Machine Learning
TabReliBench: Evaluating Tabular Foundation Models Beyond Predictive Accuracy
Tianqi Zhao, Jia Guo, Yan Shuo Tan et al. (+1 more)
Tabular foundation models (TFMs) are increasingly popular for their strong predictive performance. Many also provide posterior predictive distributions (PPDs), whose reliability remains less studied. We assess reliability through point prediction, distributional quality, marginal and subgroup calibration, and interval efficiency. We then introduce TabReliBench, a large classification and regression
benchmark for these comparisons and
dataset-level analyses. We find that TFMs' rankings change between TabArena and our
benchmark and across reliability metrics. Regression TFMs usually achieve near-nominal marginal coverage and retain narrow intervals, but this favorable marginal behavior does not imply valid conditional coverage.
Dataset-level analyses show that classification difficulty often accompanies poor calibration, whereas regression predictive performance is not a reliable indicator of calibration. Ensembling improves predictive performance but does not consistently improve calibration of PPDs; TFMs also benefit little from conformal post-processing, which does not reliably resolve subgroup gaps. Reliable TFM selection therefore requires multidimensional evaluation beyond aggregate rankings. Code is available at https://github.com/Tianqi-Zhao/TabReliBench.
Read the full abstract ▾
All authors Tianqi Zhao, Jia Guo, Yan Shuo Tan, Qiong Zhang
Machine Learning · Cross-listed from stat.ML
A Critical Look at Power Sampling when Test Time Sharpening LLMs
Abiel Malkani Talwar, Christopher Nemeth and Henry Moss
Reinforcement learning with verifiable rewards (RLVR) has substantially improved LLM
reasoning, yet growing evidence suggests it mainly performs distribution sharpening: concentrating probability mass on
reasoning paths the base model already produces. Recent work therefore seeks RLVR's gains without needing reward signals, by sampling from the power distribution $\pi_\alpha \propto \pi_{\text{base}}^\alpha$ via Markov chain Monte Carlo or sequential Monte Carlo. We show that these samplers fail to converge to $\pi_\alpha$ under realistic compute budgets, and that their gains are largely recreated via simple Best-of-$N$ with appropriate grading . This suggests they succeed by selecting high-probability sequences, not by approximating $\pi_\alpha$, raising a natural question: is the power distribution itself a worthwhile target? We test this hypothesis directly by fine-tuning the base model to match $\pi_\alpha$, sidestepping sampler convergence issues entirely. Despite approximating the target far more faithfully than existing samplers, this strategy yields gains that are noisy and specific to certain models and sampling tasks. Together, our results suggest that the value of power sampling lies not in faithfully targeting $\pi_\alpha$, but in the simpler goal of extracting sequences the base model already deems likely.
Read the full abstract ▾
Cryptography and Security
ToFu: Exploiting and Fuzzing the LLM-Orchestrator-Tool Trust Boundary
Md Nazmul Islam Ananto, Pierre Prevost-Avenel, Colin Smith et al. (+3 more)
Large
language models (LLMs) are increasingly deployed as
agents that interact with external software through tools, with an orchestration layer translating model outputs into structured tool invocations. This architecture creates an LLM-orchestrator-tool trust boundary where tool semantics may go unenforced and LLM-generated arguments may be incorrectly treated as trusted. We present ToFu, a tool-aware framework that automatically derives semantic security checks for LLM tools from their existing metadata and a CWE-guided sanitizer corpus. ToFu specializes these checks to individual tools and automatically generates fuzzing harnesses and semantic oracles using runtime introspection and outcome sanitization. We evaluated ToFu across 273 MCP servers containing 3405 tools and identified verified bugs corresponding to CWE weakness classes in 343 tools across 118 of the servers. Our findings show that real-world tools frequently fail to enforce their documented input semantics, making it unsafe to treat LLM-generated arguments as trusted.
Read the full abstract ▾
All authors Md Nazmul Islam Ananto, Pierre Prevost-Avenel, Colin Smith, Kyungtae Kim, Taegyu Kim, Arslan Khan
Computational Geometry
A Bicriteria Approximation Scheme for the Minimum Well-Separated Pair Decomposition
Guilherme D. da Fonseca
A well-separated pair decomposition (WSPD) of a set $P$ of $n$ points covers all $\binom{n}{2}$ pairs of points by a family of dumbbells $\{X,Y\}$, each covering the pairs of $X\times Y$, such that the diameters of $X$ and $Y$ are at most an $\varepsilon$-fraction of their distance. In $\mathbb{R}^d$, WSPDs with $O(n/\varepsilon^d)$ dumbbells always exist, but many inputs admit much smaller ones. However, computing a WSPD of minimum size is NP-hard in the plane. For $d\ge2$, the only known approximation that preserves the separation has a factor of $2^{O(d)}$, and Buchin et al. (ESA 2026) asked whether a $(1+\delta)$-approximation exists. We give a bicriteria approximation scheme: for fixed $d$ and $\delta\in(0,1]$, we compute in $f(d,\delta)\cdot n^{6}$ time a $(1+\delta)\varepsilon$-WSPD whose size is at most $(1+\delta)$ times the minimum size of an $\varepsilon$-WSPD. The same holds in metric spaces of bounded doubling dimension. The
algorithm combines shifting over scales with the local-growth technique of Nieberg, Hurink, and Kern, applied to a graph whose vertices are the pairs of points. In contrast, we show that in general metric spaces, relaxing the separation by a factor below $2$ keeps the problem hard: for every fixed rational $\varepsilon\le1/4$, it is NP-hard to compute such a relaxed WSPD whose size is within a factor $n^c$ of the minimum, for every $c<2/3$.
