arXiv subjects

Zhibin Wang

Publications and source records attributed to Zhibin Wang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Stability and Decay for the 2D Anisotropic Navier-Stokes Equations with Fractional Horizontal Dissipation on $\mathbb{R}^2$

The stability problem for the 2D Navier-Stokes equations with dissipation in only one direction on $\mathbb R^2$ is not fully understood. This dissipation is in the intermediate regime between the fully dissipative Navier-Stokes and the inviscid Euler. Navier-Stokes solutions in $\mathbb R^2$ decay algebraically in time while Euler solutions can grow rather rapidly in time. This paper solves the fundamental stability and large-time behavior problem on the anisotropic Navier-Stokes with fractional dissipation $Λ_1^{2s}$ for all $0\leq s<1$. The case $s=1$ corresponds to the standard one directional dissipation $\partial_1^2$. Different techniques are developed to treat different ranges of fractional exponents: $0\leq s\leq \frac34$, $\frac34<s<\frac{11}{12}$, and $\frac{11}{12} \leq s <1$. The final range is the most difficult case, for which we introduce the spatial polynomial $A_2$ weights and exploit the boundedness of Riesz transforms on weighted $L^2$-spaces.

math.AP↗

arXiv2026

Chameleon: Adaptive Fault Tolerance for Distributed Training via Real-time Policy Selection

Training large language models faces frequent interruptions due to various faults, demanding robust fault-tolerance. Existing backup-free methods, such as redundant computation, dynamic parallelism, and data rerouting, each incur performance penalties, whether from ongoing overhead, lengthy reconfigurations, or post-recovery inefficiencies. We propose Chameleon, an adaptive fault-tolerant system that intelligently selects optimal recovery strategies when a failure occurs. Chameleon achieves this through a unified performance model, expedient execution plan search, accurate performance estimation, and efficient communication optimizations. Experiments on a 32-card cluster show that Chameleon maintains a performance gap of within 11.00% between post-recovery and failure-free training, while preserving model convergence and efficient memory usage. Compared to state-of-the-art methods, Chameleon achieves up to 1.229x and 1.355x higher average throughput than Oobleck and Recycle, respectively.

cs.DC↗

arXiv2026

AdaSpark: Adaptive Sparsity for Efficient Long-Video Understanding

Processing long-form videos with Video Large Language Models (Video-LLMs) is computationally prohibitive. Current efficiency methods often compromise fine-grained perception through irreversible information disposal or inhibit long-range temporal modeling via rigid, predefined sparse patterns. This paper introduces AdaSpark, an adaptive sparsity framework designed to address these limitations. AdaSpark first partitions video inputs into 3D spatio-temporal cubes. It then employs two co-designed, context-aware components: (1) Adaptive Cube-Selective Attention (AdaS-Attn), which adaptively selects a subset of relevant video cubes to attend for each query token, and (2) Adaptive Token-Selective FFN (AdaS-FFN), which selectively processes only the most salient tokens within each cube. An entropy-based (Top-p) selection mechanism adaptively allocates computational resources based on input complexity. Experiments demonstrate that AdaSpark significantly reduces computational load by up to 57% FLOPs while maintaining comparable performance to dense models and preserving fine-grained, long-range dependencies, as validated on challenging hour-scale video benchmarks.

cs.CV↗

arXiv2026

CoDec: Prefix-Shared Decoding Kernel for LLMs

Prefix-sharing among multiple prompts presents opportunities to combine the operations of the shared prefix, while attention computation in the decode stage, which becomes a critical bottleneck with increasing context lengths, is a memory-intensive process requiring heavy memory access on the key-value (KV) cache of the prefixes. Therefore, in this paper, we explore the potential of prefix-sharing in the attention computation of the decode stage. However, the tree structure of the prefix-sharing mechanism presents significant challenges for attention computation in efficiently processing shared KV cache access patterns while managing complex dependencies and balancing irregular workloads. To address the above challenges, we propose a dedicated attention kernel to combine the memory access of shared prefixes in the decoding stage, namely CoDec. CoDec delivers two key innovations: a novel shared-prefix attention kernel that optimizes memory hierarchy and exploits both intra-block and inter-block parallelism, and a comprehensive workload balancing mechanism that efficiently estimates cost, divides tasks, and schedules execution. Experimental results show that CoDec achieves an average $1.9\times$ speedup and $120.9\times$ memory access reduction compared to the state-of-the-art FlashDecoding kernel regarding attention computation in the decode stage and $3.8\times$ end-to-end time per output token compared to the vLLM.

