Computer Vision
FastBench: Can Streaming VLMs Perceive High-Dynamic Real-World Streams?
Yuxuan Hu, Weikang Shi, Yang Bo et al. (+10 more)
Streaming
Video Large
Language Models (VLMs) enable continuous
video understanding, yet existing
benchmarks focus on low-dynamic scenarios. Under bounded context budgets, models must balance temporal history, spatial resolution, and temporal granularity; sparse sampling at 1–2 FPS misses fast events. We introduce FastBench to evaluate high-dynamic perception in real-world
video streams. Its trajectory-grounded pipeline combines QA generation from high-FPS clips, filtering of questions answerable at 2 FPS, answer verification using SAM3 and CoTracker3 trajectories, and three rounds of human inspection. FastBench contains 306 QA pairs across eight domains, six capabilities, and forward, instant, and backward temporal scopes, with human-annotated evidence intervals. We also present ProactiveFrame, a training-free baseline that adjusts incoming frame rates through text
tokens. A dual-tier sliding window retains recent high-FPS observations while downsampling older ones into sparse history. Experiments reveal substantial limitations: the strongest model, Gemini-3.5-Flash, scores only 50.7%. Denser sampling improves Qwen3-VL-8B from 32.9% at 2 FPS to 44.6% at 24 FPS, but gains saturate as history is compressed. ProactiveFrame outperforms sparse uniform sampling by 5.4 and 1.5 percentage points, yet remains well below oracle-guided focusing, showing that current VLMs struggle to determine from the stream alone when finer temporal perception is needed. FastBench provides a testbed for high-dynamic streaming
video understanding. Code and data: https://github.com/Ashone3/FastBench.
Read the full abstract ▾
All authors Yuxuan Hu, Weikang Shi, Yang Bo, Xudong Lu, Xintong Guo, Shuhan Li, Yuyang He, Huankang Guan, Peiwen Sun, Yunqiao Yang, Wenbo Li, Rui Liu, Hongsheng Li
Disordered Systems and Neural Networks
Lyapunov spectrum of random neural networks
David G. Clark
The Lyapunov spectrum of a nonlinear recurrent neural network with random asymmetric couplings is calculated in the limit $N\to\infty$. The calculation is based on a finite-$N$ identity that expresses the cumulative distribution of Lyapunov exponents as a response function of the tangent-space dynamics, with the tangent trajectory selected by a minimum-norm condition rather than by an initial condition. A cavity method then determines this response function at large $N$ through a self-consistent single-site problem. This result establishes that the chaos in this network is extensive and gives access to diffeomorphism-invariant properties of the dynamics. This work was done in collaboration with the AI models GPT-6 Astra and Claude Opus 5.5.
Read the full abstract ▾
Probability
On Large Drift Enhancement for the Relaxation Time of Averages
Alain-Sol Sznitman and Klaus Widmayer
In this note we prove in a broad analytic set-up a general result characterizing the asymptotic vanishing of the $L^2$-relaxation time of averages in the presence of a large incompressible drift-like term. The broad analytic set-up is analogous to that of Constantin-Kiselev-Ryzhik-Zlatos '08 or of Coti Zelati-Delgadino-Elgindi '20. A crucial role is played by the singular value gap method of Chatterjee '25, further developed by Eberle-Lörler '26.
Read the full abstract ▾
Robotics
RoboRSI: Stable, efficient, and reusable robot self-evolution in complex real-world environments
Zimo Wen, Yijin Chen, Yuxuan Cao et al. (+8 more)
A generalist
robot should not only perform diverse tasks but also improve through experience, turning what it learns during execution into capabilities that later tasks can reuse.
Robot agents that act through code can already repair programs from execution feedback, yet it remains a central challenge to organize this experience around the task structure that gives it meaning, so that each repair is attributed to the responsible capability, supported by execution evidence, and validated before it is reused. We introduce RoboRSI, a
robot self-improvement system built on Top-Down Skill Refinement (TSR). TSR decomposes tasks into compound, atomic, and base skills with scoped responsibilities and explicit input–output contracts, attributes each execution outcome to the responsible branch, and confines revision to that branch. Building upon this structure, a Manager, Planner, Engineer, and Reviewer coordinate planning, execution, diagnosis, and the validated release of new skills, while people steer the process through objectives and corrections; stable skill sequences are further consolidated into reusable compound skills. On a mobile manipulator, RoboRSI develops multi-object household cleanup over 104 rounds. In simulation, it achieves the highest success rate on LIBERO, LIBERO-PRO, LIBERO-Plus, and RoboTwin, exceeding the strongest baseline by 2.7 to 11.0 percentage points.
