arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2026

A deterministic Fokker--Planck/lattice-Boltzmann micro--macro solver for dilute Hookean polymer solutions

Deterministic micro--macro simulation of polymer solutions requires the coupled evolution and spatial transport of a configuration distribution, together with the conversion of its moments into macroscopic stress and feedback to the flow. We develop a Fokker--Planck/lattice-Boltzmann solver for this coupling in two-dimensional dilute Hookean polymer solutions. The solver represents the distribution on the unbounded configuration space by Hermite coefficient fields, advances local configuration dynamics and physical-space transport in separate steps, and recovers Kramers stress from the second moment to achieve two-way coupling with a purified two-relaxation-time lattice-Boltzmann flow solver. To assess whether this configuration description recovers the corresponding macroscopic response, we use the exact second-moment closure of the continuous Hookean model as a macroscopic reference for comparison with analytical and independently discretized macroscopic solutions. The calculations reproduce velocity overshoot and damped oscillations in start-up Poiseuille flow, with velocity profiles approaching the analytical start-up transient under grid refinement; in four-roll flow with spatially nonuniform extension and stress feedback, the maximum full-domain relative differences in velocity and polymer stress are approximately $0.0309\%$ and $3.84\%$, respectively, over the tested Weissenberg numbers at corresponding sampling times. These comparisons support the solver's ability to recover the Hookean macroscopic response from configuration-distribution evolution. Further calculations of wall-bounded recirculation and open cross-slot flow exhibit the corresponding conformation responses and symmetric and asymmetric flow states, extending this deterministic micro--macro method to viscoelastic flows under different boundary constraints.

physics.flu-dyn↗

arXiv2026

Improving Video Sparse Attention with Fine-grained Router and Sparse Rebasing

We present VSA2, a frontier trainable sparse attention for video DiTs. VSA2 includes a variety of new architectural features and training procedures that we apply across all stages of the DiT development cycle, including pretraining, RL, and inference, to produce a DiT with comparable or better quality than a full attention counterpart. Architecturally, VSA2 introduces a fine-grained router that improves the precision of identifying critical tokens and supports dynamic computation by allowing each query to attend to a variable number of key-value pairs. In training, we identify a Hard-to-Easy Curriculum, where models trained under high sparsity and later evaluated with lower sparsity during inference not only generalize effectively, but also outperform models trained with full attention in motion quality. VSA2 is also flexible: it can replace full attention during the middle of progressive low-to-high resolution pretraining, rebasing early-stage full-attention checkpoints. Experiments show that VSA2 reduces attention computation by half over VSA with lower loss. On 720p videos, it accelerates attention by 8.9x and end-to-end generation by 4.62x compared to the FlashAttention-3 baseline, while achieving comparable or better video quality.

cs.CV↗

arXiv2026

Multimodal Thinking with Renderable Programs

Current vision-language models (VLMs) excel at visual content understanding and text-based reasoning, yet their structure limits the advancement of incorporating images into the reasoning chain. Though Omnimodal models have made efforts in unifying text and image generation, they focus on visual tasks in the open-domain, lacking tractability due to rasterized or latent representations of images. We introduce SVGLM, a framework that uses scalable vector graphics (SVG) primitives to connect text and image in reasoning tasks. We exploit the duality of SVG as both image description and text instructions, yielding a more compact, interpretable solution to equip general VLMs with the capability of generating images within the reasoning process. We provide a large curated dataset of SVG-based image editing dataset, as well as the paradigm to tune open-source VLMs. Experiments on a mathematical reasoning benchmark demonstrate that SVGLM achieves strong SVG generation power as well as think-with-image intelligence. Our results highlight SVG as a suitable medium for building more robust digital domain agents, bridging the gap between text-based thinking and pixel-based images.