Read the full abstract ▾
Machine Learning · Cross-listed from physics.chem-ph
Electron-Density-Informed Molecular Property Prediction Using Surrogate Functionals
Manuel V. Klockow, Simon Wagner, Roman Remme et al. (+2 more)
Machine learning methods are increasingly used to predict chemical properties directly from the molecular structure to bypass the often prohibitive cost of
quantum mechanical calculations. Recent work has shown that including the ground-state electron density as input to the machine learning model can improve accuracy in this setting. This work employs the recent surrogate functional framework originally developed for orbital-free density functional theory and extends it to predicting molecular properties. We show that passing the ground-state density as input to the model, while simultaneously enforcing it to be the minimum of the surrogate functional, surpasses state-of-the-art performance for six out of seven extensive properties of the QM9
dataset. Additionally, the approach shows improved generalization to larger molecules taken from the QM40
dataset. On all extensive properties of both
datasets, our approach outperforms models that are informed by the ground-state density but do not enforce the surrogate functional constraints.
Read the full abstract ▾
All authors Manuel V. Klockow, Simon Wagner, Roman Remme, Dominic Plein, Fred A. Hamprecht
Information Retrieval
SPARQL-LLM in the TEXT2SPARQL 2026 Challenge
Panayiotis Smeros, Vincent Emonet, Ruijie Wang et al. (+2 more)
Querying distributed data represented as Knowledge Graphs (KGs) via SPARQL remains a difficult task, requiring fluency in query syntax, endpoint specifications, and schema details. Knowledge Graph Question Answering (KGQA) systems address this by automatically translating natural
language into SPARQL, and recent advances in Large
Language Models (LLMs) have made this translation increasingly feasible. In this paper, we describe SPARQL-LLM, our open-source KGQA system, in the context of the TEXT2SPARQL 2026 challenge. In this challenge, SPARQL-LLM ranks first overall among eight competing approaches across all
datasets, achieving up to 59% higher accuracy than the second-best approach and supporting multilingual question answering in English, German, and Spanish. SPARQL-LLM is publicly released at https://github.com/sib-swiss/sparql-llm and deployed in production over decentralized bioinformatics KGs at https://www.expasy.org/chat.
Read the full abstract ▾
All authors Panayiotis Smeros, Vincent Emonet, Ruijie Wang, Ana-Claudia Sima, Tarcisio Mendes de Farias
Software Engineering
A-SAW: The Agentic Software Architecture Workflow
Finn Klatt, Luca Di Grazia and Johannes Schneider
AI assistance in software engineering focuses on implementation and verification, leaving the design phase, and in particular the maintenance of architecture documentation, comparatively underserved. Architecture as Code makes architectural artifacts versionable, diffable, and machine-readable, but it does not reduce the effort of keeping them updated, so design documentation still ends up as a fragmented, author-dependent, and only partially trusted record of why a system looks the way it does. In this paper we introduce the Agentic Software Architecture Workflow (A-SAW), a process-level artifact that extends Architecture as Code with a persistent
agent layer: an IDE-integrated
agent drafts architecture documents from a version-controlled C4 model, a typed element catalog, and instrumented templates. Every draft passes a mandatory human review gate before it is accepted. We evaluate it in a large industrial organization along two tracks: a rubric assessment of 45 architecture documents including 13 matched source-regeneration pairs, and 13 task-anchored practitioner interviews. Our results show that the workflow collapses the between-author variance that made quality unpredictable. Intrinsic quality also rises, by a median of +19 points across the matched pair. The cohort has a mean of 58% and 89% bundle templates, catalog,
agent, and a richer regeneration context. So they bound the
agent's contribution from above rather than isolating it. In the
agent-generated documents, completeness (100%), consistency (93%), and machine-readability (99%) reach or approach the rubric maximum, while traceability reaches 79%. Eleven of the 13 participants rejected
agent output that no human had reviewed, and one warned that polished output may lead reviewers to skip the review. Our work supports keeping human review mandatory and concentrating it where the
agent must infer beyond the repository.
Read the full abstract ▾
Machine Learning
SelectArena: A Modern Benchmark for Feature Selection on Real-world Tabular Data
Bastian Schäfer, Dominika Matus, Lennart Purucker et al. (+5 more)
Selecting a subset of features (columns) from a tabular
dataset is a ubiquitous challenge in real-world applications. Industry experts seek to minimize model training costs,
climate scientists require an interpretable subset of factors to understand extreme weather events, and clinicians need to choose among diagnostic tests due to financial and time constraints. When looking at the scientific literature to determine which feature selection method to use, practitioners across domains face an unrepresentative landscape of benchmarking. Despite extensive research, existing feature selection
benchmarks provide limited guidance because they rely on outdated models, ad hoc evaluation protocols, or artificially generated data that do not reflect real-world problems. To close the gap between scientific benchmarking and real-world applications, we introduce SelectArena, a modern
benchmark for feature selection. We compare 15 widely used feature selection methods on 71 manually curated real-world tabular
datasets, covering classification and regression problems, using standardized evaluation protocols and state-of-the-art tabular (foundation) models. Our results reveal clear trade-offs across predictive performance, runtime, stability, and validity. Tree-based methods, CART and RFImportance, along with F-Test are Pareto-optimal when taking into account predictive performance and runtime. F-Test, information-theory-, and distance-based methods lead on stability and validity, with F-Test emerging as a balanced choice across our metrics. Search-based methods are slow, unstable and achieve low performance. SelectArena turns these findings into practical guidance and, as an extensible platform, supports future feature selection research.
