an Idea Evolution Loop in which experimental outcomes inform subsequent ideation. Evaluated on recommendation models, Auto-RecSys significantly reduces the human time required per experiment cycle and improves execution reliability as its playbooks mature.
Read the full abstract
Data Structures and Algorithms
Quasi-Monte Carlo Beyond Hardy-Krause II: $(1 + \varepsilon)n$ Samples Suffice
Numerical integration studies how well one can estimate the integral of a function $f$ over $[0,1)^d$ using $n$ sample points. The two classical methods, Monte Carlo (MC) and quasi-Monte Carlo (QMC), have complementary strengths and weaknesses, and a fundamental question is to design an approach that combines the benefits of both. Recently, building on the transference principle in discrepancy theory, Bansal and Jiang~\cite{BJ25a} gave a randomized QMC method that bridges MC and QMC guarantees using only i.i.d.\ samples. Their method also goes beyond the classical Koksma–Hlawka inequality: it achieves integration error $\widetilde{O}_d(\sigma_{\mathsf{SO}}(f)/n)$, where the smoothed-out variation $\sigma_{\mathsf{SO}}(f)$ can be substantially smaller than the Hardy–Krause variation that governs the classical bound. However, their algorithm requires $n^2$ i.i.d.\ samples as input, and this quadratic blowup is inherent to any method based on the transference principle. In this work, we bypass the quadratic blowup: for any constant $\varepsilon > 0$, we show that $(1+\varepsilon)n$ i.i.d.\ samples suffice to both obtain the beyond-Hardy–Krause guarantee of~\cite{BJ25a}, resolving an open problem posed there, and to produce low-discrepancy point sequences. Our algorithms are variants of the online Haar-thinning method of Dwivedi, Feldheim, Gurel-Gurevich, and Ramdas~\cite{DFG+19}.
Read the full abstract
Atmospheric and Oceanic Physics
Windowed Envelope Statistics for Time-Domain Significant Wave Height Estimation From HF Radar
Significant wave height (SWH) retrieval from high-frequency (HF) radar typically relies on a weak second-order Doppler continuum that is sensitive to noise, interference, and spectral leakage. This letter presents a Windowed Envelope Statistics Estimator (WESE) that operates directly on beam-formed time-domain voltages. A second-order term obtained from a Neumann expansion of the rough-surface field equation motivates quadratic compensation of localized radar features. WESE extracts the mean, standard deviation, or variance from overlapping windows of the in-phase, quadrature, or envelope-magnitude sequence, followed by quadratic compensation, rank ordering, least-squares regression, and causal smoothing. Evaluation used 335 synchronized hourly observations from a 13.385 MHz, 12-element WERA system at Argentia, Newfoundland and Labrador. The optimal configuration used quadrature variance, a 16-sample window, 896 retained chronological samples, and 30-h smoothing, achieving an RMSE of 0.152 m and a Pearson correlation of 0.978. This represents RMSE reductions of 32.1% and 18.7% relative to previously reported linear and second-order compensated ordered-statistics models, respectively. The results demonstrate robust time-domain SWH estimation without explicit Doppler-spectrum construction.
Read the full abstract
Numerical Analysis
Stress-divergence, Laplacian, and rotational forms of the incompressible Navier–Stokes equations with variable viscosity
In the Navier–Stokes equations, incompressibility allows rewriting the viscous term in various forms leading to distinct numerical properties and flow descriptions. Furthermore, models accounting for non-Newtonian, thermal or turbulent effects often break the constant-viscosity assumption, thereby producing additional consistency terms. In this context, the present work compares the classical symmetric-gradient diffusion term with more recent variable-viscosity generalizations of the Laplacian and rotational forms. We discuss, analyze and test their differences with respect to implementation, efficiency, numerical stability and outflow boundary conditions. With a focus on time-dependent flows, we consider second-order implicit-explicit (IMEX) temporal discretizations aimed at improving efficiency and numerical stability. Through a rigorous stability analysis, we show how selected explicit treatments can bypass algorithmic nonlinearities without inducing CFL conditions. Our numerical results highlight important differences between the three viscous formulations—especially in the presence of outflow boundaries, for which the generalized Laplacian form proves more suitable in diffusion-dominated regimes. %(as widely known for constant viscosity).
Read the full abstract
Robotics
ObstaDiff: Generalizable Diffusion Policy Learning via Obstacle-aware Representations
Imitation learning has achieved impressive results in robotic manipulation, yet most existing approaches assume clean backgrounds and lack explicit mechanisms for obstacle-aware motion generation. Extending such policies to cluttered, real-world scenes with unstructured obstacles remains a key generalization challenge. We present ObstaDiff, a decomposed diffusion-policy framework with a lightweight obstacle-aware visual encoder. ObstaDiff extracts a structured target-obstacle-background representation, enabling the downstream alignment policy to generate end-effector trajectories toward a target-centered bottleneck pose while reasoning about surrounding obstacles. We evaluate ObstaDiff on 61 real-robot greenhouse trials per method (366 executions in total). ObstaDiff achieves 75.41% average task success and 8.20% average obstacle collision rate, outperforming representative imitation-learning baselines and improving generalization in cluttered agricultural scenes.