cs.LG↗

arXiv2026

SpatialReward: Verifiable Spatial Reward Modeling for Fine-Grained Spatial Consistency in Text-to-Image Generation

Recent advances in text-to-image (T2I) generation via reinforcement learning (RL) have benefited from reward models that assess semantic alignment and visual quality. However, most existing reward models pay limited attention to fine-grained spatial relationships, often producing images that appear plausible overall yet contain inaccuracies in object positioning. In this work, we present \textbf{SpatialReward}, a verifiable reward model explicitly designed to evaluate spatial layouts in generated images. SpatialReward adopts a multi-stage pipeline: a \emph{Prompt Decomposer} extracts entities, attributes, and spatial metadata from free-form prompts; expert detectors provide accurate visual grounding of object positions and attributes; and a vision-language model applies chain-of-thought reasoning over grounded observations to assess complex spatial relations that are challenging for rule-based methods. To more comprehensively evaluate spatial relationships in generated images, we introduce \textbf{SpatRelBench}, a benchmark covering object attributes, orientation, inter-object relations, and rendered text placement. Experiments on Stable Diffusion and FLUX show that incorporating SpatialReward into RL training consistently improves spatial consistency and overall generation quality, with results aligned more closely to human judgments. These findings indicate that verifiable reward models hold considerable potential for enabling more accurate and controllable optimization in text-to-image generation models.

cs.CV↗

arXiv2026

Contribution-aware Token Compression for Efficient Video Understanding via Reinforcement Learning

Video large language models have demonstrated remarkable capabilities in video understanding tasks. However, the redundancy of video tokens introduces significant computational overhead during inference, limiting their practical deployment. Many compression algorithms are proposed to prioritize retaining features with the highest attention scores to minimize perturbations in attention computations. However, the correlation between attention scores and their actual contribution to correct answers remains ambiguous. To address the above limitation, we propose a novel \textbf{C}ontribution-\textbf{a}ware token \textbf{Co}mpression algorithm for \textbf{VID}eo understanding (\textbf{CaCoVID}) that explicitly optimizes the token selection policy based on the contribution of tokens to correct predictions. First, we introduce a reinforcement learning-based framework that optimizes a policy network to select video token combinations with the greatest contribution to correct predictions. This paradigm shifts the focus from passive token preservation to active discovery of optimal compressed token combinations. Secondly, we propose a combinatorial policy optimization algorithm with online combination space sampling, which dramatically reduces the exploration space for video token combinations and accelerates the convergence speed of policy optimization. Extensive experiments on diverse video understanding benchmarks demonstrate the effectiveness of CaCoVID. Codes are available at https://github.com/LivingFutureLab/CaCoVID.

cs.CV↗

arXiv2026

A Corrected Open Boundary Framework for Lattice Boltzmann Immiscible Pseudopotential Models

The pseudopotential lattice Boltzmann method (LBM) is a prominent approach for simulating multiphase flows, valued for its physical intuitiveness and computational tractability. However, existing immiscible pseudopotential methods for modeling dynamic multi-component immiscible fluid systems involving open boundaries face persistent challenges, notably the influence of spurious currents on interface formation and breakup, as well as the effects of inlet and outlet boundary configurations on simulation stability. Therefore, this paper proposes a corrected open boundary framework based on Multiple-relaxation-time (MRT) for the immiscible pseudopotential model. Our method includes three key improvements: (1) For the accurate recovery of macroscopic quantities at the inlet boundary, correction coefficients are introduced to reconstruct the distribution function; (2) Based on real-time mass flow rates at the inlet and outlet, the outlet boundary velocity is adjusted to ensure global mass conservation in the computational domain; (3) The relaxation coefficient related to numerical stability is adjusted based on the viscosity of two-phase fluids to reduce spurious currents. To validate the reliability of the proposed corrected method, four benchmark cases were simulated: Laplace tests and Taylor deformation, two-phase Poiseuille flow, migration of droplets in microchannels, as well as droplet generation in T-shaped and co-flow devices. The results demonstrate that the corrected method reduced the average spurious currents at the phase interface by 65.8%, and controlled the average mass deviation of the fluid system at around 3.5%. In addition, the morphology of the droplets differs by less than 5% compared to the benchmark examples and experiments.