Read the full abstract ▾
All authors Zimo Wen, Yijin Chen, Yuxuan Cao, Wendi Chen, Yanwen Zou, Wenye Yu, Fuhang Kuang, Han Xue, Jun Lv, Chuan Wen, Cewu Lu
Computer Vision
Pumpire: Unified Benchmark for Metric Distance Estimation
Siyu Chen, Zehan Wang, Jiayang Xu et al. (+6 more)
We present Pumpire, a unified
benchmark for evaluating metric point-pair distance estimation capability of both image- and
video-level 3D foundation models, with or without depth priors. In contrast to previous approaches that normally evaluate depth and camera intrinsics separately or evaluate point-clouds with
geometric similarity metrics, which cannot directly reflect models' point-to-point distance estimation capability, Pumpire directly assesses point-to-point distances from the reconstructed geometry. To this end, we collect a large-scale and diverse
dataset (pumpire-6k) comprising 100 real-world scenes, each annotated with physically measured point-pair distances and containing 64 frames, for a total of 6,400 frames. Building on this
dataset, we establish a holistic evaluation protocol that covers both image- and
video-level 3D foundation models and enables direct assessment of point-pair distance errors and cross-setting comparison. We conduct extensive experiments across 29 baseline configurations of representative 3D foundation models and provide a comprehensive analysis of the results. By offering this
benchmark, we target the more fundamental ability to perceive and estimate physical scale in the reconstructed 3D space, which prior evaluation protocols have largely overlooked. The project page can be found at https://pumpire.github.io/
Read the full abstract ▾
All authors Siyu Chen, Zehan Wang, Jiayang Xu, Yihan Wu, Jialei Wang, Junming Chen, Ziang Zhang, Yutong Ying, Zhou Zhao
Nuclear Theory
A Density Functional for Quark Matter with Screened Confinement Interactions
Oliver Heymer and David Blaschke
We derive a density functional for
quark matter with linear confinement interaction, applying the principle of saturation of color interactions. For the saturation parameter $\mu$ the ansatz $\mu=k_F$ is suggested, where for the
Fermi momentum $k_F \propto n^{1/3}$ going to zero the Cornell potential is obtained and at finite
quark densities $n$ an interaction energy density functional $\propto n^{2/3}$ results.
Read the full abstract ▾
Computer Vision
Beyond Spatio-Temporal Priors: A Generalizable Approach for Dense Correspondence Matching
Luping Liu, Bingyi Kang, Yifan Wang et al. (+1 more)
Dense correspondence matching has historically been bounded by simplifying spatio-temporal priors, such as smooth motion and rigid geometry. While effective for classical tasks, these assumptions break down in image editing and reference-guided generation (IEG), where transformations can preserve visual identity while breaking physical continuity. To establish identity-preserving correspondence across such transformations, we introduce FreeMatching, a generalizable framework combining generative and semantic foundation representations with heterogeneous supervision from classical
datasets, tracked
videos, and synthetic scenes. Teacher-guided iterative refinement further improves correspondence in IEG without dense correspondence annotations. Experimentally, a single FreeMatching model substantially improves correspondence quality on challenging IEG image pairs while retaining competitive performance on classical
benchmarks. Furthermore, we demonstrate its utility as a quantitative metric for evaluating identity preservation, with scores that correlate with human judgment. The code is available at https://github.com/luping-liu/FreeMatching.
Read the full abstract ▾
All authors Luping Liu, Bingyi Kang, Yifan Wang, Dong Xu
Machine Learning
A Unified Bellman Operator for Safety-Critical Reinforcement Learning
Nishanth Arun Rao, Royina Karegoudra Jayanth, Benjamin Eysenbach et al. (+1 more)
Reinforcement learning in safety-critical domains requires maximizing task performance while strictly adhering to safety constraints. Existing safe reinforcement learning paradigms typically force a trade-off: they either require a priori knowledge to provide strict safety guarantees (e.g., safety filters), or they enable joint learning but only satisfy safety constraints on average. In this work, we propose a novel Bellman operator that unifies performance and safety objectives into a joint value function. We show that temporal difference learning with the joint Bellman operator converges under a two-timescale
stochastic approximation framework. On the fast timescale, the safety value of the learning joint policy is estimated, while the joint value is estimated on the slow timescale. Convergence is ensured by formulating the limiting dynamics as an occupation-averaged differential inclusion, and showing that it asymptotically converges to a set of limiting optimal safety-constrained task value functions. Theoretically, once converged, the resulting optimal policy maximizes task return while maintaining safety at all times. Empirical evaluations on continuous control tasks with neural approximations demonstrate stable convergence with near-zero safety violations at test time.
Read the full abstract ▾
All authors Nishanth Arun Rao, Royina Karegoudra Jayanth, Benjamin Eysenbach, Jaime Fernández Fisac
Computer Vision
OneSearch-VL: Unified Multimodal Deep Research Agent for Image and Video
Hongyu Li, Manyuan Zhang, Kaituo Feng et al. (+11 more)
Single-image, multi-image, and
video deep research require different visual operations but share a workflow of visual grounding, external