Read the full abstract ▾
All authors Bastian Schäfer, Dominika Matus, Lennart Purucker, Ricardo Knauer, Erik Rodner, Frank Hutter, Anton Sokolov, Arnaud Liefooghe
Robotics
DexBioLab: Benchmarking Dexterous Manipulation with Dynamic Experimental Phenomena in Realistic Laboratories
Xuanye Wu, Yue Shi, Yongcan Yang et al. (+8 more)
Automating biological laboratories requires
robots to perform dexterous tool interactions under realistic visual conditions while handling dynamic experimental phenomena, presenting significant challenges for robotic manipulation. To enable systematic evaluation under these conditions, we introduce DexBioLab, a simulation
benchmark and framework for dexterous manipulation in realistic biological laboratory environments. DexBioLab comprises 14 tasks across 4
robot embodiments, spanning functional tool use, precision manipulation,
language-conditioned visual grounding, and instrument operation. We introduce controllable dynamic experimental phenomena, including flames, vapor, and bubbles, to capture the evolving visual conditions of laboratory procedures. To further enhance environmental realism, we develop a hybrid rendering pipeline that integrates interactive simulated foreground objects with Gaussian-splatting backgrounds reconstructed from 6 real laboratory scenes. For efficient large-scale data generation and evaluation, we further build a parallel simulation framework on top of mjlab, proposing Parallel Finite-State Machine (PFSM) and integrating batched rendering with decoupled trajectory replay for demonstration generation and model-independent closed-loop policy evaluation. Our evaluation reveals substantial variation in policy performance across tasks, exposing distinct strengths and weaknesses of current manipulation policies. Real-world experiments further demonstrate a reduced sim-to-real gap, supporting the validity of AutoBio+ for realistic manipulation evaluation. Supplementary materials are available at https://cumulusalpha.github.io/DexBioLabWebPage.
Read the full abstract ▾
All authors Xuanye Wu, Yue Shi, Yongcan Yang, Yiqun Wang, Zhaoqi Wan, Dayu Xia, Yanming Shao, Jiezhang Cao, Xiangfei Li, Yao Mu, Guangtao Zhai
Artificial Intelligence
Learning Size Independent Constraint Models from Input Data and Positive and Negative Example Solutions
Helmut Simonis
In this paper we describe a Constraint Acquisition Tool that generates size independent constraint programs from input data and positive and negative example solutions of multiple problem instances. The system finds size specific models based on typical constraint patterns over global and binary constraints, then generalizes the models to a size independent model that can be run on unseen input data. The system produces MiniZinc code as well as a textual description of the model. We evaluate the tool on the example problems of the PTHG21 Challenge and show the feasibility of the approach.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from eess.IV
Model-agnostic Knowledge Augmented Machine Learning for Weakly Annotated Bioimaging Data
Jayanie Bogahawatte, Sachith Seneviratne, Simon Maksour et al. (+6 more)
Machine learning models are increasingly used for automated feature extraction from weakly annotated bioimaging data, such as microscopy and cytomorphology images. Due to the sole dependency of automated feature extraction, existing methods yield inconsistent classification performance across weakly annotated
datasets. Here, we propose a model-agnostic knowledge augmentation method that enables ML models to exploit domain knowledge beyond what automated feature extraction can capture. The proposed approach improves classification performance across seven ML architectures. When integrated with masked autoencoder-based self-supervised learning, our method achieves the strongest overall performance, improving precision by 2.9% over the next-best method on three in-house microscopy
datasets and by an average of 6.8% across three public cytomorphology
benchmarks. We perform
nucleus-level classification based on mislocalization of TAR DNA-binding
protein 43 (TDP-43) from the
nucleus to the cytoplasm, a phenotype linked to Amyotrophic Lateral Sclerosis (ALS), using three independent microscopy
datasets of motor
neurons derived from human induced pluripotent stem
cells. Accurate quantification of this phenotype at scale remains challenging and may contribute to inconsistencies across studies. Application of the best-performing model to these microscopy
datasets reveals biological interpretations related to ALS, including dose-dependent effects of the proteasomal stressor, MG132, and the oxidative stressor, sodium arsenite (SA), on TDP-43 mislocalization, variable pathology in motor
neurons from sporadic and familial ALS donors, and a lack of protection of the pharmacological compounds, edaravone and niclosamide, against SA-induced TDP-43 mislocalization. Together, our pipeline enables in-vitro cellular image analysis to advance understanding of ALS-associated pathology and potential treatments.
