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

Systems and Control

Cooperative Adaptive Cruise Control with Constant Distance Gaps for Retrofitted Legacy Vehicles

Kay Massow, Emanuele Crisostomi, Ilja Radusch et al. (+1 more)

A Constant Distance Gap (CDG) spacing controller for Cooperative Adaptive Cruise Control (CACC) designed to increase the capacity of oversaturated signalized intersections is presented. In a departing platoon of production passenger vehicles, the controller maintains close inter-vehicle distances below 3 m within limited platoon lengths, using one-vehicle-look-ahead (OVLA) communication only. Extensive real-world experiments on heterogeneous, retrofitted legacy production cars demonstrate robust operation. To the best of our knowledge, this is the first experimental demonstration of such close distances under this combination of conditions. A CDG policy can substantially increase discharge rates at oversaturated signalized intersections when vehicles preserve their standstill spacing during startup, as shown in our previous simulation study. Because this benefit unfolds only at high penetration rates, it calls for retrofit-capable solutions deployable in existing combustion-dominated fleets. However, practical realization in such fleets remains an open challenge. The proposed MPC-based controller combines a first-order longitudinal prediction model with precise powertrain feedforward employing OEM-specific torque models, without exchanging confidential vehicle parameters. It avoids explicit platoon coordination to minimize air-interface load in dense traffic. To mitigate disturbance amplification and increase feasible platoon length at very small gaps, two controlled spacing-relaxation strategies are compared: (i) a micro-time-gap policy near the theoretical string-stability bound, and (ii) an acceleration-dependent spacing relaxation based on a virtual mass-spring-damper model. We introduce the Desired Acceleration Overshoot (DAO), a time-domain metric relating local actuator-demand amplification to achievable platoon length.
Read the full abstract

All authors Kay Massow, Emanuele Crisostomi, Ilja Radusch, Robert Shorten

Game Theory

Learning in Continuous Games from Pairwise Preference Feedback

Anas Barakat

We study learning in continuous games when players receive only pairwise preference feedback, revealing which of two actions is preferred but neither payoff values nor preference magnitudes. We first show that standard external regret and coarse correlated equilibria (CCE) are not identifiable from this ordinal information: the same sequence of play can incur zero and linear regret in two ordinally equivalent games, while the distributions that remain CCE across all cardinal representations consistent with the same preferences are exactly those supported on pure Nash equilibria. Motivated by this gap, we develop a first-order ordinal theory based on normalized unilateral preference directions, introducing an ordinal directional regret benchmark and corresponding equilibrium notions. We show that block-normalized pseudogradient dynamics achieve sublinear ordinal regret and, under additional structure, Nash-convergence guarantees. We then use a single-comparison estimator to implement these dynamics from finite pairwise comparisons. With one comparison per player and round, the resulting algorithm achieves sublinear finite-resolution ordinal regret against arbitrary opponent behavior and, in ordinal potential games, almost-sure last-iterate convergence to the Nash set. Our results provide regret, dynamics, and equilibrium guarantees directly from preference feedback without reconstructing cardinal utilities.
Read the full abstract

Computer Vision

Unmentioned Checklist Findings Change How Reinforcement Learning Appears to Improve Chest Radiograph Report Checking

Ali Vosoughi, Akhil Kasturi, Chenliang Xu et al. (+1 more)

Automated checks of radiology reports may rely on AI-generated checklists that leave findings unmentioned. We used reinforcement learning to train a vision-language model to fill in a 12-finding checklist from a chest radiograph without seeing the sentence under test; a separate checking model judged the sentence from the checklist. On held-out patients, a rule-based check and an independent medical checker, neither used in training, measured discrimination gains (Youden index) of 12.6% and 11.8%; only the rule-based check met the prespecified false-alarm criterion. Switching to the training format, which fixes finding order and enters unmentioned findings as absent, raised the training checker's measured gain and lowered the independent checker's, a prespecified comparison that yielded 6.2% (95% interval 2.0% to 10.5%) and, post hoc on held-out patients, 7.7%. Across 8 checking models, acceptance of a label-consistent negative statement about an unmentioned finding ranged from 1.0% to 97.0%. Labels were report-derived, not radiologist-adjudicated.
Read the full abstract

All authors Ali Vosoughi, Akhil Kasturi, Chenliang Xu, Axel Wismueller

Robotics

EvoMem-VLA: State-Evolution Memory for Long-Horizon Robot Manipulation

Yuheng Na, Zhide Zhong, Junjie He et al. (+7 more)

Most vision-language-action (VLA) models rely on current observations and lose task-relevant evidence once it leaves view, limiting performance on long-horizon, memory-dependent tasks. Existing efforts incorporate compressed historical features or sparse visual keyframes. However, isolated snapshots can leave the policy uncertain about what changed during past interactions and which action should follow. To overcome this limitation, we propose EvoMem-VLA, which constructs state-evolution memory by explicitly encoding and retaining observed changes between historical states. These change representations preserve evidence of interaction outcomes, allowing the policy to track task progress beyond isolated snapshots. Specifically, we introduce conditional delta tokenization to encode ordered frame pairs into directional, source-conditioned delta tokens, each associated with its corresponding state evidence. A shared VLM backbone supports task-adaptive routing: normal long-horizon tasks follow a direct action route, whereas multi-stage tasks use a subtask route that generates an executable subtask as an additional input for action generation. With a single jointly trained policy for each simulation benchmark, EvoMem-VLA achieves success rates of 80.7% on RMBench, 82.0% on RoboMME and 83.8% across four real-world tasks spanning two robot embodiments. These results represent substantial improvements over the previous state of the art in all three evaluation settings.
Read the full abstract