retrieval, and fact composition. A key challenge is to preserve the dependencies linking localized visual anchors, entity relations, source-supported facts, and answer-producing operations. We introduce OneSearch-VL, a unified
agent centered on the Visually Grounded Evidence Graph (VGEG), which encodes these dependencies as a shared task-level reference for data construction, process supervision, and operation-level evaluation. Our VGEG-based data engine constructs and verifies multi-image and
video questions and filters expert trajectories. Using these data, we assemble OneSearch-VL-SFT-110K and OneSearch-VL-RL-10K for SFT and RL, respectively. We further derive the Evidence-aware Visual-Grounded Rubric reward (EVGR) from VGEG annotations to supervise evidence traceability and visual grounding during RL. For fine-grained evaluation, we construct OneSearch-MI-Bench and OneSearch-
Video-Bench, organizing questions by the research operations encoded in their VGEGs. Experiments show that OneSearch-VL-8B improves over Qwen3-VL-8B with tool access by 20.2 and 17.6 percentage points on the two new
benchmarks, respectively, while also achieving substantial gains across 7 image
benchmarks and VideoDR. Project repository: https://github.com/appletea233/OneSearch-VL
Read the full abstract ▾
All authors Hongyu Li, Manyuan Zhang, Kaituo Feng, Shu Chen, Dian Zheng, Hao Li, Hao Yu, Zhangquan Chen, Zoey Guo, Ray Zhang, Shaofei Huang, Tianrui Hui, Linjiang Huang, Si Liu
Cosmology and Nongalactic Astrophysics
Cosmic topology. Part IId. Eigenmodes and correlation matrices of lens spaces
Mikel Martin Barandiaran, Catherine Petretti, Stefano Anselmi et al. (+16 more)
The global
topology of the Universe is a longstanding open question. In this work, we examine the statistical signatures of a positively curved universe with a Friedmann–Lemaitre–Robertson–Walker metric and the
topology of a lens space $L(p,q)$. Since these
manifolds are generally statistically anisotropic and inhomogeneous, their
cosmic microwave background (CMB) covariance
matrices contain non-zero off-diagonal entries that depend on observer location. We compute these full harmonic-space covariance
matrices for scalar perturbations for arbitrary lens spaces and observer position. Using the Kullback–Leibler divergence, we assess the distinguishability of these spaces from a simply connected three-sphere with the same
curvature. The results show that topological signatures can still be significant in a
cosmic-variance-limited regime even when the length, $d_{\rm NC}$, of the shortest loop around the Universe through the observer exceeds the diameter, $d_{\rm LSS}$, of the last-scattering surface by up to $10\%$. This distinguishability depends on the
curvature, the lens space and the observer position mainly through the ratio $d_{\rm NC}/d_{\rm LSS}$, which is set not only by the
curvature radius but also by $p$, $q$ and the observer location. Therefore, for any value of the spatial
curvature, however small, there are lens spaces and observer positions for which $0.985\,d_{\rm LSS}<d_{\rm NC}<1.1\,d_{\rm LSS}$. Such observers would find no matched circles in the CMB large enough to have been detected to date, yet the
topology would remain potentially discoverable through its statistical signatures.
Read the full abstract ▾
All authors Mikel Martin Barandiaran, Catherine Petretti, Stefano Anselmi, Deyan P. Mihaylov, Anna Negro, Glenn D. Starkman, Andrius Tamosiunas, Yashar Akrami, George Alestas, Javier Carrón Duque, Craig J. Copi, Fernando Cornet-Gomez, Linn Htat Lu, Andrew H. Jaffe, Arthur Kosowsky, José Javier Ortega Gómez, Ricardo G. Rodrigues, Amirhossein Samandar, Cynthia Trendafilova
Computer Vision
WOVEN: Weaving Visual World Modeling into Multimodal LLMs
Zheyu Fan, Yue Zhang, Mingkai Deng et al. (+9 more)
Multimodal large
language models (MLLMs) struggle with spatial,
embodied, physical, and temporal
reasoning. We hypothesize that these failures reflect a shared deficit in visual transition
reasoning, and test whether this capability can serve as a shared training primitive, one that different models can learn from different supervision sources and reuse across different tasks, with a systematic training recipe. Existing
benchmarks document these deficits separately but do not support controlled comparisons across scenes, actions, and
reasoning operations. We therefore introduce WOVEN, a training source and
benchmark for visual transition
reasoning that organizes transition supervision by scene, action, and
reasoning type, using diverse, realistic rollouts from
video-pretrained generative models: 36,076 examples across 20 scene types, 5 action types, and 8
reasoning types. We first evaluate 38 frontier MLLMs (e.g., GPT-5.4 and Qwen3-VL-235B-A22B) and find a substantial and systematic deficit: even the strongest models fall far below humans, and the failures recur across model families and persist with scale. We then train MLLMs at multiple scales on WOVEN and find that they learn a shared capability that transfers broadly: training subsets of only about 2,000 items each collectively improve 22 of 26 external
benchmarks by up to 27.3 percentage points, and WOVEN data can replace 30-50% of a task's own training data with comparable accuracy. Controlled comparisons further yield a training recipe for visual world modeling, validated prospectively on held-out
benchmarks: select supervision by the
reasoning operation it teaches rather than by the actions, scenes, or domains it shows, and prefer larger changes to the visual state for robustness. Our work establishes visual transition
reasoning as a reusable foundation for systematic visual world-model training in MLLMs.
