Peer Revenue Past citations are no guarantee of future returns

Papers

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

Computer Science

cs Watch Computer Science

Sunday 4 October

Computer Vision

Hierarchy-GBP: Accelerating Factor Graph Inference via Abstraction and Recovery

Yuzhou Cheng, Tom Yates, Ignacio Alzugaray et al. (+3 more)

Gaussian Belief Propagation (GBP) is a distributed inference algorithm that passes messages in graphical models, making it attractive for scalable spatial intelligence. However, we find GBP most effective locally: it rapidly smooths message errors that vary sharply between neighbor variables, but corrects global errors across distant graph regions incrementally through long-range message propagations. We propose Hierarchy-GBP (H-GBP), an iterative, two-stage framework that accelerates GBP by first solving these global errors with a coarse graph approximation (abstraction) and projecting the results back to the original graph (recovery), then refining the remaining local errors with GBP. We prove H-GBP convergence to the optimum by deriving the combined matrix operator of our abstraction and recovery steps and analyzing its spectral radius. Experiments on linear sparse graphs show that H-GBP converges fundamentally faster than standard GBP. Moreover, we validate H-GBP on two important spatial problems: Pose Graph Optimization (PGO) and Bundle Adjustment (BA). H-GBP markedly accelerates large-scale PGO and achieves state-of-the-art runtime across all tested BA scales.
Read the full abstract

All authors Yuzhou Cheng, Tom Yates, Ignacio Alzugaray, Danyal Akarca, Pedro A. M. Mediano, Andrew J. Davison

Computer Vision

Visual-Invariance-Augmented Feature Optimal Alignment for Transferable Adversarial Attacks against Closed-Source MLLMs

Xiaojun Jia, Simeng Qin, Yiming Li et al. (+5 more)

Multimodal large language models (MLLMs) remain vulnerable to transferable adversarial examples, especially in black-box settings where only open-source surrogate models are accessible. Existing targeted transfer attacks mainly align adversarial and target samples using global image-level features, such as encoder [CLS] embeddings. However, such coarse alignment insufficiently exploits patch-level visual structures, limiting transferability across heterogeneous closed-source MLLMs. We propose IAU-FOA, a visual-invariance-augmented feature optimal alignment attack with adaptive unbalanced transport, to improve targeted transferability against closed-source MLLMs. IAU-FOA aligns adversarial and target samples at both global and local levels: a cosine-based objective narrows their global semantic gap, while patch tokens are clustered into compact local patterns and matched through optimal transport for fine-grained feature alignment. Balanced optimal transport enforces fixed marginal masses even for local clusters without reliable counterparts, potentially introducing misleading alignment gradients. We therefore introduce confidence-adaptive unbalanced transport to relax these constraints for weakly matched clusters, aiming to reduce unreliable local alignment and improve adversarial transferability. We further study the effect of input transformations and propose visual-invariance augmentation, which applies bidirectional pixel-intensity rescaling and per-channel white-balance adjustment to simulate exposure, contrast, illumination, and color-temperature variations. This strategy encourages adversarial perturbations to generalize across different visual encoders. Extensive experiments on open-source and closed-source MLLMs show that IAU-FOA consistently outperforms state-of-the-art transferable attack methods. Code is available at https://github.com/jiaxiaojunQAQ/IAU-FOA.
Read the full abstract

All authors Xiaojun Jia, Simeng Qin, Yiming Li, Jie Liao, Sensen Gao, Ke Ma, Yang Liu, Xiaochun Cao

Artificial Intelligence

The GenAI4IDN Benchmark 3.0 - a Public Tool to Assess Generative AI Tools for the Design of Interactive Digital Narratives

Hartmut Koenitz, Jonathan Barbara and Mirjam Palosaari Eladhari

This paper presents GENAI4IDN Benchmark 3.0, the third iteration of an evaluation framework to assess Generative AI tools for creating Interactive Digital Narratives (IDNs). Moving beyond manual testing, this iteration introduces AI-assisted evaluation through a publicly accessible web application (https://genai4idn.com), enabling the community to run benchmarks on demand, add new models, and propose new tasks. The revised evaluation framework is "blinded" to avoid model-bias, and can handle complex media such as music, videos, and full IDNs that previously required human raters. A significant addition - responding to concerns raised during ICIDS 2025 - is the addition of fact-checking and bias detection with dedicated tasks and rubrics, validated by human raters with lived experience in the depicted contexts. Findings from a diverse range of models report on maturing creative capabilities while observing runaway thinking and overzealous safety filters as limitations. Fact-checking reliably caught subtle historical inaccuracies, anachronisms, and fabricated claims while the bias rater consistently exposed structural assumptions, tropes, and marginalized group erasures.
Read the full abstract

Machine Learning

Delay-coordinate reconstruction and conditional-moment causal diagnostics in stochastic systems

