Search arXiv⌕ Search

arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 19 recordsLinked to original sources
arXiv2026

WMEval: Evaluating Interactive World Models Grounded on Physical Simulations

Interactive world models hold immense potential for high-stakes applications in robotics, game engines, and immersive digital twins, yet existing benchmarks primarily focus on perceptual similarity rather than the physical-causal coherence essential for real-world utility. We introduce WMEval, a comprehensive benchmark that evaluates interactive world models against Unity-based physical simulations. By leveraging Unity as a high-fidelity ground-truth oracle, we shift evaluation beyond perceptual visual assessment toward the objective state-based verification required for downstream applications. Spanning 9 diverse scenes, WMEval assesses action controllability, scene and character consistency, prompt adherence, kinematics, fluid mechanics, physical causality, and lighting. By executing matched action sequences in both the simulator and the world model, WMEval enables evaluation at both the perceptual level and the physical-causal level, using simulator-derived states as reference. Experiments on state-of-the-art models reveal significant room for improvement in dimensions that remain hidden by standard perceptual metrics but are vital for real-world deployments. Our code is available at: https://github.com/hao-ai-lab/WMEval.

cs.CV↗
arXiv2026

CASEL: Consequence-Aware Selective Execution for Reliable Long-Horizon Tool-Using Agents

Tool-using language agents increasingly modify external systems, yet most runtime safeguards allocate scrutiny by generic uncertainty scores or coarse mutation heuristics. We introduce CASEL (Consequence-Aware Selective Execution Layer), a model-agnostic runtime allocation policy that combines calibrated failure probability, state-dependent consequence, reversibility, ambiguity, and evidence freshness to choose, for each proposed tool call, among direct execution, verification, sandboxing, clarification, approval, and abstention by minimizing a per-mode cost. We formalize direct execution as selective prediction and state and test the validity conditions of a finite-sample threshold and a trajectory risk ledger. In a controlled simulation over 20 seeds with 120,000 test actions each, in which every method has the same intervention menu and the same compute cost, CASEL reduces harmful-side-effect trajectories from 14.75 percent for an uncertainty-ranked gate to 7.02 percent and mean normalized harm from 0.184 to 0.123. A gate ranked by the product of failure probability and consequence reaches similar harm (5.82 percent, 0.120) but sends 45 percent more actions to human approval, and removing reversibility-aware routing raises the rate to 15.53 percent. A trajectory-level risk budget, in contrast, increases harm at equal cost (11.82 percent). Intervention effectiveness is assumed, not measured: the results rank allocation rules under the simulator's assumptions and do not establish deployment effectiveness.

cs.AI↗
arXiv2026

JEPA-Inspired Latent Field Prediction for Sparse-Pilot MIMO-OFDM Channel Reconstruction

Accurate acquisition of full channel state information (CSI) is essential for multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) wireless systems, while dense pilot-based estimation incurs substantial overhead. Directly minimizing the prediction error over raw CSI coefficients is fragile in realistic propagation environments, since small-scale fading, mobility, local scattering, and small spatial perturbations can induce large phase-sensitive variations. Nevertheless, wireless channels contain strong spatial, temporal, and frequency-domain correlations that can be more predictably represented in a latent propagation-field domain. Motivated by this observation, we propose a joint-embedding predictive architecture (JEPA)-inspired field world model, defined as a latent predictive model of the wireless propagation field, for sparse-pilot channel reconstruction. The model maps CSI observations into latent time-frequency tokens, predicts masked or future latent field states from historical multi-band context, and fuses the predicted latent prior with sparse noisy current pilots to reconstruct the full MIMO-OFDM channel. Ray-tracing experiments show that the proposed method improves normalized mean square error (NMSE) over pilot-only and direct reconstruction baselines, supporting latent field prediction as an effective prior for sparse-pilot channel acquisition.