physics.flu-dyn↗

arXiv2026

DART: Diffusion-Inspired Speculative Decoding for Fast LLM Inference

Speculative decoding is an effective and lossless approach for accelerating LLM inference. However, existing widely adopted model-based draft designs, such as EAGLE3, improve accuracy at the cost of multi-step autoregressive inference, resulting in high drafting latency and ultimately rendering the drafting stage itself a performance bottleneck. Inspired by diffusion-based large language models (dLLMs), we propose DART, which leverages parallel generation to reduce drafting latency. DART predicts logits for multiple future masked positions in parallel within a single forward pass based on hidden states of the target model, thereby eliminating autoregressive rollouts in the draft model while preserving a lightweight design. Based on these parallel logit predictions, we further introduce an efficient tree pruning algorithm that constructs high-quality draft token trees with N-gram-enforced semantic continuity. DART substantially reduces draft-stage overhead while preserving high draft accuracy, leading to significantly improved end-to-end decoding speed. Experimental results demonstrate that DART achieves a 2.03x--3.44x wall-clock time speedup across multiple datasets, surpassing EAGLE3 by 30% on average and offering a practical speculative decoding framework. Code is released at https://github.com/fvliang/DART.

cs.CL↗

arXiv2025

A region-specific brain dysfunction underlies cognitive impairment in long COVID brain fog

Long COVID "brain fog" is a common and debilitating subjective syndrome often associated with persistent cognitive impairment after COVID-19 infection. Here we identify a specific regional brain dysfunction that mediates this cognitive impairment and provide evidence that targeted neuromodulation improves this deficit. In 120 patients with long COVID brain fog, we found an aberrant perceptual processing pattern. Patients with more severe brain fog committed significantly more false alarms (impulsive responses to non-signals) despite preserved overall accuracy. Both high-density (128-channel) EEG and structural MRI analyses provided converging evidence of a right inferior insula deficit, characterized by a blunted neural monitoring signal and cortical atrophy. We confirmed this deficit in a separate 796-participant UK Biobank longitudinal COVID re-imaging cohort, where COVID-19 survivors also showed selective impairment on a perceptual processing task and corresponding longitudinal atrophy of the right inferior insula compared with healthy controls. Finally, in a proof-of-principle randomized, sham-controlled trial (n = 40), a non-invasive, excitatory theta-burst ultrasound stimulation protocol targeting the right inferior insula rescued the perceptual deficit by reducing false alarms. These findings provide evidence of a causal role for right inferior insula dysfunction in long COVID-related perceptual impairment and show that modulation of this region can rescue the deficit, establishing it as a novel therapeutic target for long COVID cognitive impairment.

q-bio.NC↗

arXiv2025

StreamKV: Streaming Video Question-Answering with Segment-based KV Cache Retrieval and Compression

Video Large Language Models (Video-LLMs) have demonstrated significant potential in the areas of video captioning, search, and summarization. However, current Video-LLMs still face challenges with long real-world videos. Recent methods have introduced a retrieval mechanism that retrieves query-relevant KV caches for question answering, enhancing the efficiency and accuracy of long real-world videos. However, the compression and retrieval of KV caches are still not fully explored. In this paper, we propose \textbf{StreamKV}, a training-free framework that seamlessly equips Video-LLMs with advanced KV cache retrieval and compression. Compared to previous methods that used uniform partitioning, StreamKV dynamically partitions video streams into semantic segments, which better preserves semantic information. For KV cache retrieval, StreamKV calculates a summary vector for each segment to retain segment-level information essential for retrieval. For KV cache compression, StreamKV introduces a guidance prompt designed to capture the key semantic elements within each segment, ensuring only the most informative KV caches are retained for answering questions. Moreover, StreamKV unifies KV cache retrieval and compression within a single module, performing both in a layer-adaptive manner, thereby further improving the effectiveness of streaming video question answering. Extensive experiments on public StreamingVQA benchmarks demonstrate that StreamKV significantly outperforms existing Online Video-LLMs, achieving superior accuracy while substantially improving both memory efficiency and computational latency. The code has been released at https://github.com/sou1p0wer/StreamKV.