cs.CV↗

arXiv2026

A Harness for Synthesizing Diverse Naturalistic Full-Duplex Conversations

Full-duplex dialogue systems, which listen while speaking, must distinguish a completed turn from a pause within a turn and an interruption that requests a turn from a brief acknowledgment or speech addressed to a third party. Yet existing conversational corpora provide limited control over these events and limited labels for their intent. We present a pipeline for synthesizing intent-labeled, two-channel conversational speech from relational event lists. An LLM authors each event's speaker, text, conversational act, and attachment to an earlier event without predicting absolute timestamps. Events are synthesized independently, aligned with their source text, and placed on a shared clock, so turn-taking landmarks are measured from the rendered signal while silence durations are specified or sampled from turn-taking distributions. The pipeline covers 42 phenomena across eight families in English and Mandarin, derives frame-level system actions from authored intent, and promotes diversity using small, diverse sets of prior examples and batch prompts that request alternatives with self-reported probabilities. Ablations show gains in each targeted diversity dimension. On a four-action label space for taking, holding, releasing, and not holding the conversational floor, a semantic voice-activity detector using only current and past audio reaches start-speaking and start-listening F1 scores of 0.819 and 0.802. When generating its own responses, the full-duplex speech model Moshi takes 0.85 of the reference turns after fine-tuning on the generated corpus, compared with 0.44 before fine-tuning. Its frame-level precision for predicting system-floor occupancy rises from 0.46 to 0.88. With reference context at each step, its frame-level floor F1 rises from 0.893 to 0.962. These results show that controlled synthesis can provide learnable and transferable supervision for full-duplex turn management.

eess.AS↗

arXiv2026

3dgs-sc: a controlled static screen-content benchmark for 3d gaussian splatting

Does high image fidelity imply readable screen content in 3D Gaussian Splatting (3DGS)? We introduce 3DGS-SC, a controlled static screen-content dataset and benchmark for examining this mismatch. Ten procedural scenes provide fixed multi-view splits, exact cameras, screen masks, text boxes, and transcripts. The protocol separates whole-image fidelity, screen-region fidelity, OCR readability, and edge preservation. In the reported five-method comparison, LightGaussian exceeds Mip-Splatting in screen PSNR by only 0.08 dB, yet trails it in OCR accuracy by 12.7 percentage points. Across all ten method pairs, screen-PSNR and OCR orderings disagree in seven cases. These aggregate results expose a method-selection failure of fidelity-only evaluation. Complementing prior text-aware 3DGS research, 3DGS-SC targets controlled monitor interfaces and exact annotations; scene-wise robustness and acquisition effects remain open validation questions.

cs.CV↗

arXiv2026

Hypercontractivity for a family of quantum Ornstein-Uhlenbeck semigroups

We show that a family of quantum Ornstein-Uhlenbeck semigroups is hypercontractive. We also obtain the optimal order of the optimal time up to a constant. The main ingredient of our proof is Meixner polynomials. The goal of this paper is twofold: to provide more examples of hypercontractive quantum Markov semigourps on non-tracial von Neumann algebras, and to determine the optimal order of the optimal time for quantum Ornstein-Uhlenbeck semigroups.

math.FA↗

arXiv2026

Functionalization via Structure Completion and Motion Rectification

Acquisition and creation of 3D assets have been largely view- or appearance-driven. As a result, existing digital 3D models often lack the requisite structural components to function as intended, such as joints, supports, interiors, or interaction elements. At the same time, even human-annotated motions are frequently error-prone, leading to physically implausible behavior. We introduce object functionalization, a novel task aimed at transforming visually plausible but non-functional 3D models into functional and physically operable ones. We formulate functionalization as a graph completion problem over a new functional graph representation, where labeled nodes represent object parts, labeled edges encode functional and contact relations, and movable nodes carry motion attributes, so that structural functional deficiencies manifest as missing nodes or incorrect edges. We develop a neural Graph Functionalizer (GraFu) to complete an incomplete graph representing a non-functional 3D object. The completed graph then drives a geometry realization stage that instantiates predicted connectors and structural elements in 3D, with the compelling side effect of rectifying erroneous human-annotated and predicted motions. To support training and evaluation, focusing on furniture as a rich and challenging target category, we introduce FurFun-233, a dataset of 233 paired non-functional and functionalized furniture models. On PartNet-Mobility ("zero-shot") and HSSD test sets, our method matches state-of-the-art methods in motion prediction accuracy while substantially improving functionality in terms of collision and connectivity. Project page: https://mingrui-zhao.github.io/Functionalization/

cs.CV↗

arXiv2026

Ultra-fast Neural Inference for Stochastic Gaussian Splatting Denoising

Stochastic rendering eliminates the sorting and alpha blending process in Gaussian splatting, at the cost of introducing spatial noise. Formulating temporal denoising over the pixel stream shared by view-consistent stochastic splatting renderers, we propose a temporal neural denoiser validated on stochastic 2D Gaussian Splatting rendering, combining dual-path exponential moving average accumulation, per-pixel learned trust prediction for history validation, a fixed anisotropic spatial filter and a variance-gated composition with stabilization. The denoiser suppresses the noise, achieving temporally stable, visually compelling outputs during free camera navigation, all while retaining the sort-free, blend-free rasterization performance. The combined pipeline retains a PSNR gap to sorted alpha-blending renderers, but the denoiser's overhead stays below the time saved by removing sorting and blending.