Read the full abstract ▾
All authors Zheyu Fan, Yue Zhang, Mingkai Deng, Kangrui Wang, Qineng Wang, Canyu Chen, Jie Hao, Xing Fan, Chenlei Guo, Eric P. Xing, Mohit Bansal, Manling Li
Computer Vision
MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances
Mingyuan Lei, Yoonchang Sung and Tat-Jen Cham
Generating realistic human-object interactions (HOI) in complex 3D scenes requires two complementary capabilities:
reasoning about interaction feasibility in the environment and synthesizing realistic human-object motion. However, supervision for these capabilities is rarely available jointly at scale. Human-scene
datasets provide rich information about environment-aware motion, while human-object
datasets capture detailed interaction dynamics, yet paired human-object-scene data remain scarce. We present MAMHOI, an affordance-mediated factorization for scene-aware human-object interaction generation. MAMHOI factorizes scene-aware HOI generation through an explicit motion-affordance interface between scene understanding and motion synthesis: a scene-conditioned model first predicts where and how an interaction can be feasibly executed, and an affordance-conditioned HOI model then generates the corresponding human-object motion. This factorization allows scene understanding and interaction dynamics to be learned from complementary sources of supervision without requiring paired human-object-scene data. Experiments in complex indoor environments show that MAMHOI reduces object–scene penetration while better preserving human–object interaction quality, yielding more realistic and physically feasible scene-aware interactions. Project page: https://leimingyuan.github.io/MAMHOI-project-page/
Read the full abstract ▾
Cryptography and Security
ORCAGen: Orchestrating Context-Aware Malware Deception with RAG-Guided Generative AI
Shihab Ahmed, Md Sajidul Islam Sajid, Teryl Taylor et al. (+2 more)
Malware defenses often remove or isolate suspicious programs as quickly as possible. While effective for containment, this approach can also waste an opportunity to observe attacker behavior and deploy targeted countermeasures. ORCAGen takes a different approach: it uses GenAI to build malware-specific deception playbooks offline, validates them before deployment, and enforces only the verified logic at runtime. ORCAGen combines
Retrieval-Augmented Generation (RAG) with structured
prompt engineering to generate both proof-of-concept (PoC) malware and corresponding deception orchestration code. A curated knowledge base (KB) of malware procedures and active defense strategies grounds the generation process, helping the LLM produce threat-specific and executable deception logic rather than generic or hallucinated outputs. The generated PoC malware provides a safe and reproducible way to test whether a deception strategy can disrupt, redirect, or suppress targeted malware behavior before the strategy is added to the runtime playbook. We evaluate ORCAGen across GPT-4o, GPT-5.5, Gemini 3.5 Flash, Qwen3-Coder, and Claude Sonnet 4.5 using execution success,
hallucination rate, refinement effort, deception effectiveness, and runtime overhead. The evaluation covers synthesized malware scenarios and 150 real-world malware samples across keyloggers, information stealers, and ransomware. Across the evaluated scenarios, GPT-5.5 required the fewest refinements and produced no observed hallucinated APIs, while Gemini 3.5 Flash achieved the lowest response time and runtime overhead. The results show that RAG-guided structured prompting can support the scalable construction of malware-specific deception playbooks that remain lightweight and deterministic during runtime enforcement.
Read the full abstract ▾
All authors Shihab Ahmed, Md Sajidul Islam Sajid, Teryl Taylor, Frederico Araujo, Tariqul Islam
High Energy Physics (Theory)
Energy-energy correlators in the large-charge limit of $\mathcal{N}=4$ super Yang-Mills theory
Alessandro Georgoudis, Max Jackson and Congkao Wen
We study the energy-energy correlator (EEC) in heavy states of four-dimensional $\mathcal N=4$ super Yang–Mills theory in the large-charge 't Hooft limit, in which the dimension $\Delta$ of the heavy operator (or, equivalently, its $R$-charge) is taken to infinity with the effective coupling $\lambda_Q \propto g_{\mathrm{YM}}^2\Delta$ held fixed. The relevant heavy-heavy-light-light four-point correlators are known to all orders in $\lambda_Q$ in this limit. For certain classes of operators, the perturbative series can be resummed to obtain exact finite-coupling expressions. We use these results to determine the corresponding EEC, whose dynamical part can be extracted from a double discontinuity of the four-point correlator. We determine the EEC at finite effective coupling and study its physical behavior as a function of $\lambda_Q$ and the angular separation between the detectors, both at generic angles and in the collinear and back-to-back limits.
Read the full abstract ▾
Information Theory
Commitment over Gaussian channels with deterministic identification codes
Pau Colomer, Christian Deppe, Holge Boche et al. (+1 more)
We establish a previously unnoticed connection between commitment (BC), a fundamental cryptographic primitive, and deterministic identification (DI), a post-Shannon communication setting. Specifically, for additive white Gaussian noise channels $\mathcal{G}$ we show that a reliable commitment protocol can be obtained from deterministic identification codes in both the honest but curious and fully dishonest security settings. This viewpoint allows BC schemes to inherit the asymptotic performance of DI codes. Indeed, we show that commitment is naturally achievable in the linearithmic regime, i.e., with message sets of size $N_n=\exp [\Theta(n\log n)]$, and we establish a lower bound on the linearithmic-scale BC capacity of $\dot C_{\text{hBC}}(\mathcal G)\geq\frac12$ in the honest but curious case and $\dot C_{\text{BC}}(\mathcal G)\geq\frac14$ in the fully dishonest picture. This refines the best previously known characterisation of the commitment capacity over $\mathcal{G}$, which established only an infinite linear-scale capacity.
Read the full abstract ▾
All authors Pau Colomer, Christian Deppe, Holge Boche, Andreas Winter
Computer Vision
WorldCast: Distributed Multiplayer World Models
Ziyang Ye, Junchao Huang, Evelyn Zhang et al. (+9 more)
Multiplayer world models must generate independently controlled views with consistent representations of both players and their shared environment. Most existing approaches coordinate multiple players through joint multi-view generation, whose cost grows with each additional player. We present WorldCast, a distributed multiplayer world model in which each player runs a local client comprising a
video generator and a state model. Using recorded player positions and map geometry during training, the state model estimates the player's position from generated
video and control inputs. Clients exchange player states and project them into camera-aligned player state fields that guide where and how other players are rendered. Shared scene state enables clients to reuse one another's generated observations to maintain consistent scene appearance across views. Experiments on Counter-Strike 2 demonstrate WorldCast's consistency, real-time performance, and distributed scalability. The camera-aligned player state field improves player rendering rates by over an order of magnitude over joint-generation methods, while shared scene state improves visual consistency over whole rounds. Each client runs in real time and exchanges only player and scene states, enabling scalable multiplayer generation without a centralized computational bottleneck. Image quality remains stable over hour-long rollouts.