Read the full abstract
Computer Vision
HiPerViT: A Hierarchical Perceiver-Vision Transformer Architecture for Multi-Scale Texture Recognition
Texture recognition remains challenging for modern vision models because discriminative evidence is often carried by higher-order spatial statistics rather than by object shape alone. While Vision Transformers provide strong long-range modeling capacity, their standard object-centric representations do not explicitly expose such statistical structure, which limits texture sensitivity in fine-grained recognition settings. We present HiPerViT, a compact vision-only architecture that injects an explicit second-order statistical prior into a transformer-based recognition pipeline. The method combines global and local image views with a compact bilinear descriptor encoded as a statistical token, and integrates this token with first-order spatial representations through Perceiver-style latent distillation. This design enables direct interaction between spatial tokens and second-order feature co-occurrence statistics, providing the model with explicit access to texture-relevant information without requiring multimodal pretraining or ensemble construction. Across six texture recognition benchmarks, HiPerViT achieves consistent improvements over strong vision-only baselines under the reported evaluation protocols, including gains of +3.05 percentage points on DTD, +10.48 on GTOS-Mobile, and +10.10 on 1200Tex. Beyond benchmark performance, our analyses show that these gains are largely invariant to the backbone depth used to extract second-order statistics and to the ordering of interaction and distillation stages. This pattern suggests that the primary source of improvement is not a specific fusion topology, but the explicit availability of second-order statistical information as a first-class representational signal. These results support explicit statistical tokenization as an effective and robust design principle for texture-centric visual recognition.
Read the full abstract
Computational Geometry
Almost Linear Universal Point Sets for Planar Graphs
A point set is universal for planar graphs on $n$ vertices if every such graph has a straight-line drawing without crossings whose vertices belong to the set. We construct universal point sets of size $n^{1+o(1)}$, improving the previous quadratic upper bound. Our construction uses the reduction of Bannister, Cheng, Devanny, and Eppstein from universal point sets to superpatterns for $213$-avoiding permutations. We represent these permutations by ordered rooted forests and construct a small family of intervals containing every such forest. The result follows from a straightforward bound on the size of the family of intervals. GPT-6 Astra assisted in developing the construction and proof.
Read the full abstract
Robotics
IMLE-VLA: Fast Single-Step Action Generation for Vision-Language-Action Policies
Vision-language-action (VLA) policies leverage pretrained vision-language backbones to achieve strong cross-task generalization. A leading design couples this backbone with a dedicated continuous action head trained via diffusion or flow matching. However, such heads rely on iterative multi-step sampling, for example 10 Euler steps in $\pi_{0.5}$. This creates an inference bottleneck that produces stop-and-go movement in the robot and slower task completion. We introduce IMLE-VLA, which replaces the iterative action head with a single-step conditional generator trained via conditional Implicit Maximum Likelihood Estimation (cIMLE). The cIMLE objective promotes multimodal action coverage, avoiding the mode collapse of naive regression heads while eliminating multi-step sampling entirely. When IMLE-VLA is applied to $\pi_{0.5}$, it increases inference frequency 3.67x (55 Hz vs. 15 Hz), enabling up to 11x higher action throughput. On the 40-task LIBERO benchmark, IMLE-VLA achieves the highest average success rate (98.0%) among all baselines while leading in inference frequency. Under the test-time perturbations of LIBERO-plus, IMLE-VLA retains $\pi_{0.5}$'s robustness while other baselines degrade sharply, confirming that the cIMLE head preserves generalization. Real-world experiments on a Franka Emika Panda across four tasks demonstrate smoother motion (2.2x to 3.0x lower jerk) and faster task completion, with IMLE-VLA outperforming $\pi_{0.5}$ on every task and reducing average VLA inference time per episode by 3.9x to 6.6x. Videos and code are available at https://kianhk6.github.io/IMLE-VLA/
Read the full abstract
Wednesday 9 September
Physics and Society
All edges lead to Rome: analyzing Late Antiquity regime shifts using coin hoard data
Complex network analysis and other statistical techniques can be fruitfully applied to historical data to understand dynamics that brought the territory under study to a certain state. In this paper we use the FLAME database to understand when a drastic change of trade regime occurred in the Italian peninsula by looking at mint-hoard links that are evidence of a physical connection between them. By using changepoint analysis to pinpoint precise dates in the time series of these links, and then historical as well as network data to establish the sequence of network collapse, we find that the demise of the Western Roman Empire started at the beginning of the 5th century AD, yielding a more precise sequence of how and when it reached Italia.