cs.CV↗

arXiv2026

Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning

Agentic reinforcement learning requires rapid experimentation with agents and learning algorithms, yet large policies and long, multimodal trajectories demand substantial distributed infrastructure. We present MOLT, a lightweight, PyTorch- and Hugging Face-native framework that brings these goals together through four contributions. MOLT combines direct loading of Hugging Face models with experimentally validated trillion-parameter scalability in approximately 9.2K lines of framework code. Unified OpenAI- and Anthropic-compatible interfaces integrate existing agents with automatic handling of context compaction. Fully asynchronous training overlaps agent rollouts and policy optimization, accommodating variable agent execution times. Distributed experience storage removes centralized rollout-memory bottlenecks for long, multimodal trajectories. We experimentally validate the complete RL training pipeline on a one-trillion-parameter policy and demonstrate sustained learning with a 30B mixture-of-experts agent, establishing MOLT as a lightweight foundation for large-scale agentic RL research.

cs.LG↗

arXiv2026

Lingshu: A Generalist Foundation Model for Unified Multimodal Medical Understanding and Reasoning

Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in understanding common visual elements, largely due to their large-scale datasets and advanced training strategies. However, their effectiveness in medical applications remains limited due to the inherent discrepancies between data and tasks in medical scenarios and those in the general domain. Concretely, existing medical MLLMs face the following critical limitations: (1) limited coverage of medical knowledge beyond imaging, (2) heightened susceptibility to hallucinations due to suboptimal data curation processes, (3) lack of reasoning capabilities tailored for complex medical scenarios. To address these challenges, we first propose a comprehensive data curation procedure that (1) efficiently acquires rich medical knowledge data not only from medical imaging but also from extensive medical texts and general-domain data; and (2) synthesizes accurate medical captions, visual question answering (VQA), and reasoning samples. As a result, we build a multimodal dataset enriched with extensive medical knowledge. Building on the curated data, we introduce our medical-specialized MLLM: Lingshu. Lingshu undergoes multi-stage training to embed medical expertise and enhance its task-solving capabilities progressively. Besides, we preliminarily explore the potential of applying reinforcement learning with verifiable rewards paradigm to enhance Lingshu's medical reasoning ability. Additionally, we develop MedEvalKit, a unified evaluation framework that consolidates leading multimodal and textual medical benchmarks for standardized, fair, and efficient model assessment. We evaluate the performance of Lingshu on three fundamental medical tasks, multimodal QA, text-based QA, and medical report generation. The results show that Lingshu consistently outperforms the existing open-source multimodal models on most tasks ...

cs.CL↗

arXiv2026

Preserving Full 6-DOF Actuation Under Abrupt Total Rotor Failures: Passive Fault-Tolerant Flight Control Using a Biaxial-Tilt Hexacopter

Abrupt total rotor failures can invalidate nominal allocation despite remaining wrench capability. This paper develops an allocator-layer passive fault-tolerant control (AL-PFTC) scheme for overactuated vector-thrust aerial vehicles. Analysis of convexified attainable force and torque spaces at horizontal hover compares biaxial- and uniaxial-tilt configurations and screens fault cases using prescribed reserve thresholds. Instantaneous wrench loss relative to these reserves characterizes the initial redistribution burden. AL-PFTC adapts the allocation matrix using a virtual error derived from filtered wrench commands and an onboard external wrench estimate, with adaptive parameter bounds constructed from right inverses. Joint analysis including estimator residuals and filter dynamics establishes local uniform ultimate boundedness for position and attitude regulation under the stated conditions. Controller-layer passive fault-tolerant control (CL-PFTC) provides a reference with nominal allocation and observer compensation. Both share a nominal tracking baseline and run at 600 Hz on an STM32H753 without explicit fault identification, fault-mode switching, or online optimization. Comparisons show that AL-PFTC satisfies the recovery criterion in more tested hovering cases than CL-PFTC. Additional experiments using only onboard sensing, including indoor tracking under wind disturbance, outdoor tracking under wind and low-temperature conditions, narrow-frame traversal, and contact-based aerial writing, further validate the robustness of AL-PFTC in complex operational environments.