Read the full abstract ▾
All authors Ziyang Ye, Junchao Huang, Evelyn Zhang, Zhihao Xie, Ruicheng Zhang, Boyao Han, Litao Ban, Ziye Wang, Xinting Hu, Shaoshuai Shi, Zhuotao Tian, Li Jiang
Robotics
GLIO2: A GPU-Parallelized Tightly-Coupled LiDAR-Inertial-GNSS System for Robust and Real-Time Global Localization and Mapping
Qi Zhang, Xikun Liu, Qijun Qin et al. (+5 more)
Globally consistent, real-time state estimation in large-scale, perceptually degraded environments is essential for autonomous vehicles and aerial
robots, and requires fusing
LiDAR, inertial, and GNSS measurements. Existing fusion methods, however, share a scan-to-map front-end with two failure modes. First, each scan is aligned to an incrementally built map that drifts under degeneracy, and once the estimate diverges the error is irrecoverable. Second, even without divergence, a registration biased by dynamic objects or wrong correspondences is propagated as a single pose constraint with an over-confident covariance, leaving its correspondences unavailable for GNSS to re-weight or relinearize. We propose GLIO2, a tightly-coupled
LiDAR-Inertial-GNSS system whose GPU-parallel front-end jointly optimizes scan-to-multiscan
LiDAR, IMU pre-integration, and raw GNSS measurements in a single sliding-window factor graph, sustaining real-time operation on edge hardware. A complementary offline back-end reuses the same cached factors to refine the entire trajectory in batch, completing the 30-min, 4.51-km UrbanNav Whampoa sequence in about 24 s. Across three public
benchmarks (UrbanNav, MARS-LVIG, M3DGR) and self-collected
UAV and vehicle data, GLIO2 attains the best overall accuracy among evaluated systems. On a 5.66-km bridge traversed at up to 96 km/h, where every competing baseline diverges under
LiDAR degeneracy, it maintains 1.6 m horizontal accuracy. On an NVIDIA Jetson Orin NX, the full pipeline runs at about 25 Hz (39.60 ms per scan). The source code and
datasets will be released.
Read the full abstract ▾
All authors Qi Zhang, Xikun Liu, Qijun Qin, Xiangru Wang, Junzhe Wang, Naigui Xiao, Jianhao Jiao, Weisong Wen
Computation and Language
Predicting Alignment Generalization with Value Representations
Andy Liu, Mehar Bhatia, Karolina Stanczak et al. (+3 more)
LLM developers post-train their models to exhibit prosocial values and behavioral traits, which are enumerated in an
alignment target. However, while recent post-training developments have yielded models that score highly on
alignment evaluations, training models on sets of narrow behaviors still influences their behavior across unseen contexts and environments in unexpected ways. In this paper, we establish the task of
alignment generalization prediction, i.e., predicting how fine-tuning a model to follow a given value changes its behavior across a wide range of held-out values. We conduct a large-scale analysis of
alignment generalization effects across 66 values found in modern
alignment targets, and
benchmark representational techniques on the
alignment generalization prediction task. We find that representations based on model activations when applying values in context significantly outperform methods based on textual descriptions of the values. Specifically, the best activations-based methods achieve correlations of 0.45 with our generalization
matrix, compared with 0.05 from description-based baselines. We then show the applicability of representations that predict
alignment generalization toward downstream tasks by using them to measure how similar the values in a multi-value
alignment target are, which we find is significantly correlated with model robustness. Finally, we show initial evidence towards a shared, model-independent value space, which we use to develop the first taxonomy of LLM values grounded in empirical generalization dynamics. Our work demonstrates the importance of studying value generalization in LLMs and its application toward the more empirical design and training of model behavior.
Read the full abstract ▾
All authors Andy Liu, Mehar Bhatia, Karolina Stanczak, Mona Diab, Vered Shwartz, Daniel Fried
Artificial Intelligence
Searching for "Harmful Refusal": A Psychometric Audit of an AI Safety Benchmark
Christopher M. Stewart, Preston Botter, Natalie Sarabosing et al. (+5 more)
Safety
benchmarks typically report one overall score for a suite of
datasets, each of which may target one or more safety-related attributes, so models with similar overall scores can have very different attribute profiles. Comparing models is more tractable at the level of individual attributes, yet it is often unclear whether even a single
dataset's scores isolate any single attribute. One plausible candidate for such an attribute is harmful refusal, a model's tendency to refuse dangerous or policy-violating
prompts. We examine whether it constitutes a single, measurable attribute in HELM Safety. Using a construct validity framework that stipulates that an attribute must exist before a test can measure it, we start with HELM Safety's four
datasets that might plausibly target harmful refusal, but find that three are saturated. We subject the remaining
dataset, HarmBench, to two psychometric tests to determine if a single attribute like harmful refusal could stand behind its score. First, multidimensional item response theory modeling strongly suggests that HarmBench does not measure a singular attribute. Second, a differential item functioning analysis finds items where models from different developers with the same refusal ability score differently. These flags largely disappear under scope-specific matching, a pattern consistent with aggregation effects but not sufficient to rule out domain-specific developer differences. Zooming out, HarmBench collapses distinct harm behaviors into one score, and the overall HELM safety aggregate further collapses HarmBench and scores from other
datasets into a single top-line number. Any safety score that averages over
datasets and items can hide saturation and conflate behaviors this way. We argue that a score should earn its single-attribute reading before models are compared with it.