Read the full abstract ▾
All authors Jayanie Bogahawatte, Sachith Seneviratne, Simon Maksour, Dzung Do-Ha, Elizabeth Qian, Samuel Brennan, Matthew Keon, Lezanne Ooi, Saman Halgamuge
Computer Vision
Benchmarking Answer-Grounding Consistency in Remote Sensing Visual Question Answering
Ping Guo, Zhiqi Huang, Xinran Li et al. (+3 more)
In grounded remote sensing visual question answering, a model produces a textual answer and identifies relevant objects in the image as grounding evidence. Existing evaluations measure answer correctness and localization quality, but do not directly assess answer-grounding consistency. We propose the Answer-Evidence Support Protocol (AESP) to assess this consistency through evidence support: whether the grounding evidence supports the model's own answer. AESP applies task-specific rules to existence, counting, and comparison questions. We further construct the Grounded Remote Sensing Visual Question Answering (G-RSVQA)
benchmark, containing 15,027 samples with questions, reference answers, and all related object-level annotations. Experiments on seven grounded
vision–
language models reveal failures of evidence support across all three question types. Up to 50.82% of correct answers in the evidence evaluation subset are not supported by the provided grounding evidence.
Read the full abstract ▾
All authors Ping Guo, Zhiqi Huang, Xinran Li, Linkun Yu, Jingyin Pei, Kamata Seiichiro
Machine Learning
PANDA: Prior-guided Attentional Dual-path Architecture
Rishabh Jain, Federico Belotti, Stefano Coniglio et al. (+2 more)
Hypergraph Neural Networks (HNNs), designed to model higher-order relations beyond pairwise interactions, typically rely either on fixed structural weights or on normalised
attention. Fixed weights can propagate misleading information in heterophilic settings, while normalisation makes
attention coefficients competitive, in the sense that, in a message-passing framework, increasing the weight of one sender must reduce the weight of others in the same aggregation set. Our main contribution is the introduction of PANDA (Prior-guided AttentioNal Dual-path Architecture), a plug-in for two-stage message passing HNNs applicable to both undirected and directed settings. The main path interpolates between learned receiver-normalised
attention and structural incidence priors, while a non-competitive (not normalised) auxiliary path lets each sender contribute its own transformed information. We also show that, with fixed structural coefficients, PANDA recovers a certain complex-valued convolution operator adopted in previous works. Across four HNNs spanning spatial and spectral methods in both directed and undirected settings, plugging PANDA in yields an average improvement of 5.18 percentage points across eight real-world
benchmark datasets. The code is available at https://github.com/jainris/panda.
Read the full abstract ▾
All authors Rishabh Jain, Federico Belotti, Stefano Coniglio, Pietro Lio, Stefano Fiorini
Artificial Intelligence
Humanize Olympiad Agents: A Multi-Session Workflow for Olympiad Problem Solving with Solve, Review, Wiki, Verify
Zhengyang Zhang, Jing Xiong, Hongzhou Lin et al. (+5 more)
Large
language model (LLM)
agents are rapidly improving on long-horizon scientific and olympiad tasks. Recent systems have produced a Lean-formalized resolution of the Navier–Stokes problem and autonomous solutions to four open questions in the Erdős
Conjectures database. Yet neither system releases both its underlying model and complete inference stack. More broadly, a single
agent session remains brittle on difficult problems: early errors can propagate, missing domain knowledge can stall progress, and apparent success can mask a misinterpretation or weakening of the original task. We introduce HOA (Humanize Olympiads
Agent), an open, flow-level framework that coordinates multiple
agent sessions through three complementary feedback channels. First, an independent reviewer evaluates each candidate and provides actionable feedback. Second, a structured wiki supplies task-specific knowledge that may be missing or unreliable in model pretraining. Third, a verifier checks generated Lean proofs against trusted challenge statements. With GPT-5.6 Sol, HOA achieves perfect scores on the IMO, IPhO, and IOI 2026, near-full marks on the IChO 2026 problem set, and solves all 672 problems (100%) on PutnamBench. On the Lean-Eval leaderboard, HOA ranks first, solving 251 of 304 problems (82.6%). We further evaluate HOA with open models. Using Kimi-K3, HOA achieves full scores on the IMO, IPhO, and IOI, and a gold-level score of 97.1% on the IChO. These results closely match the closed-source GPT-5.6 Sol + /Goal baseline on the IMO and IPhO, while exceeding it on the IChO and IOI, demonstrating that systematic feedback can help bridge substantial differences in base-model capability.
Read the full abstract ▾
All authors Zhengyang Zhang, Jing Xiong, Hongzhou Lin, Juihui Chung, Zijian Zhang, Song Han, Sihao Liu, Ligeng Zhu
Machine Learning
Private and Distributionally Robust Federated Multiclass Classification
Andrea P. Hernandez, Michael Ibrahim and Nagi Gebraeel
Fault diagnosis in asset-intensive industries requires multiclass models trained on noisy, sensitive data collected across geographically distributed client sites.