cs.RO↗

arXiv2026

AquaOrbit: Sim-to-Real Reinforcement Learning for Underwater Target Orbiting under Intermittent Visual Feedback

Intermittent visual loss disrupts target-relative feedback during underwater orbiting, making it difficult to maintain coordinated motion and reacquire a moving target. We present AquaOrbit, a reinforcement-learning controller with a recovery module for underwater target orbiting under interrupted visual feedback. During detection loss, the recovery module uses latched line-of-sight, roll, and depth references to support stabilization and target reacquisition. We train the controller in Isaac Sim with dynamics, observation, and vision-loss randomization. Evaluated without retraining in Gazebo/ROS2 under a different physics engine and perception perturbations, AquaOrbit completes 20/20 orbiting trials in each of the static- and moving-target conditions on an unseen variable-depth 3-D trajectory. In the moving-target condition, it reduces mean line-of-sight error by approximately 46% relative to a PID-based visual servoing controller with recovery while maintaining comparable path-tracking accuracy; removing the recovery module reduces completion to 9/20. Zero-shot physical deployment with fully onboard perception and control demonstrates elliptical, figure-eight, and variable-depth circular trajectories, including the latter two trajectory types absent from training. The robot maintains attitude stability during manual occlusions lasting up to 8s and reacquires the target within 2.5s in the reported attitude-induced field-of-view loss events.

cs.RO↗

arXiv2026

Sparse and weak-measurement certification of graph-edge entanglement in PXP scar wavepackets

An imperfect many-body revival does not by itself certify the entanglement of the returning state. We give a finite-record protocol for graph-edge localizable entanglement along scar wavepackets of a graph-dressed PXP chain. The target cluster state has nonzero energy variance, so neither an exact target eigenstate nor a dark-state embedding is assumed. Fresh binary probes of the undeformed Hamiltonians Pauli terms have a fully separable explanation, even within the dressed blockade sector. Adding local graph-stabilizer probes makes established entanglement witnesses accessible with simultaneous confidence bounds. For a specified square-root instrument, an imposed worst-case disturbance budget fixes the strength that minimizes the equal-allocation Hoeffding sufficient sampling cost. A local commutator bound accounts for finite-duration ancilla pulses while the Hamiltonian remains active. We test the protocol on chains through twenty spins, under perturbed dynamics, and with independent implementations. At the first twenty-spin return, synthetic weak records certify all nineteen graph edges. Generator, two-color, bounded-weight, and full-group benchmarks separate this localizable resource from genuine multipartite certification. The protocol measures recoverable entanglement in a known scar wavepacket, with explicit calibration and limits on its physical and statistical interpretation.

quant-ph↗

arXiv2026

Nondemolition filtering of an embedded cluster-state scar under continuous local monitoring

Identifying a low-entanglement eigenstate inside a many-body spectrum and preserving it during measurement are distinct tasks. We construct an explicit local ring Hamiltonian with an exact cluster-state eigenvector and study continuous monitoring of its stabilizer defects. For arbitrary mixed inputs, the conditional cluster fidelity is the initial target weight divided by the no-observed-click probability. A positive defect-operator gap gives finite-time bounds that hold for noncommuting Hamiltonian dynamics, nonnormal effective generators and imperfect detection. At fixed total monitoring rate, the guaranteed exponent falls inversely with system size; high conditional fidelity does not remove the preparation cost set by the initial overlap. Exact diagonalization up to eleven qubits gives finite-size evidence for a cluster-state outlier in a chaotic spectral background. Independent matrix and trajectory calculations verify the dynamics and a conservative coherent-error bound. This construction specializes established scar embedding and nondemolition verification frameworks, with explicit measurement assumptions, finite-time guarantees and resource limitations.

quant-ph↗

arXiv2026

Dimension-free Riesz transform estimates on biased hypercubes via non-tracial baby Fock models