Read the full abstract ▾
All authors Christopher M. Stewart, Preston Botter, Natalie Sarabosing, Muye Zhang, Rachel Phinnemore, Shalini Ghosh, Hong Shen, Hoda Heidari
High Energy Physics (Experiment)
A compact, interrogable Tree Tensor Network Jet Tagger on Lund declustering trees
Fabrizio Napolitano, Luca Della Penna, Tommaso Tedeschi et al. (+1 more)
Jet taggers built on the Lund declustering tree encode QCD radiation structure in the inputs and process it with graph networks. We ask whether the tree itself can also constrain the computation, and what becomes accessible when it does. For this, we construct a novel
quantum-inspired Tree Tensor Network (TTN), where a single tensor core is shared among the internal nodes of the Cambridge/Aachen declustering tree, employing Lund coordinates, and with no particle-identification or detector-level information. The construction is based on a strong prior on the radiation history rather than on the constituents, complementary to compact equivariant taggers. Because the same learned map acts identically at every node, the trained model can be evaluated on truncated trees and its latent space compared across recursions, and used to localise where on the Lund jet plane the Pythia and Herwig jet representations diverge, showing the regions of generator-dependent concentration. The $\chi=8$ model has 4014 trainable parameters against approximately 391,000 for LundNet. We evaluate boosted top tagging and
quark/
gluon discrimination at particle level on Pythia and Herwig samples, in both native (Pythia$\to$Pythia) and transfer (Pythia$\to$Herwig) configurations. Despite differing in capacity by two orders of magnitude, the TTN reaches native performance close to LundNet in
quark/
gluon discrimination, slightly underperforming in top tagging, with a similar Pythia-to-Herwig transfer behavior in both tasks. Finally, we show the behavior of the network as a function of the bond dimension, and as a function of the perturbative scale.
Read the full abstract ▾
All authors Fabrizio Napolitano, Luca Della Penna, Tommaso Tedeschi, Livio Fanò
Robotics
LeWAM: A JEPA World Action Model with Diffusion-Steering-Based MPC
Shashank Hegde, Alexander Popov, Elie Aljalbout et al. (+1 more)
World action models (WAMs) predict actions and future observations, typically from a reconstruction-based representation that carries noisy, redundant information which can complicate downstream predictions. We introduce LeWAM, a bidirectional
transformer for forward, backward, inverse dynamics and policy prediction, on a decoder-free JEPA latent trained end-to-end through all four modes. We see the following benefits: 1)
Alignment: linear probes read
robot and object state from LeWAM's latent better than from a regular Le World Model (a forward-only JEPA world model), while the latent ignores visual distractors as well as LeWM does and far better than a reconstruction-based WAM. 2) Acting: Closed-loop evaluations of LeWAM match a regular flow-matching policy trained on the same encoder at matched size, while also providing a world model. 3) Planning: Sampling raw actions when planning with WAMs lets MPC exploit dynamics-model inaccuracies; planning in the noise space of the policy head instead improves the closed-loop performance of these WAMs.
Read the full abstract ▾
All authors Shashank Hegde, Alexander Popov, Elie Aljalbout, Nikolai Smolyanskiy
Atomic Physics
Local polarization control of optical tweezer arrays
Xiangkai Sun, Richard Bing-Shiun Tsai, Andrew Winnicki et al. (+4 more)
Optical tweezer arrays trapping ultracold atoms and molecules are a versatile
quantum science platform with broad impact in
quantum simulation,
quantum computation, and
quantum metrology. The polarization dependence of anisotropic vector and tensor light shifts offers a degree of freedom for precise
quantum state engineering of trapped particles. Similarly, state-selective operations depend on the polarization of the addressing light through the atom–light coupling strength. Yet, methods for controlling the polarization of individual tweezers or local addressing beams remain an experimental challenge. Here, we demonstrate independent, site-resolved linear polarization rotation across an optical tweezer array by exploiting the local birefringence tunability of a spatial light modulator. Applying this capability to an array of $^{88}\rm{Sr}$ atoms, we homogenize differential light shifts imparted by the 813-nm tweezers to the narrow ${^1{\rm S}_0}\leftrightarrow{^3{\rm P}_1}$ transition, enabling sideband cooling under a magic-angle condition. Furthermore, we show dynamic transport of trapped atoms across distinct polarization zones with high survival and preserved coherence. Finally, we characterize tweezer polarization noise and demonstrate closed-loop stabilization referenced to the atomic transition frequency, reaching mrad-level stability. Our work establishes a technique for manipulating polarization-sensitive atomic and molecular transitions in tweezer architectures, with immediate applications to optical tweezer clocks and multi-zone
quantum processors.
Read the full abstract ▾
All authors Xiangkai Sun, Richard Bing-Shiun Tsai, Andrew Winnicki, Yuan Le, Kon H. Leung, Nelson Darkwah Oppong, Manuel Endres
Networking
OctoSense: Building a Unified Ecosystem for Open-Source Wireless Sensing
Weiying Hou, Xie Zhang and Chenshu Wu
As AI enters the physical world, wireless sensing has emerged as a key modality for enabling non-intrusive physical intelligence. However, unlike the mature ecosystems of
vision and
language models, wireless sensing research remains highly fragmented. The field faces a growing "reproducibility wall" caused by heterogeneous data formats, non-interoperable processing pipelines, and the lack of standardized
benchmarks. To dismantle this barrier, we present OctoSense, a unified platform designed to propel wireless sensing toward an "ImageNet-style" research ecosystem. OctoSense introduces a holistic framework that decouples high-level model logic from ad-hoc data nuances through three distinct components: a unified data abstraction for streamlined
dataset access, standardized signal operators for efficient model development, and a rigorous
benchmark engine for rapid and fair comparison. To support existing
datasets and models, OctoSense integrates a comprehensive suite of widely used
datasets and models, reducing the effort required for data and model adaptation to a few lines of declarative code. We demonstrate the efficacy of OctoSense through a set of usage examples that reproduce representative models on popular
datasets. By providing a foundational open-source infrastructure, OctoSense enables community efforts to accumulate rather than fragment, paving the way for wireless sensing research that is interoperable, reusable, and reproducible by design.