Jun Ohkubo

Partial observation and delay-coordinate reconstruction are rooted in deterministic dynamical-systems theory, whereas there are many systems with intrinsic stochasticity. We propose a conditional-moment interpretation of delay-coordinate reconstruction for stochastic systems, in which a delay vector is used to reconstruct conditional moments of a future distribution rather than a unique future sample path. Two complementary arguments motivate this viewpoint. First, the probability density of a stochastic differential equation obeys a deterministic Fokker-Planck equation and, under certain assumptions, is represented by an infinite deterministic hierarchy of moments. Hence, a finite-moment closure suggests a Takens-like finite-dimensional approximation. Second, a discussion based on the Koopman operator theory clarifies that the time evolution of an observable in the Mori-Zwanzig formalism yields a conditional expectation in stochastic systems. Then, the orthogonal "noise" term in the coefficient-space Mori-Zwanzig equation vanishes in the stochastic cases; this result is consistent with the moment-based argument. As an application of this stochastic delay-reconstruction viewpoint, we revisit convergent cross mapping (CCM) for diagnosing certain causal relationships. Although CCM based on the embedding theorem cannot generally be applied to stochastic systems, it is possible to examine certain types of causal relationships by using conditional moments. Using coupled logistic systems with additive and multiplicative coupling mechanisms, we discuss how causal relationships are embedded in stochastic systems.
Read the full abstract

Computation and Language

Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection

Nallathambi Vethiappan, Derya Soydaner and Gijs Wijnholds

Visual entailment (VE) asks whether an image supports, contradicts, or leaves undecided a textual hypothesis. Strong results come from fine-tuning large vision-language models on labelled data, while zero-shot and hybrid approaches remain far behind. A VE hypothesis often bundles several visual claims, yet existing zero-shot methods reason over it as a single unit. We propose Atomic Visual Entailment (AVE), which decomposes the hypothesis into atomic facts, produces candidate predictions from both the full hypothesis and its facts using frozen vision-language models, and predicts the final label with a lightweight classifier trained only on how those candidates behave. We find that decomposition helps only when the hypothesis context is preserved: judging facts in isolation is worse than not decomposing at all. Full-hypothesis and atomic prediction make complementary errors, and learning which to trust recovers far more of that complementarity than majority voting, reaching 0.803 test accuracy on SNLI-VE without fine-tuning any vision-language model. AVE also localises the visual evidence behind its prediction without region-level supervision. These results suggest that learning which candidate prediction to trust can close much of the gap to fine-tuned systems, offering a practical alternative where fine-tuning a vision-language model directly would need more labelled data or compute than is available.
Read the full abstract

Emerging Technologies

Autonomous Receptor Tuning via Integral Feedback for Adaptive Molecular Communication Receivers

Yagiz Celhan and Murat Kuscu

Molecular communication (MC) uses molecules as information carriers among engineered cells and artificial devices in the Internet of Bio-Nano Things. MC receivers commonly employ ligand receptors, whose discrimination between the two received concentration levels of binary signaling depends on their position relative to the dissociation constant. Changes in the transmitter-receiver distance, ligand degradation, and interference can shift both levels below or above the dynamic range of fixed receptors, impairing detection. Previous work restored performance by externally tuning the dissociation constant to the geometric mean of the estimated received levels. In this paper, inspired by receptor adaptation in living cells, we investigate a receiver whose intracellular antithetic integral feedback network senses the receptor occupancy and autonomously tunes the dissociation constant through a modulator molecule. For constant received levels and equiprobable bits, we prove that, without knowing these levels, the network sets the dissociation constant at equilibrium to their geometric mean, which minimizes the bit error probability (BEP) for any fixed receptor cooperativity. We derive the gain, bandwidth, noise, and stability boundary of this feedback loop. The random bits averaged by the network induce self-noise, which sets an error floor independent of the network molecule count and, balanced against the channel-tracking error, yields a loop-speed design rule. For received levels varying fivefold over a coherence time of 24 loop time constants, adaptation reduces the BEP 61-fold for non-cooperative and 410-fold for cooperative receptors, compared with the best non-adaptive receiver. This benefit grows with the coherence time and peaks at moderate variation depths.
Read the full abstract

Numerical Analysis · Cross-listed from quant-ph

Ground state energy estimation and preparation by the truncated power method

Savero Lukianto Chandra, Daochen Wang and Carolyn Zhang

The Guided Local Hamiltonian Problem (GLHP) is the problem of ground state energy estimation and preparation given a guiding state having good overlap with the true ground state. The problem is known to be BQP-complete for small error but is classically tractable for larger error. In this work, we identify an orthogonal regime in which the problem is classically tractable, namely when the true ground state $\psi$ has polynomially bounded $\ell_1$-norm. In particular, we show that the Truncated Power Method (TPM), a classical algorithm introduced by Kirby et al. in the quantum context, has a runtime that is polynomially comparable to quantum algorithms for GLHP in all relevant parameters, except for an extra factor of $\|\psi\|_1^2$. In other words, if $\|\psi\|_1^2$ is polynomially bounded, then TPM dequantizes quantum algorithms for GLHP.
Read the full abstract

Systems and Control · Cross-listed from math.DS

On global Lyapunov functions being Morse

Wouter Jongeneel

Lyapunov functions have been a cornerstone of dynamical systems theory ever since the early 1900s. Moreover and akin to Morse theory, Lyapunov functions have been the key in linking topology to dynamical systems theory. In this note we continue along these lines. Suppose that some vector field on Euclidean space has a unique equilibrium point that is globally asymptotically stable (GAS). Under these conditions, it is known that there is always a smooth Lyapunov function to certify stability. One may wonder if this function can always be chosen to be Morse. We show that if the equilibrium point is hyperbolic, this is indeed true. Moreover, we show that if this would be true in general, then the generalized 4-dimensional smooth Poincaré conjecture (SPC4) must be true. Even so, we provide an example that is GAS, but does not admit a Morse Lyapunov function, closing this route of proving SPC4.
Read the full abstract