cs.CV↗

arXiv2025

GMoPE:A Prompt-Expert Mixture Framework for Graph Foundation Models

Graph Neural Networks (GNNs) have demonstrated impressive performance on task-specific benchmarks, yet their ability to generalize across diverse domains and tasks remains limited. Existing approaches often struggle with negative transfer, scalability issues, and high adaptation costs. To address these challenges, we propose GMoPE (Graph Mixture of Prompt-Experts), a novel framework that seamlessly integrates the Mixture-of-Experts (MoE) architecture with prompt-based learning for graphs. GMoPE leverages expert-specific prompt vectors and structure-aware MoE routing to enable each expert to specialize in distinct subdomains and dynamically contribute to predictions. To promote diversity and prevent expert collapse, we introduce a soft orthogonality constraint across prompt vectors, encouraging expert specialization and facilitating a more balanced expert utilization. Additionally, we adopt a prompt-only fine-tuning strategy that significantly reduces spatiotemporal complexity during transfer. We validate GMoPE through extensive experiments under various pretraining strategies and multiple downstream tasks. Results show that GMoPE consistently outperforms state-of-the-art baselines and achieves performance comparable to full parameter fine-tuning-while requiring only a fraction of the adaptation overhead. Our work provides a principled and scalable framework for advancing generalizable and efficient graph foundation models.

cs.LG↗

arXiv2025

Accelerating Mixture-of-Experts Inference by Hiding Offloading Latency with Speculative Decoding

Recent advancements in Mixture of Experts (MoE) models have significantly increased their parameter scale as well as model performance. Extensive offloading techniques have been proposed to address the GPU memory limitations of MoE inference. However, due to the I/O bottleneck and sparse computation of MoE models, existing offloading techniques still suffer from low hardware utilization. To fully utilize the hardware resources, we propose SpecMoEOff, which employs the speculative decoding technique to enlarge the workload of each expert. SpecMoEOff orchestrates the GPU and CPU by both theoretical and empirical roofline analysis. In addition, we develop a dedicated CPU chunked attention verification kernel to fit the speculative decoding in offloading scenarios as well as minimizing the additional overhead led by draft models. SpecMoEOff further integrates an optimizer to automatically tune the hyperparameters of speculative decoding under given hardware and workload. Experimental results show that SpecMoEOff achieves up to 2.5x decode throughput improvement over the state-of-the-art MoE offloading techniques.

cs.DC↗

arXiv2025

Revisiting Service Level Objectives and System Level Metrics in Large Language Model Serving

User experience is a critical factor Large Language Model (LLM) serving systems must consider, where service level objectives (SLOs) considering the experience of individual requests and system level metrics (SLMs) considering the overall system performance are two key performance measures. However, we observe two notable issues in existing metrics: 1) manually delaying the delivery of some tokens can improve SLOs, and 2) actively abandoning requests that do not meet SLOs can improve SLMs, both of which are counterintuitive. In this paper, we revisit SLOs and SLMs in LLM serving, and propose a new SLO that aligns with user experience. Based on the SLO, we propose a comprehensive metric framework called smooth goodput, which integrates SLOs and SLMs to reflect the nature of user experience in LLM serving. Through this unified framework, we reassess the performance of different LLM serving systems under multiple workloads. Evaluation results show that our metric framework provides a more comprehensive view of token delivery and request processing, and effectively captures the optimal point of user experience and system performance with different serving strategies.

cs.LG↗

arXiv2025

BLM$_1$: A Boundless Large Model for Cross-Space, Cross-Task, and Cross-Embodiment Learning

Multimodal large language models (MLLMs) have advanced vision-language reasoning and are increasingly deployed in embodied agents. However, significant limitations remain: MLLMs generalize poorly across digital-physical spaces and embodiments; vision-language-action models (VLAs) produce low-level actions yet lack robust high-level embodied reasoning; and most embodied large language models (ELLMs) are constrained to digital-space with poor generalization to the physical world. Thus, unified models that operate seamlessly across digital and physical spaces while generalizing across embodiments and tasks remain absent. We introduce the \textbf{Boundless Large Model (BLM$_1$)}, a multimodal spatial foundation model that preserves instruction following and reasoning, incorporates embodied knowledge, and supports robust cross-embodiment control. BLM$_1$ integrates three key capabilities -- \textit{cross-space transfer, cross-task learning, and cross-embodiment generalization} -- via a two-stage training paradigm. Stage I injects embodied knowledge into the MLLM through curated digital corpora while maintaining language competence. Stage II trains a policy module through an intent-bridging interface that extracts high-level semantics from the MLLM to guide control, without fine-tuning the MLLM backbone. This process is supported by a self-collected cross-embodiment demonstration suite spanning four robot embodiments and six progressively challenging tasks. Evaluations across digital and physical benchmarks show that a single BLM$_1$ instance outperforms four model families -- MLLMs, ELLMs, VLAs, and GMLMs -- achieving $\sim\!\textbf{6%}$ gains in digital tasks and $\sim\!\textbf{3%}$ in physical tasks.