Federated learning enables collaborative training without centralizing data, but most approaches do not provide formal
privacy guarantees for individual records stored at client sites and can degrade under label noise. To address these limitations, we develop a differentially private, distributionally robust framework for
federated multiclass classification that provides sample-level
privacy for records stored at each client. We build on a
federated Wasserstein distributionally robust multiclass objective with the Crammer-Singer loss, solved using an Alternating Direction Method of Multipliers (ADMM)-based
algorithm. To protect the released client updates, our method injects calibrated Gaussian noise into the proximal linearized client updates of both the local model and the local distributionally robust variable before aggregation at the central server, and we quantify the cumulative
privacy loss using the moments accountant. We provide per-round sample-level $(\epsilon,\delta)$-differential
privacy guarantees and an overall composed
privacy guarantee across all training rounds, together with convergence results and an explicit
privacy-utility tradeoff. Experiments on simulated industrial fault data and UCI
benchmark datasets show that our method outperforms differentially private
federated empirical risk minimization and
stochastic gradient descent baselines in most tested settings, with the largest gains under strong
privacy-induced noise.
Read the full abstract ▾
Computation and Language
Multi-Stream Attention Residuals
Xiaohan Qin, Xiangdong Zhang, Tuo Dai et al. (+7 more)
Attention Residuals (AttnRes) enables content-dependent
retrieval over network depth, which can be viewed as single-stream softmax
attention along the depth axis. Inspired by the success of multi-head
attention along the sequence axis, we introduce Multi-Stream
Attention Residuals (MSAR), an extension of AttnRes that generalizes this mechanism to multi-head depth
attention, yielding substantial performance gains with minimal additional overhead. MSAR draws stream diversity from sequence context through multi-scale causal convolutions, allowing different heads to retrieve historical representations through distinct contextual views while preserving the integrity of the hidden representation. The retrieved streams are combined through lightweight adaptive mixing to generate the single-stream layer input. By retaining single-stream sources and expanding them on demand, MSAR supports multi-stream computation without requiring persistent multi-stream storage or communication. Across all 12 downstream
benchmarks at the 18B and 28B MoE scales, MSAR consistently outperforms prior residual variants, including standard residual connections,
manifold-constrained Hyper-Connections (mHC), and AttnRes. Scaling-law experiments yield an iso-loss compute speedup of $1.58\times$ over the standard residual baseline, compared with $1.35\times$ for AttnRes and $1.30\times$ for mHC. These results make MSAR a promising residual architecture for large-scale pre-training.
Read the full abstract ▾
All authors Xiaohan Qin, Xiangdong Zhang, Tuo Dai, Sunan Zou, Huaijin Wu, Yebin Yang, Zhuo Xia, Lin Yao, Yuliang Liu, Junchi Yan
Numerical Analysis · Cross-listed from math.AP
An inverse spectral problem for the wave equation with a varying Robin parameter at the boundary
Antti Hannukainen, Petr Kulikov, Matti Lassas et al. (+1 more)
An inverse spectral problem for a $(1+1)$-dimensional wave equation with wave speed $c(x)$ is considered. The spectral data consist of the
eigenvalues of the Neumann–Dirichlet problem for the operator $-c^2(x)\partial_x^2$ together with their derivatives with respect to a Robin parameter at the boundary. We propose a two-step recovery procedure. In the first step, the Neumann-to-Dirichlet map $\Lambda$ is reconstructed from the spectral data using the results of arXiv/0406582. In the second step, the reconstructed $\Lambda$ is used within the regularization strategy for the boundary control method presented in arXiv:1509.04478. As an improvement, the regularization of the control problems employs a weighted combination of $L^2$ and $H^1$ penalty terms, and we prove a uniform convergence result for the family of optimal controls, thereby allowing the regularization parameters to be tuned simultaneously. We also prove that the forward spectral map, which assigns to a wave speed its spectral data, is Lipschitz continuous, and that the inverse spectral map underlying the reconstruction procedure is Hölder continuous on a subclass of wave speeds. The method is implemented numerically and validated on both smooth and piecewise smooth wave speed profiles. Numerical experiments demonstrate that the recovery of the wave speed exhibits high accuracy and robustness.
Read the full abstract ▾
All authors Antti Hannukainen, Petr Kulikov, Matti Lassas, Lauri Oksanen
Robotics
JitterVLA: A Replay Benchmark for Motion Quality in VLA-Driven Human-Robot Manipulation
Faith Ogunwoye, Didar Seyidov, Konstantin Gubernatorov et al. (+1 more)
Vision-
language-action (VLA) policies issue sparse targets whose arrival at the
robot controller is delayed and irregular, so a reasonable predicted action can still produce abrupt motion near people. Remedies are evaluated on different streams, delays and metrics. JitterVLA replays identical commands and arrival traces through alternative controllers, isolating the controller's effect. We apply it to synthetic reaches, hand trajectories from 20 people, three open VLA pipelines (838 episodes) and a physical UR3. Interpolation lowered modeled p99 jerk for every source, by 83% (615 to 106 m/s^3) on human motion with 200 ms delay, at a 2.0 cm tracking cost, and the effect replicated on 598 new episodes; jerk-limited generation lowered jerk further at a larger cost.
Entropy-discontinuity associations had opposite, replicated signs in two pipelines, and
entropy gating helped in one but not the other. On the UR3, interpolation lowered accelerometer-derived jerk by 3.0-14.0% in all six exploratory pairs, while a position-derived metric did not resolve the contrast. JitterVLA supports screening before human-centred studies; code and data will be released.