Artificial Intelligence

A Framework for Automated Multi-Source Satellite Data Analytics and LLM-Based Report Generation

Hind Yousif Alhammadi and Isam Mashhour Al Jawarneh

This paper presents the workflow for building an automated ArcGIS Pro tool using ArcPy to extract the Land Surface Temperature (LST) from Landsat 7, 8 and 9 datasets. The tool eliminates the need for manual band selection and repetitive raster computations by automating the multi-step workflow of radiometric calibration, NDVI-based emissivity correction, and thermal conversion. In addition to supporting batch and single-scene processing, the tool has an optional Large Language Model (LLM) for statistical result interpretation and reporting. Depending on batch size, the tool reduced the processing time from around 11-58 minutes when done manually to around 4-11 minutes using the tool. We tested the tool with data from Ras Al Khaimah (RAK) in the UAE, and the LST obtained for Ras Al Khaimah ranged from approximately 25C to 50C, demonstrating an accurate LST mapping compatible with the weather conditions of RAK. In summary, our tool reduces human errors and improves processing accuracy and efficiency for thermal and environmental remote sensing applications, in addition to providing an interactive LLM-based interface for result interpretation.
Read the full abstract

Software Engineering

UndoBench: Separating Task Competence from Recovery Capability in Tool-Using AI Agents

Dolly Sah, Tanmay Sah, Harshul Jain et al. (+1 more)

Tool-using AI agents are increasingly deployed across enterprise software systems, yet widely used benchmarks primarily evaluate nominal task completion, conflating baseline planning competence with operational fault recovery. We introduce UndoBench, a benchmark spanning 36 base workflows and 36 fault scenarios across 8 enterprise domains, decoupling task competence from recovery capability via counterfactual paired trials under identical seeds alongside wire-level effect-history and environment-state oracles. On 12 held-out TEST workflows across two open-weight models, two frameworks, and three recovery paradigms (5,760 executions / 2,880 paired trials) in the frozen lost-acknowledgment study, nominal competence reached 83.54% while conditional recovery success rate (CRSR) fell to 46.72%, with naive retry producing duplicate external effects in 53.33% of trials. Extensions to commercial API models reproduced this competence-recovery separation. Evaluations across complementary execution boundaries show that recovery is phase-dependent: before mutation, methods perform similarly without duplicate effects among capable trials; during partial mutation, naive retry, per-call idempotency, and zero-privilege journaling collapse on the evaluated composite workflows; after commit but before acknowledgment, verification and server-side idempotency substantially improve safety. These findings demonstrate that evaluating nominal completion alone masks critical, phase-dependent recovery vulnerabilities in autonomous agents.
Read the full abstract

All authors Dolly Sah, Tanmay Sah, Harshul Jain, Tanya Sah

Information Retrieval

SCOUT: Supply-Aware Cold-Start Proactive Query Suggestion for Travel Search

Hao Li, Shashank Reddy, Kedar Bellare et al. (+2 more)

Generative query suggestion, powered by Large Language Models (LLMs), has become increasingly popular in search and conversational systems to reduce user friction and guide intent formulation. Existing approaches align suggestions with user preferences (e.g., clicks or conversions). This works for open-ended applications like chatbots and personal assistants, where the result space is unconstrained or historical user free-text queries are abundant. However, applying these methods to travel search presents two limitations. First, travel search is fundamentally constrained by physical inventory; a query (e.g., "romantic beachfront villa") may yield abundant results in Bali but few in Tokyo, so aligning with user preferences is not by itself grounded in what can be offered. Second, travel platforms traditionally rely on faceted search interfaces with no free-text queries. This creates a cold-start problem: without historical query logs there is no demand-side data for alignment, and without a seed query at request time, suggestions must be generated proactively from structured context alone. To address these challenges, we propose SCOUT, a bootstrapping framework for supply-aware proactive query suggestion. SCOUT overcomes the data gap by substituting missing demand-side user feedback with supply-side system feedback. It treats the search engine as a reinforcement learning environment, deriving a dense reward from the production reranker's query-listing match scores, and optimizes the policy with Group Relative Policy Optimization (GRPO). SCOUT improves inventory match rate (IMR@18) by 12.3% while preserving diversity, matching a compute-intensive best-of-8 policy at zero marginal inference cost and making supply-aware suggestion deployable on a real-time travel search path.
Read the full abstract

All authors Hao Li, Shashank Reddy, Kedar Bellare, Ashish Jain, Stephanie Moyerman

Machine Learning · Cross-listed from stat.ME

Causal Lag Structure Discovery in Confounded Time Series via Orthogonalized Adaptive Estimation

Hong Kiat Tan, Isaac-Neil Zanoria, James Chen et al. (+2 more)

Finding which variables cause which others in multivariate time series, and at what lags, is central to science and policy, yet existing methods force a choice between flexible confounder adjustment, data-driven lag selection, and inference that controls the false discovery rate (FDR). ORACLE-VARX does all three in one pipeline. First, double/debiased machine learning (DML) removes nonlinear confounder effects from the outcomes and the lagged series. Second, adaptive causal lag estimation (ACLE) picks the lag order at each time step by sequential significance tests, tracking regime changes. Third, entry-wise $z$-tests with Benjamini–Hochberg correction select directed edges at a target FDR. We prove that in each rolling window, the debiased coefficients are asymptotically normal around a window-averaged target, so their $z$-tests are asymptotically valid. On a synthetic benchmark with time-varying structure and nonlinear confounding, ORACLE-VARX (LightGBM) tracks the true lag order best (RMSE $0.96$ vs $1.1$–$1.5$), has edge FDR $0.047$, close to PCMCI ($0.045$) and below VAR ($0.129$) and VAR-LiNGAM ($0.187$), and forecasts better than all three. On nine U.S. sector ETFs with macroeconomic confounders, it yields interpretable causal graphs whose lag order rises in high-volatility regimes.
Read the full abstract