cs.AI↗

arXiv2025

ReWatch-R1: Boosting Complex Video Reasoning in Large Vision-Language Models through Agentic Data Synthesis

While Reinforcement Learning with Verifiable Reward (RLVR) significantly advances image reasoning in Large Vision-Language Models (LVLMs), its application to complex video reasoning remains underdeveloped. This gap stems primarily from a critical data bottleneck: existing datasets lack the challenging, multi-hop questions and high-quality, video-grounded Chain-of-Thought (CoT) data necessary to effectively bootstrap RLVR. To address this, we introduce ReWatch, a large-scale dataset built to foster advanced video reasoning. We propose a novel multi-stage synthesis pipeline to synthesize its three components: ReWatch-Caption, ReWatch-QA, and ReWatch-CoT. A core innovation is our Multi-Agent ReAct framework for CoT synthesis, which simulates a human-like "re-watching" process to generate video-grounded reasoning traces by explicitly modeling information retrieval and verification. Building on this dataset, we develop ReWatch-R1 by post-training a strong baseline LVLM with Supervised Fine-Tuning (SFT) and our RLVR framework. This framework incorporates a novel Observation \& Reasoning (O\&R) reward mechanism that evaluates both the final answer's correctness and the reasoning's alignment with video content, directly penalizing hallucination. Our experiments show that ReWatch-R1 achieves state-of-the-art average performance on five challenging video reasoning benchmarks. Project Page: https://rewatch-r1.github.io

cs.CV↗

arXiv2025

Improvement of Solid-Fluid Interaction Scheme in Lattice Boltzmann Immiscible Pseudopotential Models

The pseudopotential model within the Lattice Boltzmann Method (LBM) framework has emerged as a prominent approach in computational fluid dynamics due to its dual strengths in physical intuitiveness and computational tractability. However, when modeling wettability phenomenon, existing solid-fluid interaction schemes exhibit persistent challenges in multi-component immiscible fluid systems, notably manifested through spurious velocity artifacts and unphysical mass-transfer boundary layers. This study presents an improved interaction scheme that preserves implementation simplicity while effectively mitigating these numerical artifacts. Furthermore, leveraging the enhanced isotropy characteristics of eighth-order discrete schemes, we develop a novel boundary treatment methodology addressing second-layer lattice data reconstruction at complex interfaces. To verify the universality of the proposed optimization scheme, four benchmark scenarios, including static contact angle measurement on cylindrical surface, droplet dynamics through confined geometries, immiscible displacement processes, and co-current flow in microchannels, are simulated to demonstrate the proposed scheme's capability. The results show that the improved scheme can well simulate various complex immiscible multiphase flows.

physics.flu-dyn↗

arXiv2025

Three-dimensional Helical-rotating Plasma Structures in Beam-generated Partially Magnetized Plasmas

Azimuthal structures emerging in beam-generated partially magnetized plasmas are investigated using three-dimensional particle-in-cell/Monte Carlo collision simulations. Two distinct instability regimes are identified at low pressures. When the gas pressure is sufficiently high, quasi-neutrality is attained and 2D spiral-arm structures form as a result of the development of a lower-hybrid instability, resulting in enhanced cross-field transport. At lower pressures, quasi-neutrality is not achieved and a 3D helical-rotating plasma structure forms due to development of the diocotron instability. Analytical formulas are proposed for the critical threshold pressure between these regimes and for the rotation frequency of the helical structures. Preliminary experimental verification is provided.

physics.plasm-ph↗

arXiv2025

Chordless Structure: A Pathway to Simple and Expressive GNNs

Researchers have proposed various methods of incorporating more structured information into the design of Graph Neural Networks (GNNs) to enhance their expressiveness. However, these methods are either computationally expensive or lacking in provable expressiveness. In this paper, we observe that the chords increase the complexity of the graph structure while contributing little useful information in many cases. In contrast, chordless structures are more efficient and effective for representing the graph. Therefore, when leveraging the information of cycles, we choose to omit the chords. Accordingly, we propose a Chordless Structure-based Graph Neural Network (CSGNN) and prove that its expressiveness is strictly more powerful than the k-hop GNN (KPGNN) with polynomial complexity. Experimental results on real-world datasets demonstrate that CSGNN outperforms existing GNNs across various graph tasks while incurring lower computational costs and achieving better performance than the GNNs of 3-WL expressiveness.

cs.LG↗