eess.SP↗
arXiv2026

Diagnosing and Correcting Bias in MLLM for Long Video Understanding

Long-video question answering is challenging since the answer often hinges on a few decisive moments scattered throughout the video, while memory constraints force Multimodal Large Language Models (MLLMs) to sample frames at extremely low rates, resulting in sparsity that risks omitting key evidence. Though aggressively enlarging the context size is intuitive, we argue that performance remains fundamentally constrained by the bias, \emph{i.e.}, the misleading cues caused by spurious \textit{linguistic} correlations and salient yet irrelevant \textit{visual} observations. In this paper, we first introduce {S-MME}, a benchmark designed to systematically diagnose shortcut-biased behavior in MLLMs. To mitigate this failure, we further propose a training-free framework, Counterfactual Long-video Evidence-Aware Reasoning (CLEAR). By treating the full video-question pair as a complete view, CLEAR estimates bias effects via augmented views by only keeping the question or locally salient visual clips. Then, we intervene hidden states to mitigate the discrepancy at inference such that the correct answer can be predicted confidently. With experimental studies across five benchmarks, we show that our method is generalizable and can be integrated in different MLLMs, yielding consistent gains in prediction and robustness. We hope CLEAR can offer a complementary direction in test-time scaling for long-video understanding.

cs.CV↗
arXiv2026

AgentBaton: Efficient Model Collaboration in Agentic Tasks

Solving long-horizon agentic tasks requires language agents to interact with their environments over multiple steps, and reliable performance often comes from the expensive capable models available through cloud providers. Meanwhile, alternative efficient models with inferior performance but a much lower cost are emerging, including lower-tier cloud models and smaller open-weights models deployable on local hardware. This raises a natural question that remains less explored than traditional query-level routing: can capable and efficient models collaborate within a single long-horizon agentic task to substantially reduce cost while preserving task accuracy? In this paper, we systematically investigate key design choices for such collaboration through model handoffs, where one model takes over from another, focusing on handoff direction and context transfer. We find that efficient-to-capable handoffs generally underperform capable-only execution with limited or even no cost reduction, whereas capable-to-efficient handoffs at suitable stages of task progress largely preserve or even improve accuracy while substantially reducing cost. These findings suggest that a one-time, simple, well-timed handoff strategy is good enough for cost-effective agent collaboration, motivating AgentBaton, a lightweight detector that learns when to hand off from recurring patterns in task progress. Across 11 model pairs and two representative benchmarks, AgentBaton reduces API costs by 65.9% on average while largely preserving task accuracy, with accuracy gains of up to 8.8 percentage points over capable-only execution. Experiments across various reasoning-effort settings further show that AgentBaton extends the empirical accuracy-cost Pareto frontier.

cs.MA↗
arXiv2026

WorldCast: World-Consistent Character Video Generation with Large-Angle Camera Control

Precise camera control is essential for cinematic character video generation, yet existing methods struggle to follow large-angle trajectories with combined rotation, translation, and radius variation from arbitrary initial azimuths. Such viewpoint changes also reveal character regions unseen in the initial frame, causing geometric distortions, appearance degradation, and identity drift. To systematically study these coupled challenges, we introduce CLACT, a large-scale Character-centric dataset and benchmark covering diverse Large-Angle Camera Trajectories, with accurate camera geometry and multi-view character observations for training and evaluation. We further propose WorldCast, a framework for world-consistent character video generation under diverse large-angle camera motions. Specifically, we back-project the initial frame into a 3D point cloud using estimated depth, align the point-cloud scale with the target camera trajectory, and render it along the trajectory to obtain explicit geometric guidance and coverage masks. To complement the incomplete character information, we introduce Geometry-Aligned Memory Attention (GAMA). Using relative camera poses estimated jointly from the initial frame and multi-view character anchors, we express anchor and target-view geometry in a common coordinate frame. GAMA uses this geometric alignment to guide the retrieval and injection of view-relevant character features during generation. Evaluations across diverse large-angle camera trajectories demonstrate accurate trajectory following while preserving character geometry, appearance, and identity.