All authors Yuheng Na, Zhide Zhong, Junjie He, Junfeng Li, Haodong Yan, Jiaan Wang, Jiaguan Zhu, Yangyang Zheng, Tianyu Huang, Haoang Li

Computer Vision

Render to Reason: Novel-View Semantic Prediction Improves Spatial Understanding in VLMs

Yuqun Wu, Yao Xiao, Chuhang Zou et al. (+2 more)

Recent works augment Vision-Language Models with geometry features from pretrained 3D models, expecting that the geometric signal will boost spatial reasoning. However, we find that simply fusing geometry features and training on standard spatial QA yields only marginal improvements on high-level multi-hop tasks. We attribute this gap to a training-signal problem: standard spatial QA can be largely answered from visual features and language priors, so the geometry pathway receives weak gradients and fails to integrate with the visual features. To provide a training signal that requires geometry, we propose novel-view semantic rendering as an auxiliary training task that requires the model to predict the semantic layout of an unobserved viewpoint, inspired by humans' ability to mentally simulate novel viewpoints during spatial reasoning. This task encourages joint use of both pathways: geometry provides pose-dependent visibility, while vision provides semantic content. Our auxiliary task yields consistent improvements over the geometry-augmented baseline across all three benchmarks (up to +1.6 on VSI-Bench, +2.2 on ReVSI, +2.9 on our 3D-Point-QA dataset) and our full model surpasses prior open-source methods on VSI-Bench and on ReVSI. Project page: https://yuqunw.github.io/Render2Reason/.
Read the full abstract

All authors Yuqun Wu, Yao Xiao, Chuhang Zou, Shenlong Wang, Derek Hoiem

Computer Vision

Prism: Dynamic Sparse Attention for Native 2K Joint Video-Audio Generation Model Training

Shuyuan Tu, Qi Tian, Yinming Huang et al. (+8 more)

Natively training joint video-audio generation models at higher resolutions empowers them to learn richer visual details and sharper motion dynamics. However, full attention incurs quadratic cost and, as resolution increases, spreads attention over increasingly redundant tokens, diluting learning signals for informative content and disrupting pretrained priors. Existing sparse attention methods either target training-free acceleration or overlook the unique structure of joint video-audio data, where cross-modal interactions are inherently concentrated around sound-producing regions. To address this, we propose Prism, a dynamic sparse attention framework for natively training joint video-audio generation models at 2K. In particular, Prism organizes the token sequence into spatiotemporal macro-zones, enabling the attention structure to adapt to local content. For each zone, it estimates local information structure via video feature variance along the channel and feature norms from the audio-to-video cross-attention, jointly capturing how visual content varies directionally and how strongly audio influences each visual region. Based on these signals, Prism dynamically assigns a tailored block shape to each zone, applying finer partitioning along axes of rapid visual content variation and strong audio-visual coupling. This encourages tokens within each block to remain semantically coherent, allowing block-level features to capture both visual content and joint video-audio interaction patterns. Prism further adopts a hybrid block selection strategy to dynamically determine per-query sparsity. Experiments show that Prism achieves 2.5$\times$ training speedup compared to full attention, while surpassing it in generation quality.
Read the full abstract

All authors Shuyuan Tu, Qi Tian, Yinming Huang, Yue Wu, Xintong Han, Kaihang Pan, Weijie Kong, Jiangfeng Xiong, Jian-Wei Zhang, Zuxuan Wu, Yu-Gang Jiang

Robotics

RMMBench: A Comprehensive Benchmark for Robotic Mobile Manipulation

Huapeng Li, Fuxiang Feng, Jinqiu Fan et al. (+5 more)

Although the advancement of vision-language models (VLMs) has endowed robots with enhanced environmental understanding and task reasoning, a comprehensive evaluation methodology is important to advance the integration of VLMs in robotic navigation and manipulation. However, current benchmarks lack a comprehensive method to evaluate diverse robotic tasks, and evaluation metrics remain relatively constrained, making it difficult to assess the embodied capabilities of VLMs in a thorough and fine-grained manner. To address this issue, we propose RMMBench, an evaluation benchmark that requires robots to understand language instructions and perform long-horizon tasks in continuous spaces. RMMBench seamlessly integrates high- and low-level embodied tasks into a unified framework, constructing a "navigation-manipulation" task suite comprising 70 canonical task scenarios that range from localized manipulation to long-horizon composite navigation. The results reveal that leading VLMs still face major challenges in spatial localization when performing mobile manipulation tasks, and also highlight the necessity of enhancing the spatial perception capability of robots during long-horizon interactions. RMMBench can be accessed at https://mxxq-stack.github.io/rmmbench-project/
Read the full abstract

All authors Huapeng Li, Fuxiang Feng, Jinqiu Fan, Shuo Yang, Fengjiao Chen, Xuezhi Cao, Ran Song, Wei Zhang

Computer Vision

A Strong Baseline for Evaluating Vision Encoders in Multimodal Large Language Models

Yilin Yang, Jun-Tao Tang, Kengyi Wang et al. (+3 more)

Evaluating vision encoders requires metrics that reliably predict their downstream performance in multimodal large language models (MLLMs). Although recent studies have shown that cross-modal metrics can better capture such performance, unimodal metrics remain the dominant choice in practice. In this work, we revisit cross-modal evaluation of vision encoders through large-scale experiments. We identify important limitations in both the experimental design and methodological formulation of prior approaches. After addressing these limitations and introducing simple improvements, we propose RAVEL, a training-free method based on cross-modal nearest-neighbor retrieval. Despite its simplicity, RAVEL achieves state-of-the-art performance across our experiments, outperforming prior methods by a substantial margin. Our results demonstrate that simple cross-modal metrics, when evaluated under a careful and comprehensive setup, can provide a strong basis for evaluating vision encoders for MLLMs.
Read the full abstract