All authors Hong Kiat Tan, Isaac-Neil Zanoria, James Chen, Haoyang Lyu, Mihai Cucuringu

Data Structures and Algorithms

Prime factorisation of stable-matching instances: uniqueness, simultaneous products, and an exact census

Yoshiteru Ishida

Every balanced instance of the stable marriage problem with strict complete preferences has a unique finest partition into prime blocks, and that single partition simultaneously factors three different structures: the reachable execution digraph as a Cartesian product, the proposal-prefix antimatroid as a direct sum, and the stable-matching lattice as a direct product. The converse fails, and fails at every size from two on: two explicit families share the identical Boolean-cube execution while one is maximally decomposable with a single stable matching and the other is prime with n. Uniqueness yields an exact census, a recursion counting the prime instances at every size, under which exactly 88,478,208 of the 110,075,314,176 profiles with four agents on each side are decomposable and the decomposable fraction is asymptotically n! / n^(2n). The blocks are characterised as the square components of the mutual-rank filtration, so the partition is computable in polynomial time and the factorisation is a tool rather than only a fact.
Read the full abstract

Artificial Intelligence

DREAM: Dynamic Resolution Assignment For Multimodal Multi-agent Debate

Khanh-Binh Nguyen, Van Dai Do, Tien Anh Nguyen et al. (+2 more)

Multi-agent debate (MAD) has emerged as an effective paradigm to improve the reasoning capabilities of large language models (LLMs) and is increasingly being extended to multimodal settings. However, existing multimodal MAD frameworks typically expose agents to the same fixed visual input, ignoring substantial variation in the visual scale needed across samples and agents. In addition, these frameworks frequently suffer from groupthink, a phenomenon where agents prematurely abandon correct deductions to conform with confident but hallucinated peer responses. To address these bottlenecks, we introduce DREAM (Dynamic Resolution Assignment For Multimodal Multi-Agent Debate), which operates via two core components: (1) Dynamic Resolution Assignment, a zero-shot probe round where agents test multiple resolutions, quantify uncertainty using Average Normalized Log-Likelihood (ANLL), and use an adaptive threshold to assign each agent to its empirically optimal resolution; (2) Uncertainty-Guided Rollback Aggregation counters groupthink by tracking each agent's uncertainty over rounds and restoring early low-uncertainty answers overridden by group pressure. On six multimodal datasets, DREAM improves the accuracy-token trade-off over multi-agent debate baselines by 1.5-3.2% accuracy without dataset-specific tuning.
Read the full abstract

All authors Khanh-Binh Nguyen, Van Dai Do, Tien Anh Nguyen, Svetha Venkatesh, Hung Le

Sound

SEA-LM: Egocentric Spatial Audio Understanding for Wearable Microphone Arrays

Sonal Kumar, Sinan Hersek, Artem Dementyev et al. (+6 more)

Embodied, ego-centric intelligence fundamentally requires the ability to comprehend spatial audio within complex environments. While large audio-language models excel at mono-channel reasoning, they lack spatial awareness, discarding critical spatial cues that enable sound localization and that can improve the disentanglement of overlapping sound sources. To address this, we present SEA-LM, a Spatial Audio Understanding model. First, we introduce FOACODER, a layout-flexible spatial audio encoder trained on source localization and ego-centric voice activity detection objectives to encode First Order Ambisonics derived from variable-count, variable-position smart-glasses arrays via beamforming. We train a Multimodal Large Language Model (MLLM) to understand these spatial audio embeddings through a two-stage curriculum spanning six tasks, including sound localization and spatially selective transcription in settings with multiple speakers and overlapping sounds. To prevent the transcription outputs from dominating the next token prediction loss and overwhelming the direction predictions, we introduce a Spatio-temporal Weighted Cross-Entropy Loss. On our evaluation set, SEA-LM achieves lower azimuth and elevation MAE, higher temporal IoU, lower external-source hallucination and missing-source rates, and lower WER on most transcription tasks than compared baselines, while remaining robust across 1,211 smart-glasses array configurations with 4 to 9 microphones.
Read the full abstract

All authors Sonal Kumar, Sinan Hersek, Artem Dementyev, Mengzhen Pan, Ishan Chatterjee, Anurag Kumar, Ramani Duraiswami, Dinesh Manocha, Andrea Colaco

Computer Vision

Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation

Team Kandinsky, Julia Agafonova, Bulat Akhmatov et al. (+85 more)