cs.CV↗
arXiv2026

LAST-V2X: Language-Driven Adaptive Scenario Synthesis and Testing for Communication-Induced Failures in V2X Safety Applications

Vehicle-to-everything (V2X) communication enables connected vehicles to make safety-critical decisions, but delayed, missing, or incorrect information can compromise driving safety. Systematically testing such failures is challenging because traffic interactions, communication conditions, and application responses must be jointly configured to produce meaningful scenarios. This paper presents LAST-V2X, a language-driven, adaptive framework for automated scenario synthesis and safety testing of V2X applications. LAST-V2X leverages large language models (LLMs) to translate natural-language test descriptions into executable scenarios grounded in roadway geometry, V2X communication semantics, and vehicle behavior. The framework integrates physical traffic simulation with spatially aware communication modeling and message-level fault injection to evaluate how information degradation affects vehicle decisions. To establish whether an observed safety failure is caused by communication rather than the underlying traffic conditions, LAST-V2X employs matched executions under degraded and ground-truth information. Furthermore, an execution-guided refinement mechanism combines parameter search and LLM self-reflection to identify ineffective test configurations, improve scenario effectiveness, and produce minimal reproducible failure cases. Experiments across diverse CARLA environments and V2X safety applications demonstrate the framework's ability to generate reliable scenarios, distinguish communication-induced failures from unrelated safety violations, and substantially improve failure discovery through adaptive refinement. These results establish LAST-V2X as a systematic approach to language-driven, causally verified, and adaptive safety testing for connected vehicle systems.

cs.RO↗
arXiv2026

ExtremControl: Low-Latency Humanoid Teleoperation with Scaled Velocity Feedforward

Building a low-latency humanoid teleoperation system is essential for collecting diverse reactive and dynamic demonstrations. However, end-to-end latencies in existing approaches consistently hover around 200ms, roughly equal to human reaction time, that severely limits responsiveness and prevents tasks requiring rapid feedback. We identify position-only PD control as the root cause of this latency floor and propose ExtremControl, a low-latency whole-body control framework that: (1) incorporates scaled velocity feedforward in low-level control, reducing tracking delay by approximately 70ms with a provably optimal feedforward ratio; (2) operates directly on SE(3) poses of selected rigid links, primarily humanoid extremities, to avoid joint-space retargeting in the control loop. To minimize retargeting time, we derive a closed-form Cartesian-space mapping that uses a single anthropomorphic calibration step to translate human motion into humanoid link targets, accommodating operators with varying body proportions. Furthermore, we provide a theoretical framework for selecting control hyperparameters, and validate the effectiveness of velocity feedforward control in the reinforcement learning for locomotion setting through experiments in both simulation and the real world. Building on ExtremControl, we implement a teleoperation system supporting both optical motion capture and VR-based tracking, achieving end-to-end latency as low as 50ms --- well beneath the threshold at which system delay adds a full human reaction cycle to the control loop. This enables a qualitatively new class of reactive behaviors, including ping-pong ball balancing, juggling, and real-time return, that are beyond the capability of existing teleoperation systems.

cs.RO↗
arXiv2026

RayCo4D: Ray-Conditioned Continuous Motion Model for Monocular 4D Dynamic Reconstruction