All authors Yilin Yang, Jun-Tao Tang, Kengyi Wang, Siyuan Su, Gaoyong Luo, Mingda Chen

Computational Complexity

A Dichotomy for Planar Graph Homomorphisms with Nonnegative Weights

Chenghua Liu and Boning Meng

We prove a complete complexity dichotomy for planar graph homomorphism counting with any fixed symmetric nonnegative matrix of arbitrary finite order, giving an explicit criterion for tractability. We also characterize exactly which fixed positive vertex weights preserve tractability, with both classifications extending from algebraic weights to fixed real weights in a prescribed exact representation. Our proof hinges on an entropy-based continuation argument: maximal logarithmic support identifies distance kernels as maximum-entropy completions, extending their positive definiteness throughout the parameter interval. This enables distance geometry to recover hidden product coordinates even when planar gadgets cannot distinguish colors; counting-hardness arguments then force the factors to be zero-field Boolean Ising interactions. The classification also yields complete tractability criteria for clock models, coupled Ising systems, and planar contractions of stoquastic imaginary-time kernels. All results have been formally verified in Lean 4.
Read the full abstract

Computer Vision

CleanMDM: Clean Motion Diffusion Model for Multimodal Motion Cleanup

Zhe Li, Shicheng Wang, Bowen Cai et al. (+1 more)

Motion capture data is rarely directly usable, as they typically exhibit missing segments, jitter, drift and contact artifacts. Traditionally, corrupted motions are cleaned by animators through the manual identification of keyframes from noisy motion, subsequent keyframe correction, and interpolation between corrected keyframes to reconstruct coherent motion. While the rise of generative motion models has made automatic cleanup feasible, most approaches operate as black box denoisers with limited controllability, making it difficult to preserve reliable segments or enforce specific user intents. Inspired by animation workflows, we present CleanMDM, a unified multimodal motion cleanup framework that formulates cleanup as masked conditional generation with plug-and-play conditions. This single model supports arbitrary combinations of noisy 3D motion, sparse 2D keyframes, sparse 3D keyframes, and text. This design enables both automatic cleanup without additional user annotation and controllable cleanup under multimodal guidance. To further improve motion realism, we incorporate the Latent Motion Quality Discriminator (LMQD) to better match kinematic distributions and reduce skating, jitter, and interpenetration artifacts, and we apply Mesh-Aware Contact Projection as a test-time optimization step to enhance contact and physical consistency. Experiments across multiple datasets demonstrate that CleanMDM consistently outperforms prior cleanup and generation baselines, and that low cost conditions (text and 2D keyframes) provide reliable controllability gains in multimodal cleanup scenarios.
Read the full abstract

All authors Zhe Li, Shicheng Wang, Bowen Cai, Huan Fu

Machine Learning

BeliefGraph-JEPA: Structured Latent World Models for Action-Conditioned Time Series

Yue Li, Kangqi Ni, Zhen Tan et al. (+1 more)

Action-conditioned time-series forecasting requires accounting for how future actions and exogenous forcings influence multiple targets through partially observed effects with different delays and persistence. Direct conditioning leaves the evolution and target-specific influence of these effects implicit in the predictor, while static relational graphs specify connections without tracking evolving effects. This motivates representing future-driver influence through structured latent states that evolve over the forecast horizon and route information to individual targets. We introduce BeliefGraph-JEPA, a structured latent world model that factorizes driver influence into typed latent-effect states. These states are rolled forward under future drivers and routed through a graph to target-specific nodes, forming the predictive base of a joint-embedding predictive architecture. A capacity-controlled residual supplements this base with direct driver information. On four multi-target clinical, agricultural, environmental, and industrial systems, the framework outperforms a range of pretrained and supervised known-future-covariate baselines. Matched controls isolate latent dynamics, future rollout, graph routing, and residual capacity; future rollout and graph-first residual routing improve forecasting across all four systems.
Read the full abstract

All authors Yue Li, Kangqi Ni, Zhen Tan, Tianlong Chen

Robotics

TUCO: Curating Simulation Demonstrations for Sim-to-Real Robot Policy Co-Training

Ning Zhu, Mengfei Zhao, Yikai Tang et al. (+5 more)

Simulation demonstrations can supplement scarce real-world data for robot policy co-training. However, the value of using data curation to actively select these demonstrations for sim-to-real co-training remains underexplored. Existing curation methods also lack a unified criterion for measuring trajectory-level utility and set-level coverage from closed-loop target behavior. To address these gaps, we present the first systematic study of data curation for sim-to-real robot policy co-training and propose Trajectory-level Utility and set-level Coverage Optimization (TUCO). TUCO uses influence functions to trace how each source demonstration affects target-domain scoring rollouts. Our key insight is that these effects can be decomposed into an overall contribution to target return and variation across rollouts, providing a common closed-loop basis for measuring trajectory utility and set coverage. We further propose a performance-aligned subset optimizer that combines these measures in a unified curation objective to reduce redundancy and select complementary demonstrations. Extensive experiments on RoboMimic and OmniReset establish the value of active simulation data curation for sim-to-real policy co-training and show that TUCO achieves state-of-the-art performance across single-simulator, sim-to-sim, and sim-to-real settings.
Read the full abstract

All authors Ning Zhu, Mengfei Zhao, Yikai Tang, Zhangyujie Sun, Peihao Li, Dongyue Ni, Jindou Jia, Jianfei Yang

Machine Learning

Task Vector Descent: Learning from Non-IID Batches

Anton Baumann, Jonas Hübotter, Zeynep Akata et al. (+1 more)