Read the full abstract ▾
All authors Faith Ogunwoye, Didar Seyidov, Konstantin Gubernatorov, Dzmitry Tsetserukou
Computation and Language
AMLworld-Compact: Importance-Weighted Downsampling for Cost-Effective LLM Evaluation and Error Diagnosis
Natapong Nitarach, Phume Ngampornsukswadi, Pittawat Taveekitworachai et al. (+3 more)
Evaluating large
language models (LLMs) on AMLworld's smallest test splits is costly because each of over a million targets requires a separate graph
prompt. Sampling reduces this cost, but small balanced subsets change unweighted precision and F1. We introduce AMLworld-Compact, an evaluation set that retains every illicit test transaction and samples benign transactions from supervised-score strata. Its HT-Coreset protocol reduces evaluation targets by 271x and 611x on HI-Small and LI-Small. Horvitz-Thompson weights give unbiased full-test confusion-count estimates for fixed predictions; recall is exact, while precision and F1 are ratio estimates. In a held-out study using Graph Feature Preprocessor (GFP) features, one boosted-tree family defines strata for evaluating the other over 500 draws per configuration. At the released sizes, F1 RMSE is 0.00-2.58 percentage points versus 4.44-7.41 for uniform sampling with HT weights; conservative intervals remain broad. Under the evaluated graph
prompts, seven open-weight LLMs have weighted precision near the predict-all-illicit baseline. In an outcome-stratified audit of 1,000 HI-Small few-shot outputs from one run per model, lexical checks pass for 84.5% of correct and 56.2% of incorrect binary verdicts, so these markers do not establish correctness. We release the evaluation sets, construction code, and audit tools at https://huggingface.co/
datasets/typhoon-ai/AMLworldCompactEval and https://github.com/nat-nischw/AMLworld-Compact.
Read the full abstract ▾
All authors Natapong Nitarach, Phume Ngampornsukswadi, Pittawat Taveekitworachai, Surapon Nonesung, Warit Sirichotedumrong, Kunat Pipatanakul
Software Engineering
SWE-DeceptionBench: Benchmarking Deceptive Behaviors of LLMs Across the Software Development Lifecycle
Xiting Liu, Lin Yang, Yitong Zhang et al. (+2 more)
Large
Language Models (LLMs) are increasingly used throughout the software development lifecycle. This growing reliance on LLMs raises concerns about their trustworthiness, with deception posing a particularly important risk. Deceptive behavior takes the form of misleading responses or assistance in fabricating, concealing, or distorting information. To study this behavior, we introduce SWE-DeceptionBench, a
benchmark for evaluating LLM deception in software engineering. The
benchmark covers six lifecycle stages and two deception triggers: Active Deception, where models recognize infeasible or conflicting requirements but still present them as achievable, and Passive Deception, where models help fulfill explicit user requests to mislead others. We further introduce Neutral, Reward, and Pressure conditions to study the effect of external inducements, resulting in 720 test instances in Chinese and English. Evaluating seven mainstream LLMs, we find substantial deception even under the Neutral condition, with aggregate rates of 49.88% for Active Deception and 64.29% for Passive Deception. Deception rates vary across lifecycle stages, and external inducements affect Active and Passive Deception to different extents. We also evaluate lightweight defenses, finding that guard models miss many deceptive requests, and Self-Reminder reduces but does not eliminate deceptive responses. These findings highlight the need for more effective deception detection and mitigation methods to support trustworthy LLM-based software engineering.
Read the full abstract ▾
All authors Xiting Liu, Lin Yang, Yitong Zhang, Jia Li, Shi-Min Hu
Machine Learning · Cross-listed from stat.ME
Kernel-based External Validity of Experimental Studies
Federico Matteucci and Gerhard Neumann
The external validity of a study is the extent to which its findings generalize to unseen experimental conditions, such as
datasets. We formalize the results of a study together with its research question, and define external validity as the probability that two independent and same-sized replications of the study produce results whose maximum mean discrepancy (MMD) is below a given tolerance. This approach, centered on the choice of a space of results and of an appropriate reproducing kernel for the MMD, is flexible enough to be deployed on multiple different types of results and research questions. Additionally, we establish an asymptotic power-law relation between the size of a study and the
quantiles of the MMD, which we use to estimate the desired size of an externally valid study from a small-scale preliminary study. We deploy our framework to rankings with ties and discuss kernels suited to different research questions. Experiments on existing case studies and on a simulated newly designed study illustrate the applicability of the framework. We release our code as part of the genexpy Python module.
Read the full abstract ▾
Numerical Analysis
Energy conserving compatible space–time FEM for deterministic and stochastic rotating shallow water
Maneesh Kumar Singh and Colin J. Cotter
We present a fully discrete energy conserving discretisation of the rotating shallow water equations with arbitrary order in space and time. A space time finite element discretisation is used, built from compatible finite element methods in space and continuous Petrov-Galerkin schemes in time. Projected auxiliary variables for
Hamiltonian derivatives are necessary for exact energy conservation. We provide an iterative solver approach based on monolithic patch smoothers, which allows us to efficiently converge a Newton method sufficiently that energy errors due to the solver are very small. We extend this construction to the energy conserving
Stochastic Forcing by Lie Transport formulation. Then we demonstrate the convergence, conservation and robustness in numerical tests.