We present Kandinsky 6.0 Video, a family of foundation diffusion models for synchronized text-to-audio-video generation, comprising Kandinsky 6.0 Video Lite (3B parameters) and Kandinsky 6.0 Video Pro (29B parameters). Both models generate 5-second video clips with synchronized 44 kHz audio, including lip-sync, in text-to-audio-video (T2AV) and image-to-audio-video (I2AV) modes; a built-in super-resolution model raises the output resolution to Full-HD (1920$\times$1080). Building on the video generation capabilities of Kandinsky 5.0, Kandinsky 6.0 Video employs a dual-stream CrossDiT architecture that connects a pretrained video stream and a newly trained audio stream through bidirectional cross-attention for temporal and semantic alignment. Our continuous pretraining strategy first trains the audio stream from scratch on large-scale audio corpora and then trains both streams jointly on paired audio-video data while preserving unimodal fidelity; pretraining is followed by supervised fine-tuning, reinforcement-learning-based post-training, and distillation. In side-by-side human evaluation, Kandinsky 6.0 Video Pro clearly outperforms its predecessor, Kandinsky 5.0 Video Pro, and remains competitive with leading audio-video generation models, particularly in speech quality. To accelerate open research and deployment in multimedia generation, we release the code, model checkpoints, and diffusers integration under the MIT license.
Read the full abstract

All authors Team Kandinsky, Julia Agafonova, Bulat Akhmatov, Mikhail Aksyutin, Grigorii Alekseenko, Anastasia Aliaskina, Olga Androsova, Vladimir Arkhipkin, Anna Averchenkova, Alexander Belykh, Serafima Bocharova, Sofiya Bogakovskaya, Anton Bukashkin, Mark Bulygin, Kirill Buzygin, Irina Cheremnykh, Kirill Chernyshev, Mikhail Chernyshov, Vladimir Chernyy, David Chikovani, Georgy Daniltsev, Denis Dimitrov, Anna Dmitrienko, Vladimir Dokholyan, Sergey Emelyanov, Dmitry Ermilov, Georgii Fedorov, Polina Gavrilova, Nikolai Gerasimenko, Aleksandr Gordeev, Andrey Inozemtsev, Andrei Ivaniuta, Alexander Ivanov, Mikhail Karaev, Anastasiia Kargapoltseva, Ivan Kirillov, Nikita Kiselev, Valeria Kobenko, Yury Kolabushin, Denis Koposov, Anatoly Korobov, Vladimir Korviakov, Kirill Kozlov, Denis Krzhivokolskiy, Konstantin Kuklev, Alexander Kunitsyn, Sergey Kuzin, Vladislav Lakhtionov, Alexey Letunovskiy, Maxim Litvinov, Alexander Lyulkov, Georgy Makarov, Kirill Malakhov, Egor Malykh, Mikhail Mamaev, Dmitrii Mikhailov, Polina Mikhailova, Ivan Mikheev, Elizaveta Muromtseva, Nikolai Nazarkin, Tatiana Nikulina, Lev Novitskiy, Stanislav Onuchin, Nikita Osterov, Denis Parkhomenko, Anatoliy Parpara, Vladimir Polovnikov, Konstantin Reznikov, Azat Saginbaev, Nikita Samsonov, Alexander Sentsov, Nikita Shaimov, Artem Sherstyuk, Andrey Shutkin, Egor Silvestrov, Bulat Suleimanov, Matvey Suprunov, Sergey Taranov, Irina Tolstykh, Tatiana Trofimuk, Ilya Trushkin, Aleksandra Tsybina, Olga Varlashina, Viacheslav Vasilev, Ilya Vasiliev, Eugeny Vilisov, Sergey Yakubson, Konstantin Zakharov

Robotics

Emergency Obstacle Avoidance Maneuvers in Differential-Drive Mobile Robots

Kautik Mandve and Nilay Kant

Emergency obstacle avoidance requires a mobile robot to brake or change its heading within the available distance. This paper investigates the dependence of maneuver performance and odometric error on approach speed for a differential-drive robot. A footprint-clearance analysis and a normalized wheel-motion index provide a kinematic description of five braking and turning maneuvers. The principal experiment comprises 540 block-randomized trials on tile, of which 539 are retained, at commanded approach speeds of 45, 55, and 65 cm/s. An overhead camera provides an independent pose reference. When encoder, gyro, and camera heading changes are evaluated over complete motion records, the mean encoder discrepancy increases by 4.7-5.1 degrees for the two reverse-spin maneuvers between the lowest and highest speeds; the arc and brake-assisted pivot change by less than 0.4 degrees. Larger encoder discrepancy is associated with lower avoidance success after adjustment for distance, speed, maneuver, day, and turn direction. Gyro discrepancies remain approximately 3 degrees for the reverse spins. Estimated reaction distances for 90% success and their uncertainty quantify the maneuver trade-offs. The results support speed-specific empirical characterization of emergency maneuvers and distinguish encoder error from recovery motion and measurement-window mismatch.
Read the full abstract

Human-Computer Interaction

Exploring the Effects of Personality in Human-Agent Interactions: A Study on User-Agent Synchrony with Human-based Vocalics

Kai Alexander Hackney, Jhonathan Sora-Cardenas, Aibek Musaev et al. (+1 more)

User trust is paramount in human-agent interactions, as it allows users to feel comfortable being themselves around an agent. The process of building user rapport starts in how an agent was designed, from its modality to the setting, to any of the many features and characteristics that can be tailored. All these aspects can affect whether users will be able to properly interact with the virtual agent and achieve the intended purpose. One such feature that is critical in human-human interactions is personality. A person's personality can strongly influence whether those that interact with them perceive them as trustworthy. This study used virtual agents generated from human voices with known personality traits to evaluate user perceptions. We found that extraverted agents were deemed to be more likeable by users, and that there was no significant effect of user-agent synchrony on user perceptions of the agent. In addition, it was found that user personality, without regard for agent personality, affected user perceptions of the agents. Our observations suggest that user perceptions may depend more on agent-topic synchrony than user-agent synchrony, and contribute to the broader community with considerations for the design of effective human-agent interactions.
Read the full abstract