A central challenge in continual learning is to acquire new knowledge without forgetting what the model has already learned. This challenge appears in language model training when training data comes from various domain-, user-, or task-specific distributions that are encountered unevenly over time. In such settings, successive minibatches are temporally clustered by distribution instead of being sampled i.i.d. from the overall data mixture. Training on temporally clustered data induces a stability-plasticity tradeoff. Adapting the model to the active distribution can improve the model on the active distribution but may lead to a performance degradation on data it previously trained on. We find that this tradeoff intensifies with longer exposure to the same distribution. We therefore ask if the parameter displacement produced by such a sequence (the task vector) should be fully retained or applied partially. We compare applying the full displacement ($\lambda=1$) with partial integration, which scales the task vector by $\lambda$ before applying it to the continuing model and scales the optimizer state by the same coefficient. Across continual pretraining, pretraining from random initialization, supervised post-training, and reinforcement post-training, we find that intermediate values of $\lambda$ often improve average continuing-model performance relative to full integration, particularly after longer same-distribution sequences. In continual-pretraining experiments with both controlled streams and naturally defined adaptation sequences, task-vector scaling outperforms full integration at the matched learning rate, showing that its benefits are not reproduced by learning-rate scaling alone.
Read the full abstract

All authors Anton Baumann, Jonas Hübotter, Zeynep Akata, Andreas Krause

Machine Learning

No Concept Escapes the Audit: Auditing-Aware Unlearning for Verifiable Concept Erasure in Diffusion Models

Kaiyuan Deng, Yuchen Li, Gen Li et al. (+5 more)

Text-to-image diffusion models can generate prohibited content, which motivates concept erasure through machine unlearning. Most erasure methods intervene at the text interface, through prompt modification or localized updates to text-conditioning weights, and they are evaluated by what the model outputs for given prompts. Such evaluation cannot see what the network still encodes. Latent-space auditing, which bypasses text conditioning and probes the denoising network directly, shows that erased concepts remain recoverable from internal representations. We find that this also holds for methods built to be robust against adversarial prompts, and that the problem grows with the number of erased concepts. We propose Auditing-Aware Unlearning for Verifiable Concept Erasure in Diffusion Models (AVCE), a framework that grounds erasure in the model's latent representations. AVCE audits the embedding neighborhood of each concept and condenses the discovered vulnerable directions into an anchor at the weakest geometric point. It edits cross-attention and self-attention projections in closed form at this anchor, then fine-tunes the two pathways with pathway-level auditing losses, using orthogonal gradient projection to consolidate multiple concepts. Experiments on SD v1.5, SDXL, and Flux 1.0 across object, explicit-content, and artistic-style unlearning show that AVCE reduces attack success rates by 5.07x and improves auditing scores by 3.84x over the strongest baseline, while preserving competitive generation quality.
Read the full abstract

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

Artificial Intelligence

CASE: Cost-Aware Stopping for Efficient Long-Video Agents

Yiming Du, Chenghao Liu, Zhiyuan Liu et al. (+4 more)

Long-video agents can actively gather question-relevant evidence, but they typically leave a central decision implicit: when has the agent seen enough to answer? We propose CASE, a plug-in termination framework that frames this decision as policy-conditioned sequential stopping. At each causal checkpoint, CASE combines an auxiliary multiple-choice assessment of accumulated evidence with the host agent's execution state. From complete native trajectories, we construct a cost-aware target that compares answering now with stopping later along the same search path, accounting jointly for answer correctness and the full cost of continued reasoning. A lightweight Ridge regressor learns this decision gap and produces STOP/CONTINUE decisions. We evaluate three vision-language models with VideoSeek and AVP. On Video-MME, end-to-end accuracy changes by +0.67 percentage points on average while CASE reduces model-token use by 53.63%. The same frozen policies then transfer zero-shot to LongVideoBench and MLVU, with end-to-end accuracy changes of +3.38 and +4.58 points while saving 58.78% and 51.28% of model tokens, respectively. Across all agent-model-benchmark combinations, CASE attains the highest accuracy-efficiency Pareto-frontier coverage among the compared stopping methods (83.3%) at the selected operating points. Online execution preserves this favorable accuracy-efficiency trade-off and additionally reduces measured runtime by 54.1% on average. CASE provides a plug-in termination framework for long-video reasoning agents, enabling them to decide when further evidence acquisition is no longer worthwhile.
Read the full abstract

All authors Yiming Du, Chenghao Liu, Zhiyuan Liu, Fangxing Zheng, Zhao Wang, Junnan Nie, Songfang Huang

Software Engineering

Adaptive Code Revision Attacks on AI Pull Request Reviewers

Jingzhi Gong, Jie M. Zhang, Gunel Jahangirova et al. (+1 more)

Pull-request review protects software before new code reaches users, helping prevent vulnerabilities that could expose users to attacks. AI agents increasingly perform these reviews and explain which problems need fixing. However, for an attacker submitting vulnerable code, this feedback also reveals what changes may secure approval. Existing PR attacks seek such approval through persuasive text and comments while keeping executable code fixed. This leaves unclear whether an attacker can use the feedback to repair the reported problem while preserving a vulnerability in the revised code. We therefore conduct an empirical study of this threat using AFCRA (Adaptive Feedback-guided Code Revision Attack). To distinguish successful attacks from genuine repairs, we construct AFCRA-Bench from 159 disclosed vulnerabilities, with executable exploits to verify vulnerabilities in code. Across five-round interactions with Sonnet 5 and GPT-5.5 reviewers, AFCRA reaches success rates 2.5x and 12.5x those of the strongest evaluated text- or comment-based attack. Case studies of these successes show how reviewers accept repairs of reported problems while overlooking surviving vulnerabilities. These findings establish feedback-guided code revision as a threat to automated PR review. To address this threat, we derive actionable implications for researchers, AI providers, PR reviewers, and PR authors on securing AI-assisted development.
Read the full abstract