Read the full abstract ▾
Robotics
A Physics-Informed Collision Learning Framework for Collaborative Robot Motion Generation
Chen Cai and Steven Liu
Close-proximity multi-arm manipulation requires collision models that are both geometrically accurate and differentiable enough for real-time optimization. Classical geometry checkers provide reliable distances but are difficult to use inside gradient-based model predictive control, while conservative proxy models can restrict tightly coupled motion. We present PI-UDF, a physics-informed unified differentiable framework for body-to-body collision distance prediction between articulated
robots. PI-UDF combines analytical forward kinematics with learnable link-geometry
embeddings and a shared residual network to predict pairwise inter-arm distances directly from
robot configurations. To improve safety-critical fidelity, we combine quota-driven boundary mining with an asymmetric boundary-crossing penalty that emphasizes false-safe sign errors near the collision boundary. The learned distance field is integrated into nonlinear MPC as a differentiable inter-arm clearance term. We validate the framework on a real dual-Franka platform through high-speed close-proximity 14-DoF dual-arm swapping, sustained single-arm dynamic evasion, and dynamic-evasion planning configurations with frozen, predicted, and target-switching treatments of the moving arm. Hardware experiments and offline Drake/FCL replay show that PI-UDF provides a differentiable inter-arm clearance estimate suitable for closed-loop collision-aware collaborative
robot motion generation.
Read the full abstract ▾
Computer Vision
ViSkill: Reinforcing VLM Agents with Evolving Visual-Native Skills
Hongxing Li, Dingming Li, Yixin Li et al. (+6 more)
Skill-augmented
agents improve sample efficiency by distilling successful trajectories into reusable strategies. Yet most existing approaches remain text-centric, linearizing spatial layouts and action-state correspondences into
language that loses critical
geometric structure. Recent efforts have begun incorporating visual evidence, but construct and update skills separately from policy optimization, leaving their mutual improvement underexplored. We propose ViSkill, a visual-native skill learning framework that encodes successful interactions as composite visual skill cards directly accessible to VLM
agents. Retrieved skills guide both inference and reward shaping, while successful trajectories are distilled back into the library, forming a closed feedback loop in which skill accumulation and policy improvement reinforce each other. An optional cold-start mechanism further accelerates early-stage learning. Evaluated on Sokoban, FrozenLake, and PrimitiveSkill, ViSkill achieves an overall success rate of 0.89, rising to 0.91 with cold-start initialization, outperforming all evaluated proprietary and open-source baselines while converging faster than standard PPO. Our code is available at https://github.com/ZJU-REAL/ViSkill.
Read the full abstract ▾
All authors Hongxing Li, Dingming Li, Yixin Li, Yong Du, Wenqi Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen
Computer Vision
SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models
Hongxing Li, Jinyue Su, Dingming Li et al. (+5 more)
Existing spatial
reasoning benchmarks mainly test spatial perception: reading off relations already visible in the input. Yet real-world spatial intelligence demands predictive spatial
reasoning: constructing a scene from observations, anticipating how an intervention changes it, and
reasoning about the unseen outcome. We introduce SpaceCast-Bench, the first
benchmark to directly and diagnostically evaluate this capability. Built around an observe-transform-infer framework, its 3,862 questions from 182 real-world scenes span 16 task types at three levels: static perception, local prediction, and global prediction, progressively requiring scene understanding, spatial state updating, and relational inference over unobserved outcomes. Evaluating 21 models exposes a stark gap: the strongest model reaches only 58.0% against 87.2% human performance, while spatially specialized models remain near random chance. Controlled analyses further reveal that bridge views are critical for integrating distributed observations, and that explicit 3D evidence benefits models more reliably than generated outcome images or
videos. Fine-tuning on our programmatically generated data lifts Qwen3-VL-4B from 34.0% to 65.7% with macro-average gains across six out-of-domain
benchmarks.
Read the full abstract ▾
All authors Hongxing Li, Jinyue Su, Dingming Li, Wenqi Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen
Machine Learning
Learning Kilometer-Scale Weather Prediction with Global-Regional Alignment
Guowen Li, Yang Liu, Yujie Wang et al. (+5 more)
Kilometer-scale regional weather
forecasting is essential for local weather warnings and weather-sensitive decisions. Existing data-driven approaches often rely on numerical forecasts for large-scale guidance or require additional training of global
forecasting components. Pretrained global weather models offer an efficient source of large-scale forecasts, motivating their reuse to guide high-resolution regional prediction. However, this coupling requires aligning global and regional representations across different grids and integrating global guidance with local interactions to advance regional states. We propose ScaleCast, a regional
forecasting framework that addresses these challenges through Global-Regional
Alignment. Its Global-Regional Conversion module aligns joint global and regional representations with regional locations, while the Global-Regional
Alignment and Dynamics block combines aligned guidance with regional neighborhood interactions. Experiments using ERA5 global analyses on a 0.25-degree grid and CERRA regional reanalysis at 5.5 km spacing demonstrate improved regional forecasts across surface and upper-air variables, with a single trained model supporting multiple global forecast drivers (i.e., Pangu-Weather, GraphCast, and HRES) without specific retraining. Fine-tuning on HRRR at 3 km spacing further demonstrates the framework's adaptability to a different regional domain and spatial resolution. Windstorm case studies show improved cyclone positioning and core-pressure estimates, while comparisons with HadISD station observations show closer agreement with local temperature and humidity changes.