All authors Kai Alexander Hackney, Jhonathan Sora-Cardenas, Aibek Musaev, Pedro Guillermo Feijóo-García

Distributed Computing

Handling Missing Data in Performance Portability Studies

Ami Marowka

Missing data is a common challenge in many real-world datasets, often leading to biased results or reduced accuracy. Missing data is a pervasive problem in performance efficiency datasets, particularly in high-performance computing (HPC) performance portability studies. Performance portability scores rely on complete sets of performance efficiencies as inputs to their respective metrics; however, missing values can arise due to incomplete benchmarking, hardware constraints, or implementation gaps. Many established performance portability metrics lack mechanisms for handling missing inputs, which can result in incomplete analyses or the inability to compute scores altogether. In this study, we evaluate a range of imputation methods and algorithms for addressing missing performance efficiencies. We present illustrative examples for each approach and apply them to real-world HPC performance portability datasets. Our results highlight the trade-offs between accuracy, robustness, and computational cost, offering practical recommendations for researchers seeking to mitigate the impact of missing data in performance portability assessments.
Read the full abstract

Robotics

Building A Multi-Sensor Platform For Autonomous Driving Research: Challenges and Lessons Learned

Paulo Ricardo Marques de Araujo, Eslam Mounier, Qamar Bader et al. (+2 more)

This paper reports on the challenges encountered and lessons learned during the development and deployment of a flexible multi-sensor platform for autonomous driving research. It aims to serve as a reference for researchers developing new multi-sensor systems. As the need for reliable, diverse datasets increases, novel environments and sensing configurations are essential to tackle real-world operational challenges. Consequently, many research groups create custom multi-sensory data collection platforms, where fundamental issues, such as mechanical design, sensor calibration, power management, and time synchronization, arise regardless of sensor types. We reflect on these challenges and share key insights to guide future platform designs, enhancing reproducibility and robustness in autonomous vehicle testing. We also summarize the mitigation strategies we adopted, for instance, system-wide time synchronization using GNSS timing and NTP protocols, custom calibration routines for different sensor configurations, and design practices to improve reliability and data integrity during field deployments.
Read the full abstract

All authors Paulo Ricardo Marques de Araujo, Eslam Mounier, Qamar Bader, Emma Dawson, Aboelmagd Noureldin

Computer Vision

EchoDino: A pediatric foundation model for transferable echocardiographic analysis across the lifespan

Sheng Cheng, Donnchadh M. O'Sullivan, Daniel J. Penny et al. (+3 more)

Echocardiography is the most widely used cardiac imaging modality, yet interpretation demands integrating visual evidence across global anatomy, localized structures and dynamic cardiac motion. Machine-learning models have automated individual tasks, but they are typically built for a single purpose and depend on expensively labeled datasets - a barrier particularly acute in pediatric care, where data are scarce and anatomy changes with age. Here we present EchoDino, a self-supervised foundation model for echocardiography, created by adapting the DINOv3 framework to 3.7 million frames from 1.7 million unlabeled pediatric echocardiography videos. With its encoder frozen, EchoDino produces representations that capture global context, local anatomy, and dense spatial detail. We introduce Motion-biased Entropy Maximization Sampling (MEMS) to select the most informative frames for video-level analysis. Across nine pediatric and adult datasets, EchoDino outperformed strong baseline models, raising view-classification accuracy from 0.609 to 0.889 and the area under the receiver operating characteristic curve for structural-heart-disease detection from 0.811 to 0.872, while also cutting age-estimation error from 3.857 to 1.389 years, achieving the best segmentation accuracy and lowering ejection-fraction errors. By generalizing from label-free pediatric data to adult echocardiography, EchoDino offers a versatile foundation for cardiac image analysis across the lifespan.
Read the full abstract

All authors Sheng Cheng, Donnchadh M. O'Sullivan, Daniel J. Penny, Craig G. Rusin, Minh B. Nguyen, Devika Subramanian

Machine Learning

Your Unlearning Gives You Away: Identifying Erased Concepts in Diffusion Models

Kaiyuan Deng, Yuchen Li, Yang Xiao et al. (+3 more)

Existing attacks on unlearned diffusion models assume that the erased concepts are known in advance and focus on recovering them. In practice, however, model providers may not disclose which concepts have been removed, and even with access to the original base model, an adversary may still lack a clear target to attack. In this paper, we aim to answer the following critical but overlooked questions: which concepts have been erased from the model, and how many have been erased in total? To this end, we present Tracer, a framework that rapidly and accurately identifies erased concepts and estimates their number. Tracer efficiently identifies erased concepts without generating and classifying images. By combining lightweight spectral analysis of weight footprints, it enables efficient search over large candidate vocabularies. To distinguish multiple erased concepts, we introduce a footprint coverage objective that guides sequential discovery. Tracer estimates the number of erased concepts by detecting a sharp decline in candidate confidence as the selected concepts account for the erasure footprint, without requiring labeled examples for calibration. The framework requires only lightweight linear algebra and limited forward probes, with no prior knowledge of the unlearning algorithm. Experiments across text-to-image and text-to-video backbones and diverse unlearning methods demonstrate that Tracer identifies erased concepts and estimates their number in seconds, achieving 150 to 137,000 times and 133 to 20,000 times speedups over MIA and brute-force search on image and video models, respectively, with substantially higher identification accuracy.
Read the full abstract