All authors Jingzhi Gong, Jie M. Zhang, Gunel Jahangirova, Meng Wang

Artificial Intelligence

MMPostTrainBench: Benchmarking Autonomous Research for Multimodal Post-Training

Yuxin Liu, Yuxuan Wang, Zhenxin Lei et al. (+9 more)

Autonomous research seeks sustained model improvements through iterative experimentation and feedback. LLM agents show promise in automating machine learning and language-model post-training, but their ability to sustain multimodal improvement remains unclear. We introduce MMPostTrainBench, a benchmark spanning eight tasks in image, audio, video, and joint audio-video understanding and image-grounded software repair. Agents operate from a common base model within fixed budgets, using development feedback before independent evaluation of their submitted models. Evaluation covers target and non-target model outcomes, iterative model improvement and selection, and research integrity. Across all eight tasks, 52.1% of model–task means fall below the base, and evaluated submissions also exhibit non-target regressions. Model performance does not consistently improve across research iterations, and agents do not reliably select the best evaluated candidate for submission; final submissions trail that candidate by up to 5.38 percentage points. Extending autonomous research from text-only to multimodal tasks introduces additional sources of error in perception, cross-modal alignment, and temporal grounding. The observed regressions and selection gaps highlight the need to balance targeted improvements with non-target capability preservation and to retain gains across research iterations. These requirements motivate MMResearch, a multimodal research framework that connects media-grounded evidence to hypotheses and interventions, carries findings across rounds through hierarchical memory, and retains candidates using development evaluation. Added to existing code-agent runtimes, it improves submitted-model accuracy by up to 7.75 percentage points for Claude Opus 4.8 with Claude Code and 2.33 points for GPT-5.6-sol with Codex.
Read the full abstract

All authors Yuxin Liu, Yuxuan Wang, Zhenxin Lei, Lingchen Meng, Yuchong Sun, Junming Lin, Hongcheng Liu, Yunfei Chu, Qize Yang, Jin Xu, Lei Zhang, Zhendong Mao

Machine Learning

Efficient Graph Generation via Direct Prediction and Flow Matching

Susie Lu

Generative modeling of graph-structured data is crucial for tasks ranging from drug discovery to social network simulation. Among these models, denoising diffusion models have achieved great success in graph generation by learning to progressively reverse a process that adds noise to the original graph. However, the standard noise-prediction approach of diffusion models is suboptimal for graph data. The goal for a graph generative model is to learn the clean graphs' topological properties, such as connectivity and degree distribution. Because a diffusion model that predicts noise does not explicitly learn these topological properties, it is challenging for the model to output graphs with the desired structural statistics. To address this challenge, we introduce Direct Graph Flow Matching (DiGFM), a novel graph transformer model guided by two goals: predict clean graphs and improve sampling efficiency. Distinct from the prevailing diffusion approach, DiGFM employs a continuous flow-matching paradigm and integrates direct graph prediction. Specifically, DiGFM maps the prior noise distribution to the clean graph distribution via a multi-step process: the model repeatedly predicts the underlying clean graph, and a transformation is employed to convert the model output to the velocity vector that points in the direction toward the clean graph distribution. This design enables DiGFM to generate high-quality samples using only 2.5% to 15.6% of the steps required by diffusion-based models, which leads to a 5.3x to 257x speedup in wall-clock inference time. Experiments demonstrate that DiGFM outperforms or matches prior state-of-the-art models across general graph benchmarks and molecular datasets, generating graphs with strong adherence to ground-truth structural statistics at significantly faster inference speeds.
Read the full abstract

Computation and Language

The Hidden States Cookbook: A Large-Scale Ablation Study for Noise-Robust Conversational Intent Classification in Industry

Bogdan Bogachov, Nikita Letov and Yaoyao Fiona Zhao

Conversational database interfaces face a critical challenge: users naturally embed queries in conversational noise (greetings, politeness, off-topic remarks), which degrades intent classification accuracy and wastes computational resources. Despite advances in orchestration and retrieval strategies, a fundamental question remains unanswered: which pooling strategy maximizes intent classification accuracy under realistic conversational noise in production language models? This work addresses this gap through 360 controlled experiments spanning four pooling configurations (mean, max, last-token, attention, and FFT-augmented variants) using Llama-3.2-1B-Instruct on BANKING77 and CLINC150 datasets under clean/noisy conditions with ten random seeds. Key findings reveal that attention pooling consistently outperforms alternative strategies under noisy conditions (~+2.6-2.8 F1 over the default), while mean pooling degrades performance by up to ~5 F1 points. Frequency-domain filtering does not produce consistent accuracy improvements and functions primarily as a structural variation rather than an accuracy-enhancing component. These results provide concrete, evidence-based guidance for building noise-robust conversational classifiers: attention pooling is recommended for noisy interfaces, mean pooling should be avoided, and last-token pooling is appropriate for clean-query scenarios.
Read the full abstract

Machine Learning

GNN-CB: A Graph Neural Network Competition Benchmark for Human and LLM Evaluation

Murad Hossen, Tasneem Selim, Gurur Gamgam et al. (+22 more)