Read the full abstract ▾
Cryptography and Security
APEX-Harm: Evaluating Agent Safety in Shared Professional Workspaces
Ziyi Yang, Adel Bibi, Guohao Li et al. (+2 more)
Professional
agents increasingly work through command-line interfaces and multiple tools over shared workspaces containing files, messages, and business applications. These environments create safety risks that are difficult to capture with attacks placed at a single, known read location. We introduce APEX-Harm, a
benchmark for evaluating
prompt and script injection against long-horizon
agents performing finance, consulting, and legal tasks in large workspaces. Each task contains 115 to 328 files, together with email, calendar, chat, document, spreadsheet, PDF, and code-execution tools exposed through MCP servers. The
benchmark includes six attack categories covering static and dynamic injections, task-specific script replacement, and attacker-provided generic scripts, and evaluates attack success, exposure, and task performance. We evaluate 12 frontier models on 180 task instances per model. Opus 5 and Sonnet 5 have the lowest overall attack success rates (8% and 20%), while task-specific script replacement and attacker-provided generic script attacks remain substantially more effective, averaging 81-84% and 73% attack success, respectively. Gemini 3.8 Flash achieves the highest mean task score (0.877) while still exhibiting a 33% attack success rate. These results show that completing the requested task is not sufficient evidence of safe behavior: an
agent can produce a correct deliverable while also performing an unauthorized action in the workspace.
Read the full abstract ▾
All authors Ziyi Yang, Adel Bibi, Guohao Li, Bernard Ghanem, Carlos Hinojosa
Computation and Language
Training User Simulators with Latent Intrinsic Feedback
Tarek Naous, Craig Boutilier, Wei Xu et al. (+1 more)
User simulators are gaining
attention as environments to train LLM assistants that can better cope with the nuances of multi-turn conversations with real users. Prior work on user simulation has emphasized generating realistic user utterances, leaving unexplored their potential as a source of direct feedback on the assistant's responses. Such evaluation signals may not be explicitly communicated in
language, but can instead reside in the latent
reasoning of users. In this work, we introduce user simulators with intrinsic feedback, namely, models equipped with latent evaluation states: internal thoughts, reactions, and scalar ratings that evaluate the assistant from a first-person user perspective at each conversation turn. We train LLM-based simulators using interaction logs between real users and assistants and demonstrate that intrinsic feedback improves simulation realism and critiquing behaviors versus prompted simulator baselines, while also providing turn-level evaluations that existing fine-tuned simulators lack. Further, evaluations generated by our trained simulators correlate better with human judgments than those of prompted simulators of the same size. Lastly, our experiments reveal that simulators with intrinsic feedback provide effective reward signals for multi-turn RL training, matching setups that rely on external judge LLMs for rewards while cutting RL compute by more than half.
Read the full abstract ▾
All authors Tarek Naous, Craig Boutilier, Wei Xu, Ofer Meshi
Machine Learning · Cross-listed from stat.ML
Some remarks on the connection between generative diffusion and optimal transport
Giulio Biroli and Marc Mézard
We discuss three simple connections between generative
diffusion and optimal transport. First, the joint law of the data and of their Ornstein-Uhlenbeck noising is, at every time, the solution of an entropic optimal-transport problem, at a temperature set by the noise level. Backward
diffusion can consequently be read as an annealing which gradually decreases the entropic regularization. Second, we study the speciation transition–how the modes of the data distribution emerge during generation–and compare how it happens in
diffusion and flow matching with respect to optimal transport. Using Gaussian mixtures as high-dimensional example, we show that the dynamical instability through which
diffusion and flow matching separate the modes becomes, for optimal transport, an exponentially thin boundary layer at the beginning of generation, which is rounded by entropic regularization. Third, considering a generic forward dynamics with a general confining potential, we show that the backward process driven by the score also satisfies a drift-based variational principle whose minimal cost is the
entropy difference between initial and final distributions.
Read the full abstract ▾
Social and Information Networks
Non-Bayesian Social Learning with Heterogeneous and Time-Varying Aggregation Rules
Sammy Oukaci, Pol Rodríguez Farrés and Jesus Cerquides
People constantly form opinions by combining what they observe with what friends, collegues, and communities tell them. Following this idea, this article generalizes earlier results in Non-
Bayesian social learning on strongly connected social networks, where
agents repeatedly update their beliefs by combining their own private signals with the beliefs of their neighbors. Previous work assumed that every
agent aggregates information in the same way and often does not change in time, assuming
agents never change how they reason and listen to others. We show that
agents' beliefs still converge almost surely to the true state when
agents use heterogeneous and time-varying (non-
stochastic) aggregation rules under certain conditions. That is, even when people change their way to aggregate others opinions, the community can still reach the truth as long as information flows among its members. We also provide convergence rate results under this scenario and show that, for some of these functions, it is at least equal to the best convergence rate we have found in the literature. We present the widely used aggregation functions in the Constant Elasticity of Substitution (CES) family as a specific case of our general model, which include rules such as the log-linear, DeGroot, and the harmonic rules, and provide a simple illustrative example to show the capabilities of the model.