Reconstructing dynamic 3D scenes from monocular video requires inferring motion from noisy and intermittently visible observations. Long-range tracks provide valuable geometric constraints, but integrating their fragmented temporal coverage into shared, continuous motion remains challenging. We introduce RayCo4D, a ray-conditioned continuous motion model for monocular 4D reconstruction that explicitly separates visual measurements from latent 3D motion states. Each track is associated with a reference-time 3D query constrained to its camera ray, and its image and depth observations constrain shared latent motion nodes through local many-to-many associations. Each node maintains an $\mathrm{SE}(3)$ pose and body-frame velocity at sparse support times. A white-noise-on-acceleration process prior couples neighboring states and defines continuous-time interpolation, allowing observations across visibility intervals to jointly constrain the same latent state sequence. The inferred shared motion is transferred to Gaussian primitives through dual-quaternion blending, while reference-anchored local $\mathrm{SE}(3)$ residuals capture fine-scale non-rigid deformation beyond the shared scaffold. Experiments on diverse monocular dynamic scenes show that RayCo4D achieves competitive novel-view synthesis quality, improves structural and temporal consistency under rapid motion and incomplete tracks, and supports continuous-time rendering.

cs.CV↗
arXiv2026

ExGrad: Beyond Suppression - Explicit Clean-Gradient Subtraction in Adversarial Training

Adversarial training derives updates from surrogate losses, but source-loss descent does not directly characterize an update's effect on adversarial margins. We introduce ExGrad (extrapolative gradient recomposition), which constructs clean-input, adversarial-input, and adversarial-weight-perturbed (AWP) gradients. Virtual steps measure their margin responses, and a globally solvable three-variable problem selects coefficients over a bounded affine domain without nonnegativity constraints. Across affine radii on CIFAR-10, negative clean coefficients rapidly predominate and increase in magnitude with the radius. Cross-zero interventions show that setting the clean coefficient to zero is insufficient: mild subtraction improves PGD-20, while stronger subtraction trades clean accuracy for further PGD-20 gains. In CIFAR-10 WRN-28-4 controls, none of the three candidate-removal variants matches full ExGrad. Projection diagnostics show that negative clean coefficients usually coexist with positive net clean-axis projections, although both decrease as the radius grows. Norm-matched controls indicate that update magnitude alone does not explain the moderate-radius gains. Experiments on CIFAR-10, CIFAR-100, and ImageNet-100 show radius-dependent improvements over AWP. On CIFAR-100, ExGrad at $R=0.8$ improves all five reported metrics, including clean and AutoAttack accuracy by 3.84 and 1.75 percentage points. On ImageNet-100, the same radius improves all four robust metrics over both AWP and PGD-AT. These results identify clean-gradient subtraction as a response-revealed update preference whose strength and clean-robust operating point are controlled by the affine radius.

cs.LG↗
arXiv2026

Ask to Distill: Learning to Synthesize Data for On-Policy Distillation

On-policy distillation (OPD) can succeed with surprisingly few training questions, motivating a view of questions as carriers of distributional knowledge transfer. How can we assess whether a newly generated question is valuable for distillation, and learn to generate better questions---without access to its correct answer? We introduce Ask-to-Distill (\method), a framework that treats synthetic questions as distribution elicitors and optimizes the teacher supervision they expose along student-generated trajectories. A student-initialized synthesizer is optimized using a distributional surrogate for question value, computed from a frozen student--teacher pair without reference answers, correctness rewards, or per-candidate student updates. The reward combines teacher--student disagreement with teacher probability mass on the student's candidate support, together with length and diversity controls. The generated questions are filtered and deduplicated into a prompt-only bank for standard OPD. Across mathematical reasoning, code generation, and agentic tasks, our method improves average performance and teacher-gap recovery over unoptimized student-generated prompts and KL-filtered real questions, while matching or exceeding real-bank OPD in the evaluated settings. Signal analyses further show that optimized questions can elicit denser supervision and more total signal from shorter student responses, supporting elicited distributional supervision as an objective for synthetic distillation data. Our code, data and models are publicly available at https://github.com/GasolSun36/Ask_to_Distill.

cs.AI↗
arXiv2026

Joint Subcarrier and Power Allocation for Localization-Oriented Multi-Carrier Underwater ISAC Systems