Large language models (LLMs) have demonstrated strong performance on coding and reasoning benchmarks; however, their ability to solve graph-structured machine learning problems remains largely unexplored. In particular, no benchmark currently evaluates whether LLMs can autonomously solve end-to-end Graph Neural Network (GNN) coding tasks under realistic competition settings. To address this gap, this paper introduces GNN-CB, the first competition-based benchmark for evaluating both humans and LLMs on GNN coding tasks. GNN-CB consists of 18 curated competitions spanning node-, edge-, and graph-level prediction across diverse graph categories, domains, and difficulty tiers. All submissions are evaluated through a unified automated pipeline with hidden test sets and standardized scoring. Human participants solve tasks under controlled competition constraints, while LLMs are evaluated using a frozen zero-shot prompting protocol based on a plan-then-code paradigm with bounded execute-and-repair loops. The benchmark additionally supports both non-agent and autonomous agent-based evaluation within the same protocol. Under our evaluated protocol, LLMs rarely match Human Top performance and show less stable performance across competitions. No single model dominates: a few competitions are won by LLMs, yet humans still hold the top score on most tasks. We release GNN-CB as a living benchmark with automated evaluation infrastructure, dynamic leaderboards, and reproducible execution pipelines. Beyond benchmarking, GNN-CB provides a practice-oriented resource for studying GNN implementation across progressively diverse graph-learning tasks. The benchmark and evaluation framework are publicly available at https://basiralab.github.io/GNN-CB/.
Read the full abstract

All authors Murad Hossen, Tasneem Selim, Gurur Gamgam, Tuga Yousif, Abderrahmane Kasmi, Ikram Aissiou, Mubaraq Onipede, Faran Taimoor Butt, Sanae Zrigui, Rosa Y. G. Paccotacya-Yanque, Ignatius Balayo, Ikram Elhouiti, Hadil Affes, Bijay Adhikari, Sargam Goyal, Muhammad Ibrahim Isah, Mohammad Idrees Bhat, Samuel Kangoni Matia, Peguy Kem-Meka Tiotsop Kadzue, Maha Trabelsi, Emmanuel Owusu, Vinit, Nour Majdoub, Tamiru Alemnew, Islem Rekik

Computational Complexity

On the Computational Complexity of Problems: Formalizing Sensitivity to Uncertainty and Parametric Complexity Classes

Yannis Tzitzikas

Worst-case analysis, the setting of the $P \stackrel{?}{=} NP$ question, measures the cost of a problem as a function of input size alone. Yet problems that are exponential in general often become polynomial, even linear, once a checkable property of the input is known: Partition, for instance, is decidable in linear time on inputs whose elements are all equal. We develop a framework that makes such knowledge an explicit parameter $K$ of complexity statements. Costs are measured for fixed algorithms; we distinguish the promise cost from the guarded cost that also pays for checking $K$, and introduce the residual uncertainty $\Psi(n,K)$, the number of input cells that, together with $K$, determine the input. An Uncertainty Reduction Theorem decomposes the guarded cost into checking $K$, locating and reading the residual information, and solving an explicit reduced problem. The classes $P_\upsilon[K]$ and $NP_\upsilon[K]$ are restrictions to polynomial-time promises, made informative by the coverage of $K$. We show that small $\Psi$ forces sparsity but not tractability, characterize FPT as uniform membership in graded families $P_\upsilon[K_k]$, prove a free-variable threshold for SAT, and show that under the factoring assumption no checkable property of non-negligible coverage makes factoring easy. We relate the framework to promise problems, kernelization, average- and generic-case complexity and beyond-worst-case analysis, illustrate it on Partition, Graph Isomorphism, SAT, Vertex Cover and TSP, and apply it to guarded algorithm dispatch with latency contracts and optimal guard ordering, to solver routing in agentic AI, and to the role of promises in quantum speedups.
Read the full abstract

Artificial Intelligence

Sibyl: An Efficient Small-large Model Collaboration Framework for Long-horizon Tasks

Zhewei Fang, Yuxin Zhang, Zhenwei Shao et al. (+8 more)

Small language models (SLMs) offer a promising foundation for on-device agents through low-latency, resource-efficient inference, yet limited reasoning and planning capabilities constrain their performance on long-horizon tasks requiring multi-step interaction with the environment. Step-level collaboration between SLMs and larger cloud-hosted models can bridge this gap, but identifying states that warrant cloud assistance remains challenging: the contribution of each cloud call is entangled with subsequent actions and can be assessed only from the final task outcome. Compounding this challenge, the SLM must balance two competing objectives: maximizing task success and minimizing cloud calls. To address this, we propose Sibyl, an algorithm that trains SLM agents to selectively consult cloud models at the step level and internalize their guidance for subsequent decisions, achieving strong task performance with minimal cloud reliance. Sibyl follows a three-stage training pipeline that (1) builds a robust base policy through consultation-free self-evolving reinforcement learning (RL); (2) cold-starts consultation behavior via decisive-disagreement state mining; and (3) jointly optimizes consultation decisions and guidance internalization through consultation-aware RL. Experiments on ALFWorld and WebShop demonstrate that Sibyl, using only a 0.6B-parameter model, outperforms state-of-the-art baselines, including agent training and routing methods, by 95.2% and 80.4% in success rate while averaging only 0.8 and 3.9 cloud calls per trajectory, respectively.
Read the full abstract

All authors Zhewei Fang, Yuxin Zhang, Zhenwei Shao, Mengze Li, Zheng Lin, Long Chen, Zhou Yu, Zhe Chen, Zhiwen Chen, Zhaode Wang, chengfei lv

Computation and Language

Harness-Search: Guiding Long-Horizon Search through Multi-Agent Coordination

Shanyong Wang, Zhenwen Ji, Lei Jin et al. (+7 more)