Read the full abstract ▾
Computer Vision
CUBE: Class-conditioned Unknown Boundary Evidence for Open-Set Insect Species Recognition
Hao Tan, Qimin Zhang, Zhe Xiao et al. (+3 more)
Open-set insect species recognition faces significant challenges when deployed in real-world biodiversity scenarios, particularly when encountering unknown species that are visually close to known training classes. Unlike generic out-of-distribution samples, fine-grained unknown insects are often not visually anomalous. They instead tend to appear near the decision boundary of specific known species, leading to unreliable closed-set predictions. To address these challenges, we propose Class-conditioned Unknown Boundary Evidence (CUBE), a novel evidence-based framework that models unknownness around each known species rather than as a single global unknown class. Our key idea is to evaluate a test sample relative to the known species it most closely resembles, enabling local rejection for visually confusable unknown insects. CUBE learns class-conditioned boundary evidence from auxiliary species, preserves auxiliary species as shadow references during inference, and incorporates taxonomic candidate calibration to verify biological plausibility. When ecological metadata is available, we further introduce CUBE-E, which adds a lightweight location-consistency term without modifying the visual backbone. Experiments on the Open-Insect C-America
benchmark show that CUBE achieves superior performance, outperforming state-of-the-art baselines with 95.8%, 98.7%, and 99.1% AUROC on local, non-local, and non-moth unknown splits, respectively, while maintaining 94.9% closed-set accuracy. With ecological metadata, CUBE-E further improves the results to 96.0%, 99.7%, and 99.5% AUROC with 95.2% closed-set accuracy.
Read the full abstract ▾
All authors Hao Tan, Qimin Zhang, Zhe Xiao, Zekai Zuo, Wanting Liu, Jinghui Qin
Machine Learning · Cross-listed from eess.AS
Vocal Intensity Conversion with Continuous SPL Control via Adversarial Training
Quentin Le Tellier, Albert Rilliard, Olivier Perrotin et al. (+1 more)
Vocal intensity conversion renders a recorded utterance at a different vocal intensity, rather than simply scaling its amplitude, while preserving its content and speaker identity. Existing systems control intensity through a reference utterance or a few discrete categories. Instead, we control the conversion with any target sound pressure level (SPL) on a continuous dB scale. A lightweight converter, trained as a conditional cycle-consistent GAN without parallel data, transforms the features of a frozen
speech encoder, from which a pretrained vocoder resynthesizes the
speech. We compare three representations, log-mel spectrograms, WavLM, and SpeechTokenizer, and introduce an objective protocol that measures the achieved intensity change against the target one on a calibrated corpus. With WavLM, the converter achieves a measured 83% of the target change in intensity and also shifts the fundamental frequency. A perceptual evaluation confirms that perceived vocal intensity increases with the target level on amplitude-normalized stimuli.
Read the full abstract ▾
All authors Quentin Le Tellier, Albert Rilliard, Olivier Perrotin, Marc Evrard
Computation and Language
BLiMPPI: Benchmark of Linguistic Minimal Pairs for Polarity Items
Arzu Burcu Güven, Anna Rogers, Leonie Weissweiler et al. (+2 more)
Negative polarity items (NPIs), such as "anymore" and "ever", are grammatical only in select environments, predominantly in the context of negation. Because their grammaticality depends on both syntax and semantics, they have a rich, multilingual literature in theoretical linguistics, yet work on NPIs in
language models is almost entirely confined to English. We introduce BLiMPPI, a multilingual evaluation suite of over 12,000 NPI minimal pairs in 33
languages, built by a pipeline that discovers NPIs in Universal Dependencies (UD) treebanks, labels their licensing environments, and perturbs these into minimal pairs. Evaluating 12 multilingual and 2 monolingual models alongside n-gram baselines, we find that multilingual models struggle most with lower-resource
languages and with pairs that are not n-gram solvable. We further examine the effects of NPI typology,
language family, licensing variability, and NPI-licenser distance. Data: https://github.com/arzuburcuguven/blimppi-data
Read the full abstract ▾
All authors Arzu Burcu Güven, Anna Rogers, Leonie Weissweiler, Joakim Nivre, Rob van der Goot
Machine Learning · Cross-listed from stat.ML
CorrFlow: Exact and Permutation-Equivariant Flow Matching on the Correlation Elliptope
Gautier Marti
Correlation
matrices combine nonlinear positive-definiteness constraints with a simultaneous row-column permutation
symmetry. Their generation therefore involves two distinct problems. First, every output must remain on the correlation elliptope, irrespective of how its variables are ordered. Second, the generator must learn the domain-specific distribution occupied by empirical data inside that valid space. Existing financial image generators mix these problems: they learn ordered
matrix entries and repair invalid draws afterward. We introduce CorrFlow, which separates them. A generalized Fisher transform maps full-rank correlation
matrices bijectively to unconstrained coordinates, guaranteeing valid outputs without projection. A graph flow-matching network learns the empirical distribution in those coordinates and is permutation equivariant: reordering the variables only reorders the output. We compare a conditional CorrGAN reimplementation, image
diffusion, and CorrFlow on $80\times80$ Taiwan equity
matrices using five seeds per neural method. CorrFlow produces 100% valid draws without repair and obtains feature sliced-Wasserstein distance $.712\pm.141$, versus $4.339\pm.252$ for CorrGAN and $4.610\pm.191$ for image
diffusion after their invalid outputs are projected back to the correlation space. It also improves this distance over historical lookup, while lookup retains higher local coverage. These results support the combined construction under the evaluated training budget; they do not establish superiority over fully tuned image models or isolate the contribution of each component.
Read the full abstract ▾