We study dimension-free Riesz transform estimates on hypercubes equipped with non-uniform product measures. For $2\leq p<\infty$, we prove a dimension-free Meyer inequality for the associated carré du champ, with constants independent of both the dimension and the product measure. In contrast to the uniform case, the carré-du-champ square function does not in general coincide with the unweighted Riesz square function. Under an additional boundedness assumption on the weight parameters, we further obtain dimension-free estimates for the latter. Our proof is based on a transference from the biased hypercube to a non-tracial baby Fock model. We introduce new annihilation operators and the associated Riesz transforms on the biased hypercube, and derive explicit formulas relating them to the Riesz transforms on the non-tracial baby Fock model. The main analytic ingredient is a dimension-free $L_p$-estimate for the resulting noncommutative Riesz transform, proved by a rotation argument in the spirit of Pisier and Lust-Piquard. The same analytic mechanism also yields dimension-free Riesz transform estimates on $q$-Araki--Woods algebras.

math.FA↗

arXiv2026

OSWorld-Pro: Process-based Evaluation for Computer Use Agents

Evaluation of Computer-Use Agents (CUAs) is often limited to the final deliverables they create (at the end of hundreds of steps) and assessed with functional verifiers, as seen in OSWorld. However, such evaluation of end-state performance lacks transparency into how and why agents fail in various tasks, obfuscating critical insight for subsequent improvement. For instance, agents that err during keyboard inputs would require a different mitigation strategy from those that fail to precisely provide click-based inputs on the graphical UI. We introduce OSWorld-Pro: a set of over 300 tasks containing over 2800 subgoals to enable the procedural evaluation of CUAs grounded in over 67,000 human annotations. We use robust human-aligned LLM-Judges to evaluate the fulfillment of OSWorld-Pro subgoals and thereby reveal the progress that models make throughout a series of sequentially dependent subgoals. Our findings reveal that OSWorld-Pro is challenging even for state-of-the-art LLMs, with top performers like Claude Opus 5 achieving only 75.7% vs. 83.4% on OSWorld. Furthermore, we identify critical process-focused failure modes of various models (e.g. subgoal-irrelevant actions and click-based mistakes) to provide insights to improve performance and efficiency of CUAs.

cs.CL↗

arXiv2026

MotionJEPA: Preventing Temporal Feature Collapse by Capturing Visual Changes in Latent Space

Joint Embedding Predictive Architectures (JEPAs) are a promising paradigm for learning task-agnostic latent world models without visual reconstruction. However, standard JEPA training exhibits a strong inductive bias towards slow features, causing feature suppression and the collapse of latent representation. While inverse dynamics provides temporal anti-collapse, it relies on action labels and offers little incentive to embed general, unlabeled dynamics. We introduce Difference Image and Single image embedding Regularization (DISReg), a novel regularizer that builds on an inverse-dynamics-style module that predicts temporal difference image embeddings without any pixel reconstruction loss, encouraging balanced static and dynamic feature learning. DISReg consists of a static term that shapes the distribution of the image embedding and encourages slow features, and a dynamic term, which, unlike direct regularization on the embedding, imposes no constraint on the image embedding's shape or distribution and instead only incentivizes that dynamic features be present. By integrating this regularizer into a standard JEPA, we establish our new architecture, MotionJEPA. Latent probing demonstrates that MotionJEPA produces more complete representations than other methods, and our trajectory analysis shows it maintains geometrically simple latent embeddings with low curvature. We further show that MotionJEPA improves downstream planning success under static-background distractors across four environments.

cs.CV↗

arXiv2026

Skyrmions of Frustrated Quantum Dimer Systems

Magnetic skyrmions are topologically protected solitons observed in various classes of real magnets. In two-dimensional systems, where the target space of local magnetization values is the two-sphere $S^2$, skyrmion textures are classified by the homotopy classes of two-loops $S^2$ in $S^2$: $Π_2(S^2) \cong Z$. Here, we demonstrate that more general topological skyrmion textures emerge in the classical limit of quantum dimer systems, where the phase space of the relevant classical theory is $\mathbb{CP}^{N-1}$ (with $N=4$ for the case of interest), because the relevant second homotopy group, $Π_2(\mathbb{CP}^{N-1}) \cong Z$ for $N\geq 2$, remains unchanged. Building on the framework established by Zhang et al. (2023), we consider a classical limit based on SU(4) coherent states, which preserve intra-dimer entanglement. We show that the zero-temperature phase diagram of frustrated spin-dimer systems on a bilayer triangular lattice with weak inter-dimer coupling includes two magnetic-field-induced $\mathbb{CP}^{3}$ skyrmion crystal phases.

cond-mat.str-el↗