Long-horizon search requires agents to gather evidence across multiple steps and synthesize it into well-supported answers. The recent agent harnesses provide a natural and promising framework to support such long-running search processes. As interaction histories grow, one single agent in harnesses might get stuck and cause the policy to lose track of unresolved questions, overlook useful evidence, or terminate before sufficient support has been collected. One of promising way is to decouple three distinct responsibilities of proposing retrieval actions, updating persistent state, and deciding when to stop rather than concentrating them within a single policy. Targeted at it, we introduce Harness-Search, a multi-agent search harness to reduce the local errors propagating across subsequent exploration, evidence curation, and termination decisions. In particular, Harness-Search assigns these responsibilities to three permission-bounded authorities: a Retrieval Policy that proposes search operations, a Memory Operator that validates and commits persistent-state updates, and a Summary Auditor that accepts or rejects termination based on the sufficiency of the curated evidence. Together, these roles form a Propose-Commit-Audit loop in which actions are proposed, persistent evidence is selectively committed, and stopping decisions are subjected to an explicit sufficiency check. Across seven long-horizon search benchmarks, Harness-Search improves both retrieval and answer generation under the same policy backbone, increasing Recall by 4.60-27.92 points and Final-Answer Recall by 12.34-30.13 points over the strongest harness-based baseline on each evidence-retrieval benchmark. Moreover, trajectory-level analyses show that Harness-Search continues to accumulate useful evidence and expand evidence coverage with less redundant retrieval as the search history grows.
Read the full abstract

All authors Shanyong Wang, Zhenwen Ji, Lei Jin, Yining Zhao, Yicheng Qian, Chengqiang Lu, Yi Wu, Yao Hu, Lizhen Cui, Yanyu Xu

Systems and Control

Backup Control Barrier Functions with Memory for Online Set Expansion

M. Yusuf Uzun and Ersin Das

Backup control barrier functions (bCBFs) construct implicit safe sets for nonlinear systems with bounded inputs using finite-horizon predictions under a backup controller. As in model predictive control, an invariant backup set serves as a terminal set that guarantees safe continuation beyond the prediction horizon. However, reaching a small terminal set may require long prediction horizons, increasing the cost of online prediction. In this paper, we introduce an online construction that builds a certified region from sets stored along backup trajectories while keeping the prediction horizon fixed. The method stores finite collections of balls and connects each newly certified collection to the backup set or to previously stored balls. We derive conditions that guarantee forward invariance of the union of these balls and the backup set under a fixed backup controller. These conditions also ensure that every stored state reaches the backup set in finite time. The resulting union serves as an enlarged terminal set, allowing safety to be certified with shorter prediction horizons. We formulate a feasible quadratic program for safe control and prove that safety is preserved as additional certified collections are stored online. Simulations on a quadrotor demonstrate safe task completion with a shorter prediction horizon than the standard bCBF.
Read the full abstract

Programming Languages

VHDL-REPOBENCH: A Repository-Level Benchmark for Evaluating Large Language Models on VHDL Design Generation

Prashanth Vijayaraghavan, Akul Malhotra, Ashutosh Jadhav et al. (+2 more)

Large Language Models (LLMs) are increasingly applied in hardware design automation, demonstrating strong potential in generating and understanding hardware description languages. However, most existing benchmarks focus on Verilog, with limited evaluation of VHDL, which remains widely used in industry and academia for FPGA and safety-critical systems. To address this gap, we introduce VHDL-REPOBENCH, a large-scale, cross-file, repository-level benchmark for assessing LLM capabilities on realistic VHDL design generation and analysis tasks. VHDL-REPOBENCH curates ~100 open-source VHDL repositories, encompassing ~2.5k VHDL files and ~500 testbenches, and provides structured problem statements, module stubs, and self-verifying testbenches. The benchmark enables comprehensive evaluation across syntax, semantic correctness, hierarchical reasoning, cross-file dependency resolution, and functional verification. We evaluate several state-of-the-art models, including GPT-4o, Llama-3-70B, Qwen2.5-72B, CodeLlama-70B, and multi-step reasoning approaches such as Reflexion and CoDes. Results reveal that while current LLMs achieve moderate line- and block-level accuracy, substantial challenges remain in multi-file reasoning, hierarchical design understanding, and specification-to-module generation. VHDL-REPOBENCH represents the first large-scale VHDL-focused benchmark and provides a valuable resource for the hardware design community to evaluate, compare, and advance LLM capabilities for practical VHDL development.
Read the full abstract

All authors Prashanth Vijayaraghavan, Akul Malhotra, Ashutosh Jadhav, Ehsan Degan, Vandana Mukherjee

Machine Learning

Does Explainability Survive Data Drift?

Samuel Ozechi

Model performance monitoring is a standard practice in machine learning deployments. Detection performance is tracked continuously, and model decay is expected as the relationship between the feature and target variables degrades, a phenomenon known as concept drift. Explanation fidelity, however, is rarely monitored with the same discipline, even in domains such as financial systems, healthcare, and other regulated environments where explanations are required for governance purposes. This paper investigates whether explanations can decay under data drift, even when the feature-target relationship remains stable, and whether explanations produced before drift occurs remain faithful to the decisions of the model that replaces them. Using the IEEE-CIS Transaction Fraud Detection dataset, we find statistically significant covariate shift but no statistically significant evidence of concept drift under the implemented conditional-drift tests, thereby providing an empirical setting in which input distributional change can be studied separately from detectable changes in the feature-target relationship. Local explanations are generated with ExIFFI and evaluated at three levels: path validity, structural behaviour, and fidelity under controlled intervention. Results show that while prior explanations retain substantial decision relevance to a retrained model, they are consistently less faithful than newly generated explanations, with no evidence of a systematically widening gap across the evaluated windows. The study shows that explanation fidelity requires its own monitoring, that structural stability of explanations does not guarantee functional fidelity, and that explanations should be treated as artifacts tied to the model that produced them.
Read the full abstract