Multi-carrier underwater integrated sensing and communication (ISAC) realizes concurrent acoustic communication and localization, yet the frequency selectivity of the acoustic channel creates uneven positioning capabilities across subcarriers. This paper proposes a localization-oriented subcarrier and power allocation scheme for multi-anchor underwater ISAC systems. We build a unified orthogonal subcarrier multiplexing framework and derive closed-form Cramer-Rao bounds (CRBs) under frequency-selective channels with colored noise. We derive a subcarrier-level localization information contribution from the time of arrival Fisher information, which enables localization-oriented resource allocation. We derive several theoretical properties regarding the distance and frequency dependence of the localization information metric, and establish a distance-frequency matching principle. These results provide a theoretical foundation for localization-oriented resource allocation and guide the design of a low-complexity greedy algorithm that improves positioning accuracy. Specifically, the algorithm iteratively allocates subcarriers to minimize the residual positioning-error objective, thereby improving positioning accuracy while satisfying communication quality-of-service (QoS) constraints. Simulations verify that our method yields much lower positioning error bound than uniform, random and high-frequency-first subcarrier allocation in symmetric or asymmetric topologies while guaranteeing communication QoS, achieving superior communication-localization tradeoffs.

eess.SP↗
arXiv2026

MCDA: Multi-View Graph Contrastive Learning with Dynamic Topology Augmentation for Functional MRI Analysis

Resting-state fMRI (rs-fMRI) is a vital non-invasive tool for diagnosing brain disorders and uncovering pathological mechanisms. However, existing graph contrastive learning (GCL) methods often underexploit augmentation histories across training epochs and the relationships among different augmented views. In this paper, we introduce a topology-guided Multi-view graph Contrastive learning framework with Dynamic graph topology Augmentation (MCDA) for fMRI analysis. Unlike traditional GCL that uses fixed noise or edge-dropping, MCDA employs a Dynamic Augmentation Update (DAU) module that iteratively refines the augmented topology. Specifically, we incorporate self-aware and cross-sample augmentation strategies to simultaneously capture individual-specific functional signatures and inter-subject heterogeneity. We then design a topology-guided multi-view graph contrastive learning scheme that explores interrelationships not only between the original network and its augmented views, but also across different augmented views to maximize structural information preservation. We characterize the score-level optimization directions and establish a lower bound for the contrastive objective. The framework combines classification and contrastive losses for brain disorder diagnosis and examines discriminative functional connectivity patterns as potential biomarkers. Experiments on two public rs-fMRI datasets comprising 1,236 subjects demonstrate that MCDA outperforms the state-of-the-art methods. The code has been made publicly available at: \url{https://github.com/goodcodingagoodboy/MCDA}.

cs.CV↗
arXiv2026

Large values of derivatives of quadratic Dirichlet $L$-functions

e study large values of fixed derivatives of quadratic Dirichlet $L$-functions on the real segment $1/2\leqσ\leq1$. Assuming GRH, we prove lower bounds in terms of GCD sums and finite Euler products. At the central point, every fixed derivative order is compared with the same square-free GCD sum. In the mesoscopic band $σ=1/2+A/\log_2X$, a renormalized resonator estimates contributions from both halves of the differentiated approximate functional equation. The resulting lower bound has a relative error. On a smaller boundary scale, a common finite product appears in both contributions to this lower bound. This allows us to locate the crossover between their derivative weights. The coarser exponential bounds have constant $1$ at the center and $\mathrm e^{-A}$ in the mesoscopic band. For fixed $1/2<σ<1$, the lower bound includes the full typical logarithmic weight of the resonating Euler product. Burgess truncation improves the exponential constant when a shorter cutoff allows a longer resonator. Near the $1$-line, our lower bounds involve Dickman constants on the $1/\log_2X$ scale. We also prove a uniform transition estimate outside a shrinking neighborhood of the central point. All results hold in the positive even subfamily of fundamental discriminants.