Read the full abstract ▾
All authors Guowen Li, Yang Liu, Yujie Wang, Qiuyan Sun, Haoyuan Liang, Juepeng Zheng, Hong Cheng, Haohuan Fu
Quantum Physics
Measure Now, Mitigate Later: Virtual Error Cancellation for Logical Quantum Circuits
Oles Shtanko, Saswata Roy, Shashwat Kumar et al. (+1 more)
Early fault-tolerant
quantum algorithms are a key step in scalable
quantum computing, but are ultimately constrained by residual, uncorrectable logical errors. Logical
quantum error mitigation can eliminate this bias, however its standard requirements include prior noise characterization, more complex experiments, and substantial sampling overhead. Although no mitigation technique can simultaneously overcome all three constraints, we demonstrate that the first two—noise learning and need for modified
circuits—can be completely bypassed for logical
circuits by leveraging only syndrome records. Specifically, we introduce virtual error cancellation: a method that, unlike previous bias-free approaches, operates entirely in classical post-processing on universal
quantum circuits. We develop and numerically test end-to-end protocols that reduce the sampling overhead beyond fundamental limitations of existing syndrome-aware mitigation techniques, including syndrome-based postselection. In particular, we propose a scheme that pairs the low-latency decoder used in the experiment with a high-complexity decoder run in post-processing. In numerical simulations of a distance-7 rotated surface code, we show that our full-stack dual-decoder
algorithm exhibits an error that diminishes as the number of samples increases, capable of reaching more than three orders of magnitude of suppression compared to the decoded logical error rate in an experimentally achievable setting. These results establish syndrome records as a resource for correcting universal computational estimates after
quantum execution.
Read the full abstract ▾
All authors Oles Shtanko, Saswata Roy, Shashwat Kumar, Zlatko Minev
Computer Vision
SpaceFlow: Locally Controllable 3D Generation
Neil De La Fuente, Joan Lafuente, Mukhammadali Sayfiddinov et al. (+5 more)
Current 3D generation methods lack explicit local control:
geometric adherence is often defined by a global control strength, and appearance cannot be specified locally. We present SpaceFlow, a training-free pipeline for locally controllable 3D generation from text descriptions and a collection of
geometric primitives. Each primitive serves as a proxy for an object part and is assigned a local control level, enabling users to specify whether regions should strictly follow the input shape or allow generative completion. During structure generation, we enforce these spatial constraints within the generative flow process. For appearance synthesis, the generated structure is segmented and matched to the primitives. Each generated part is conditioned only on its assigned text or image cue, thereby limiting cross-part leakage. Regional geometry metrics demonstrate that SpaceFlow preserves the specified geometry in high-control regions and enables plausible shape variation in low-control areas. A user study further indicates that the resulting balance between
geometric fidelity and generative freedom remains competitive in overall quality. When evaluating appearance on fixed geometry, text-conditioned routing achieves state-of-the-art
prompt faithfulness and color/material accuracy. Qualitative results additionally show localized routing of image cues. The project page is available at SpaceFlow3D.github.io.
Read the full abstract ▾
All authors Neil De La Fuente, Joan Lafuente, Mukhammadali Sayfiddinov, Felicia Scharitzer, Marc Pollefeys, Ata Celen, Sayan Deb Sarkar, Elisabetta Fedele
Astrophysics of Galaxies
ACORN I. Massive Black Hole Seeding and Tidal Disruption Events from Star Clusters in Cosmological Simulations
Yihao Zhou, Tiziana Di Matteo, Claire E. Williams et al. (+6 more)
JWST has revealed massive
star clusters at $z\gtrsim 6$, which provide promising environments for massive black hole (MBH) seed formation. We introduce a new MBH seeding model for cosmological simulations, physically motivated by the runaway collisions in
star clusters. Subgrid
star clusters are generated according to the local ISM and evolved through
stellar evolution and two-body relaxation. The masses of MBH seeds formed through runaway collisions are determined by the properties of their host clusters, including the effects of
metallicity-dependent
stellar winds. We further follow
tidal disruption events (TDEs) with the evolving clusters, which can be enhanced in MBH binaries, and account for gravitational-wave (GW)
recoil following MBH mergers. We implement this model in a constrained simulation of a 5$\sigma$ overdense region, where seeds of $10^{3}-10^{5.6}\,M_\odot$ form in clusters with $10^{4}\leq m_{\rm SC}\leq 10^{8}\,M_\odot$. Because both the seed abundance and mass are directly connected to the local environment, we have a more physical $M_{\rm BH}\!-\!M_{\rm galaxy}$ scaling for low-mass
galaxies ($M_{\rm galaxy}\sim10^{6-9}\,M_\odot$), a regime poorly described by previous seeding models. Multiple MBHs are naturally allowed to form in-situ within individual
halos, producing a spatially clustered population. This increases the MBH merger rate by $\sim100$ relative to other seeding models, significantly enhancing the expected LISA event rate. TDEs can dominate the growth of $10^{3-4}\,M_\odot$ MBHs and produce luminous
flares of $L_{\rm peak}\sim 10^{42-43}\,{\rm erg\,s^{-1}}$, increasing the observability of low-mass seeds. This seeding prescription accelerates early MBH growth at $z\gtrsim 8$, while subsequent self-regulated gas
accretion drives the central MBH toward similar masses $\sim10^{8.6}\, M_\odot$ by $z=6$ across different seeding models.
Read the full abstract ▾
All authors Yihao Zhou, Tiziana Di Matteo, Claire E. Williams, Aklant Kumar Bhowmick, Marta Reina-Campos, Carl L. Rodriguez, Elena González Prieto, Laura Blecha, Rupert Croft