Machine Learning

Distributed Subliminal Learning: Replacing Model Updates with Random-Carrier Outputs

Dario Fenoglio, Gabriele Dominici, Martin Gjoreski et al. (+1 more)

Collaborative learning typically exchanges model parameters: federated clients communicate updates, while independently adapted foundation models are combined by exchanging adapters or checkpoints. This makes communication scale with model size and requires local specializations to be reconciled in weight space, where interference is common. We ask whether knowledge can instead be shared through model behavior on task-unrelated inputs. We introduce Distributed Subliminal Learning (DSL), a collaborative learning primitive in which participants adapt a common model locally, probe it with task-unrelated inputs, and transmit only the resulting carrier outputs. A coordinator pools these outputs and distills them into a shared model. The primitive supports one-shot foundation-model composition through carrier completions and iterative federated learning through carrier logits, without transmitting model updates or requiring task-related proxy data. In LLM composition, compared with LoRA averaging, DSL achieves higher preference retention (94.56% vs. 87.76%) and a larger GSM8K gain over the base model (22.0 vs. 0.6 points), while reducing upload by 30.6-49.0$\times$. In federated classification, DSL reaches 96.83% on MNIST with 8.9$\times$ less uplink than FedAvg and provides lower-communication operating points on CIFAR-10 and Tiny ImageNet. These results establish random-carrier outputs as a practical communication primitive for knowledge sharing across distinct collaborative learning paradigms.
Read the full abstract

All authors Dario Fenoglio, Gabriele Dominici, Martin Gjoreski, Marc Langheinrich

Systems and Control

Physics-informed neural control of heat flow rate in a counter-flow heat exchanger: modeling, identification and experimental validation

Konstantinos Skantzikas, Emmanuel Witrant and Bojan Mavkov

The advanced operation of heat exchangers plays a critical role in the development of efficient and sustainable multivector energy systems. This paper proposes a hybrid physics-informed and data-driven framework for the analysis, identification, and regulation of transient heat transfer in counter-flow heat exchangers. A parameterized physical model combines classical heat-transfer relations with transport equations to capture convective propagation, transport-induced delays, and nonlinear effects associated with flow-rate variations. The heat-transfer coefficient is identified from experimental data, enabling the characterization of flow-dependent heat-transfer dynamics over a broad operating range. The identified model is then used to develop a hierarchical real-time control strategy. An inner loop regulates the volumetric flow rate using a high-order data-driven linear model, while an outer loop regulates the exchanged heat power using a machine-learning controller initially trained on the physical model and subsequently refined with experimental data. Experimental results demonstrate accurate reproduction of transient heat-transfer dynamics and effective heat-flow-rate regulation under varying thermal boundary conditions (identification fit above 94% in normalized root-mean-square error and total variation of the flow rate reduced by half compared to a proportional-integral controller). The proposed framework combines physical interpretability with data-driven adaptation for the analysis and real-time operation of dynamic heat-transfer systems.
Read the full abstract

Artificial Intelligence

Learning Field Reconstruction from Incomplete Data by Globally Correcting Local Estimates

Renhao Zhong, Zihan Zhou, Chiyuan Ma et al. (+1 more)

Reconstructing physical fields from training samples that are always incomplete requires learning spatial structure from fragmented observations. Existing context–query work establishes how held-out observations provide valid training targets, but this does not make the complete-field distribution identifiable when every training field is incomplete. With finite data, weak evidence of sharp transitions and localized variations can further favor averaged predictions that attenuate local detail. A structural prior is therefore needed to favor plausible completions; local spatial relationships offer one grounded in the observations. We propose a locally constructed, globally revisable estimator that explicitly learns local field estimates and subsequently corrects them using full-domain observations. A shared coordinate-conditioned predictor learns from incomplete patches, allowing relatively well-observed neighborhoods to provide direct supervision of local structure. Its overlapping predictions are reconciled into an observation-conditioned consensus field. A full-domain estimator retains the original observations and learns a residual correction around this frozen field estimate, allowing locally constructed structure to be revised by broader evidence. The local estimate serves as both an explicit input, accompanied by its discrepancies with the observations, and a prediction starting point that the global model can revise. On three real-world ocean datasets with authentic observation gaps, our estimator achieves the lowest MSE and highest PSNR on withheld source-supported values, reducing MSE by 28.9%–34.5% against the strongest external baseline.
Read the full abstract

All authors Renhao Zhong, Zihan Zhou, Chiyuan Ma, Tianshu Yu

Information Theory · Cross-listed from quant-ph

Minimax Quantum State Tomography in a Known Trace-Distance Ball

Yonglong Li

We study quantum state tomography when the unknown state lies in a trace-distance ball of radius $R$ around a known reference state $\rho$. For arbitrary collective measurements, we determine the minimax expected risk and confidence radius up to universal constants, uniformly over the dimension, center spectrum, radius, and sample size. Write $\Gamma=d(\mathrm{Tr}\sqrt\rho)^2+d^2R$. For $d\ge2$, $0<R\le1$, and $n\ge1$ copies, the expected risk is $\Theta(\min\{R,\sqrt{\Gamma/n}\})$. At failure probability $0<\delta\le1/3$, the minimax confidence radius is \[ \Theta\!\left(\min\left\{R,\sqrt{\frac{\Gamma+\log(1/\delta)}{n}}\right\}\right). \] For $\varepsilon\le c_0R$ with a universal $c_0>0$, the copy complexity is $\Theta((\Gamma+\log(1/\delta))/\varepsilon^2)$. For the lower bound we use a weighted trace-norm witness to retain all spectral scales in a square-root perturbation family. For the upper bound, we give a direct spectral and concentration analysis of a modified existing purification-based procedure.
Read the full abstract