All authors Kaiyuan Deng, Yuchen Li, Yang Xiao, Bo Hui, Geng Yuan, Xiaolong Ma

Machine Learning

An LLM-in-the-loop RL Framework for Bioinformatics Feature Selection

Xinyuan Wang, Deepti Agrawal and Yanjie Fu

High-dimensional bioinformatics data, characterized by a large number of features relative to the number of samples, pose major challenges such as the “curse of dimensionality,” leading to overfitting, high computational cost, and poor generalization. Traditional feature selection methods often suffer from limited scalability and adaptability in such domains. We propose an LLM-in-the-loop reinforcement learning (RL) framework for bioinformatics feature selection, where the RL agent formulates feature selection as a sequential decision-making task, while the large language model (LLM) enhances the process in two ways: (1) guiding exploration through domain-informed advice, and (2) providing hybrid rewards that integrate data-driven performance with knowledge-driven evaluation. The LLM also produces explanations to improve interpretability for human experts without altering the RL policy update. Experiments on diverse bioinformatics datasets show that the LLM-in-the-loop framework outperforms baselines, achieves stable performance across downstream models, and converges faster than pure RL.
Read the full abstract

Machine Learning · Cross-listed from stat.ML

Gaussian Limits for SGD Without Stationary Moments

Xiaoli Li and Wei Biao Wu

Temporal dependence can separate the Gaussian approximation of stochastic gradient descent from its stationary moments. For unmodified least-squares SGD, we construct a design with standard Gaussian marginals whose stationary error has every positive moment infinite. Independent observations with the same marginals instead give finite stationary variance. Both regimes retain a Gaussian small-step limit. Our general theory establishes pathwise contraction from a finite second design moment, then uses score cancellation and localization to obtain stationary Gaussian and Ornstein–Uhlenbeck limits. Independent Gaussian regression errors yield an exact conditional Gaussian law and total-variation convergence under the same design integrability. Stronger design conditions identify a positive first-order total-variation constant and a deterministic covariance correction with $o(a)$ error. A scalar coverage expansion translates this correction into its inference consequence. Experiments examine distributional error, coverage, and calibration with dependent scores. Together, these results establish precise probability-law approximation beyond moment-based stationary analysis.
Read the full abstract

Machine Learning

When the Cross-Silo Federation Goes Offline: Continual Learning for Site Onboarding with Limited Unlabeled Data

Ahmadreza Eslaminia, Klara Nahrstedt and Chenhui Shao

An organization often holds too little labeled data to train a model that generalizes, and the records that would supply the rest sit with organizations that cannot release them. Cross-silo federated learning offers a way through, since participants exchange model parameters rather than records, but it ordinarily settles two aspects of the arrangement in advance, the participating sites and the classes the model can predict, and deployment can breach both. A new site joins after training, once the established sites have finished their engagement and gone offline, and its records arrive unlabeled, mixing conditions the model already recognizes with conditions no participant has observed. We present an autonomous three-stage procedure that expands the model entirely at the joining site: reconstruction experts screen for novelty, clustering separates the flagged records into candidate conditions, and class means describe the old classes, all inside one shared representation. Those classes were learned from records that never leave their owners, so the usual defenses against forgetting are unavailable, and the procedure supplies the evidence they would have carried from either of two dissimilar sources, prototypes held by the federation or records held by the joining site. On a real industrial condition-monitoring dataset, run end to end with no label consulted, either source holds old-class accuracy at 0.868 or above with forgetting at most 0.063, and the two differ by 0.021, so a configuration can be chosen by the disclosure it permits rather than the accuracy it delivers. Both keep old- and new-class accuracy in balance where every alternative we measure gives up one for the other, and both retain more of the old classes than distillation- and regularization-based baselines. The balance still holds with only 6 labeled records per arriving condition and 3 retained per old class.
Read the full abstract

Computation and Language

More Than Words: Compositional Tokenization for Efficient Language Models

Yuval Reif, Guy Kaplan and Roy Schwartz

Language models process and generate text sequentially in token units, and the tokenizer determines how much text each inference step covers. Under standard tokenization, a short English phrase such as "On the table." is usually produced as four separate predictions for the preposition (On), article (the), noun (table), and punctuation (.), where each consumes a sequence position and adds inference cost. We introduce CoBPE, a compositional tokenization approach that represents such phrases as a lexical base token (table) attached with a small set of reusable surface modifiers, composed in embedding space at input and predicted jointly at output. In controlled pretraining from scratch at 780M and 1.3B scales, CoBPE shortens sequences by 30% and improves average downstream performance by 1.2 points relative to standard BPE under matched training compute. Our results suggest that part of what is now expressed through token sequences can instead be modeled through structured representations, opening a broad design space for more token-efficient and capable language models.
Read the full abstract

Machine Learning · Cross-listed from stat.ML

Moment-Accurate Gaussian Mixtures for Constant-Step Stochastic Approximation

Xiaoli Li and Wei Biao Wu

Local Gaussian models of constant-step learning predict output variability and expected losses, but weak convergence alone does not justify these moment predictions. We establish moment-accurate Gaussian mixtures by matching stationary energy with local Ornstein–Uhlenbeck limits, ruling out quadratic tail mass invisible to weak convergence. For step size $a$, the second-order Wasserstein error is $o(\sqrt a)$, uniformly over invariant laws, using each law's actual root weights. The assumptions combine confinement, descent, finitely many hyperbolic equilibria and root continuity with finite-variance innovations. The result yields observable covariances, expected objective gaps and first-order mean shifts, while allowing singular covariances, compatible saddles and weights without a limit. For additive noise given by a fixed invertible transform of independent standardized Student $t_3$ coordinates, symmetry gives an order-sharp $\sqrt a$ smooth-test bound. Numerical transport calculations demonstrate the value of root-specific covariances; controlled SGD studies assess observable predictions across step sizes, batch sizes and model geometries.
Read the full abstract