Read the full abstract
Computation and Language
Intrinsic Attention as Navigator: A Non-Generative Retrieval Method for Multi-Hop Queries
For multi-hop queries in retrieval-augmented generation (RAG), it is crucial to accurately obtain evidence pieces. In this study, we propose a training-free method for multi-hop query retrieval, named Intrinsic Navigating-based Non-Generative Retrieval (IN-Retriever). Instead of generating intermediate reasoning text, the method extracts the bridging clue from the internal attention of the embedding model based on the current context and retrieved candidate documents. During iterative clue extraction and document retrieval, IN-Retriever strips away revealed clues from the embedded query vector and then reranks candidate documents in a targeted manner, thereby steering retrieval toward hidden next-hop evidence. In the experiments, we confirm the feasibility and effectiveness of using intrinsic attention signals to guide external multi-hop retrieval. Experimental results show that our method precisely collects sufficient documents to support multi-hop queries and significantly outperforms sparse and dense retrieval methods.
Read the full abstract
Computation and Language
The Judge Is Not Impartial: Self-Preference in Medical LLM Evaluation
Large language models (LLMs) are increasingly used to evaluate model-generated medical content, creating a potential circularity: an LLM judge may favor outputs generated by the same model independently of clinical quality. To characterize LLM judge self-preference in medical settings, we measured same-model affinity in single-turn specialist question answering and multi-turn standardized-patient encounters. We evaluated eight models, two from OpenAI, Anthropic, Google, and Qwen, on 620 real point of care questions from clinicians (Real-POCQi). The same eight models also generated and judged 200 standardized patient-clinician scenarios (MedSP1000) up to 8 visible turns. In total, our analysis spans 13,760 LLM judgments. We define three complementary measures of self-preference. Answer Rank Bias, our primary measure, compared the rank assigned to a fixed response by its own-model judge with its average rank from outside judges. Answer Score Bias made the analogous comparison using rubric scores. Own-Answer Win Rate measured how often a judge ranked its own model's response above a competitor. Higher values across all these measures indicate greater degrees of self-preference. In Real-POCQi, pooled Answer Rank Bias was $+1.22$ positions, and all eight judges favored their own answers. Self-preference persisted both with and without a grading rubric. Revealing generator identities modestly weakened self-preference by an average of $-0.094$ positions. In MedSP1000, pooled Answer Rank Bias ranged from $+0.53$ to $+0.62$ positions, with no clear difference by conversation length ($p=0.17$). LLM judges systematically favored responses generated by their own models across multiple formats. Clinical AI deployments should exercise increased vigilance including multiple evaluation measures and judge panels spanning differing model families when deploying LLM-as-a-judge systems.
Read the full abstract
Computation and Language
Self-Evolving Dual-Role Deliberation for Medication Recommendation from Clinical Records
Recommending medications from electronic health records (EHRs) is a safety-critical task that demands therapeutic coverage, explicit safety constraints, traceable justification, and continual adaptivity. Supervised neural models freeze at deployment and cannot reason about new phenotypes, while existing LLM pipelines are single-pass and lack a cross-case improvement loop. We propose SEMPA, a training-free agentic framework that couples a Proposer with a Critic in multi-round adversarial deliberation, grounds candidates with molecular-aware retrieval, contrastively evolves both agents' natural-language policies under a semantic-anchor constraint, and enforces rule-based DDI and contraindication checks. On 100 MIMIC-IV encounters, SEMPA reaches 0.3116 Jaccard, 0.7295 F1, and a 0.0172 DDI rate, improving F1 by 57.7% and cutting DDI by 77.5% over the strongest supervised baseline. Two LLM judges rank SEMPA highest, and six-backbone studies confirm that gains derive from the framework rather than a specific model.
Read the full abstract
Human-Computer Interaction
Fairness Explanations Improve Fairness In Human-AI Decision Making: an empirical evaluation of parole decision-making
Both humans and AI exhibit systemic biases, yet little is known about how interface design can foster fairness in hybrid human-AI decision-making. This paper introduces a method for extracting fair data samples from biased datasets and adapts a fairness-explanation technique to highlight potential AI bias during decisions. We conducted a within-subjects study (n=48) where non-experts made decisions across three conditions: independent human judgment, unexplained AI assistance, and AI assistance augmented with visual ethnic-bias explanations (Fair-XAI). To isolate effects, the AI conditions provided systematically balanced recommendations. Results show that fairness explanations significantly reduced both sex and ethnic bias compared to human-only decisions. Fair-XAI also marginally improved fairness over balanced AI recommendations while enhancing task performance. These findings demonstrate that fairness-centered explanation interfaces can meaningfully shape user behavior, providing concrete design insights for developing equitable and accountable human-AI systems.