math.NT↗
arXiv2026

Noncommutative $L_p$-Poincaré inequalities

We derive $L_p$-Poincaré inequalities for quantum Markov semigroups on finite von Neumann algebras, from a spectral gap (PI) or from a modified logarithmic Sobolev inequality (MLSI). For $p\ge2$, the corresponding coefficients have optimal order $O(p)$ and $O(\sqrt p)$, respectively. Both estimates follow from Beckner inequalities by integration along semigroups. We also establish a non-tracial extension to quantum Markov semigroups satisfying GNS detailed balance. As a direct byproduct, we apply $L_p$-Poincaré inequalities to Schatten-valued functions on graphs in order to obtain the Schatten distortion of expanders.

math.OA↗
arXiv2026

LoGRA: Scaling LLM Reinforcement Learning with Low-Rank Gradient Sketches

Reinforcement learning has greatly advanced the capabilities of large language models, but its memory demands remain a barrier to broader adoption. We introduce LoGRA, an approach to RL post-training that reduces memory by retaining useful learning signals in low-rank gradient sketches. These compact representations support both model updates and efficient policy synchronization. To prevent overly large updates from disrupting learning, we complement gradient compression with predicted-KL step control, which estimates policy changes before applying each update and adjusts its magnitude accordingly. With all techniques combined, LoGRA reduces average training memory usage by up to 45.7% across reasoning tasks without compromising performance. It also enables stable training of a 27B-parameter model for over 1,100 steps on a single eight-GPU node, where dense Adam runs out of memory, making previously memory-infeasible RL training practical. Code is available in the \href{https://github.com/skzhang1/labs-molt/tree/logra/examples/scripts/logra}{Molt library}.

cs.CL↗
arXiv2026

UBASE: An AI Search Engine for Trillion-Scale Vector Data Management at ByteDance

Since 2016, UBASE has been the foundation of ByteDance's search infrastructure, scaling to more than 7,000 clusters and 300 PB of indexed data. Driven by the demands of AI workloads, UBASE has evolved from a text search engine into a unified AI search system supporting vector retrieval, lexical matching, and predicate filtering. Its largest deployment indexes nearly one trillion high-dimensional vectors. This scale exposes two central bottlenecks in AI-era retrieval: memory-intensive graph-index construction under sustained ingestion, and the prohibitive cost of keeping vector indexes entirely in memory. UBASE addresses these bottlenecks with two techniques. First, it introduces a quantization-aware vector kernel based on SymRaBitQ, a new symmetric quantization scheme with tight theoretical guarantees that allows index construction to run directly in the quantized space accurately and efficiently without retaining a copy of full-precision vectors. Second, it provides a hybrid storage engine that supports memory-resident, hybrid, and SSD-resident deployments, with fine-grained record-level caching to trade memory for latency under operational control. On large-scale benchmarks, UBASE improves throughput by up to 3x, reduces indexing memory by 80%, and lowers serving cost by up to 99.4% compared with prior systems, while supporting trillion-vector scale, write-heavy or latency-sensitive workloads in production.

cs.DB↗
arXiv2026

Programmable Hard-X-Ray Spectral Comb Generation via Laser-Compton Scattering

Independent control of line spacing and offset underpins programmable frequency-comb spectroscopy, yet it has not been established for hard-X-ray sources based on inverse Compton scattering after electron transport and angular collection. Self-consistent multi-particle tracking and Li'enard-Wiechert radiation calculations model a transported 10 pC beam and its angle-integrated spectrum near 11 keV. Laser pulse separation programs the comb spacing, while pulse-to-pulse phase translates the entire line grid at fixed spacing. The spacing model agrees within 0.52%, while transport and optics change the phase-transfer slope by less than 0.2%. At fixed electron spot size, the largest resolvable separation scales as the inverse square of normalized emittance; subnanometer bunches also produce a mode-locked train with 13.9 as spacing. This two-coordinate programmability enables phase-stepped multicolor measurements and offers routes to hard-X-ray Ramsey interferometry, coherent control of core and nuclear excitations, and attosecond nonlinear spectroscopy.

physics.optics↗