Computation and Language

What Is a Repeated Token Worth? The Scaling Geometry of Multi-Epoch Pretraining

Yekun Chai and Haoyi Xiong

As pretraining increasingly repeats data, every run faces three questions: how many epochs to take, how that number should change with model size, and whether anything besides the epoch count matters. We answer them by pricing a repeated token against two references: one epoch on the same data, which gives its value, and fresh data at equal compute, which gives its cost. Against fresh data, the cost of repetition follows a single variable, the number of extra epochs divided by the unique tokens per parameter. Against the same data, a second epoch is worth nearly as much as a fresh one, and repeated tokens fall to half the value of fresh ones after a critical epoch count that grows with the training budget per parameter but hardly with model size. With unique data fixed, the predicted compute-optimal run grows model size and epochs together until loss stops improving, near the critical epoch count. The same variable accounts for the direction of size trends that appear to conflict: larger models tolerate fewer epochs when the corpus is fixed, from about 15 at 127M to 4 at 2B parameters, but not when unique data grow with the model. Counts alone do not determine loss: at identical counts, replaying shards consecutively raises loss by up to 0.46~bits per byte, concentrating repeats on fewer samples also raises it, lower-entropy sources degrade faster with repetition, and re-tokenizing repeats helps only under heavy repetition. These results offer an empirical guide to pretraining when unique data, rather than compute, are the binding constraint.
Read the full abstract

Machine Learning

ColdDDI: Evaluating Knowledge Utilization in Cold-Start Drug-Drug Interaction Prediction

Jiheng Liang, Chen Zhao, Di Wu et al. (+4 more)

Cold-start drug-drug interaction (DDI) prediction tests whether models can identify clinically significant interactions for drugs without training-time interaction history. Existing benchmarks mostly report aggregate edge-prediction scores, leaving a key evaluation question unanswered: when models receive molecular, textual, or knowledge-graph (KG) evidence, do they actually use the evidence that pharmacologically supports the interaction? We introduce ColdDDI, a reconstructible diagnostic benchmark built from DrugBank 5.1.13, with 1,900 approved small-molecule drugs and 565,731 positive DDI pairs. ColdDDI evaluates pairs with zero, one, or two unseen drugs. It also annotates each interaction by whether it changes drug exposure or drug effect, and by whether the biomedical knowledge graph contains shared enzymes, transporters, or targets that can plausibly mediate the interaction. These annotations separate evidence availability from predictive dependence. We evaluate eight conventional DDI methods and 13 LLMs; for open-weight LLMs, we test five prompt patterns and use masking, drug replacement, and channel-sensitivity metrics to probe knowledge utilization. ColdDDI exposes that, in the hardest split where both drugs are unseen, the main performance divide is mediator availability. A fine-tuned 1B LLM recovers 89-93% of interactions with a shared enzyme, transporter, or target, but only 40-62% without such a mediator. More importantly, KG-provided evidence is not always used; several KG-augmented baselines change little when the shared mediator is masked or disrupted, whereas fine-tuned LLMs respond strongly to this intervention. Thus, ColdDDI evaluates knowledge utilization rather than knowledge access alone, showing where cold-start DDI models rely on mechanistic evidence and where they fail despite receiving it. Code is available at https://github.com/0217ljh/ColdDDI-NeurIPS2026.
Read the full abstract

All authors Jiheng Liang, Chen Zhao, Di Wu, Chenyang Bu, Yunpeng Hong, Xingquan Zhu, Yi He

Game Theory

Sender-Receiver Games with Payment Commitments

Ivan Geffner

We study sender-receiver games in which players commit, before the state is drawn, to payments contingent on the subsequently verified state and the receiver's action. Payments may be transferred or burned, serving as contractual guarantees or promised tips. We analyze equilibrium guarantees, welfare, and computational complexity. For arbitrary finite action sets, we characterize pessimistic utility guarantees under unilateral commitments. Sender-only and sender-first commitments have worst-equilibrium welfare equal to the best constant-action welfare. Simultaneous commitments can instead reduce welfare below every equilibrium of the original game. Neither simultaneous nor sender-first commitments always admit an efficient equilibrium. Receiver-first commitments always admit an efficiently computable welfare-maximizing equilibrium, but can also reduce worst-equilibrium welfare. For binary actions, we characterize their entire equilibrium payoff region. With binary actions, optimal receiver commitments are computable in linear time, whereas optimizing deterministic sender commitments is NP-hard. Privately randomized sender commitments admit an exact $O(n\log n)$ algorithm, with an optimal lottery over at most two schemes differing at only one state. This yields an additive approximation by deterministic commitments. For arbitrary finite action sets, optimal privately randomized sender payoffs equal Bayesian persuasion values: ordinary persuasion with burning alone, and persuasion with nonnegative compliance payments when transfers are allowed. These equivalences yield polynomial-time algorithms with polynomial-size commitment lotteries.
Read the full abstract