arXiv subjects

Yanwei Li

Publications and source records attributed to Yanwei Li.

At least 19 recordsLinked to original sources

arXiv2026

ZimaBlue: Evolving Generalizable World Action Models through Scalable Video Pre-training

Robotic manipulation faces a fundamental scaling challenge: robust generalization demands broad physical experience, yet action-labeled robot trajectories are expensive to collect and inherently limited in diversity. Egocentric videos offer a far more scalable source of embodied experience, capturing object interactions, contact dynamics, tool use, and long-horizon behaviors across diverse environments. The central challenge is how to convert this abundant but action-free experience into effective robot control. We introduce ZimaBlue, a scalable framework for learning generalizable World Action Models (WAMs) from large-scale video. ZimaBlue follows a three-stage training curriculum: it first performs causal embodied video pre-training on large-scale human and robot egocentric videos, then grounds the learned visual dynamics in heterogeneous robot trajectories through video-action mid-training with a unified action representation, and finally specializes the model to a target robot for deployment. To make generative WAMs practical for real-time control, ZimaBluefurther adopts an asynchronous Slow-Fast dual-system architecture, where a high-capacity Slow world model provides generalizable spatiotemporal representations and a lightweight Fast branch enables 30 Hz action prediction on NVIDIA RTX 4090. On real-robot zero-shot evaluations, scaling from target-robot data alone to over 120,000 hours of embodied video improves success from 36.1% to 77.8%. ZimaBlue further delivers strong performance across multiple benchmarks, with particularly pronounced gains on unseen tasks.

cs.CV↗

arXiv2026

MotionAtlas: Detailed Region Captioning for Motion-Centric Videos

We propose MotionAtlas, a system for detailed captioning of motion-centric videos, comprising (1) a dedicated human-annotated benchmark, (2) a scalable, high-quality pipeline to construct training samples, and (3) a family of powerful Video-MLLMs. Unlike conventional global motion captioning datasets, we focus on region-aware motion captioning: given a video and a spatiotemporal mask, the model generates precise descriptions of motion within the target region, thereby alleviating visual clutter and motion entanglement and enabling reliable, quantifiable evaluation. Concretely, we first build MotionAtlas-Bench, a comprehensive benchmark comprising 2,073 multiple-choice questions, meticulously annotated for a curated set of high-quality, motion-centric videos, to evaluate fine-grained motion understanding of the objects in question. Second, we design a rigorous and scalable data pipeline that leverages self-bootstrap refinement to suppress fine-grained hallucinations, yielding 159k high-quality motion captioning data. Third, we design a tailored training data composition strategy, which achieves consistent and substantial performance gains across diverse baseline Video-MLLMs, including Molmo2 and Qwen3-VL. For instance, MotionAtlas-4B surpasses Qwen3-VL-4B by an average of 5.2 percentage points across general motion benchmarks. The benchmark, dataset, and code have been released.

cs.CV↗

arXiv2026

Watch, Remember, Reason: Human-View Video Understanding with MLLMs

Video understanding is being rapidly transformed by multimodal large language models (MLLMs), as research moves from short clips to long, multimodal, and knowledge-intensive video scenarios. These scenarios require models to handle sparse evidence, long-range dependencies, multimodal alignment, and reliable inference under limited computational budgets. This work presents a human-view perspective on LLM-based video understanding, organized around three functional abilities: watching, remembering, and reasoning. Rather than treating video tasks as isolated benchmarks, this view provides a unified structure for analyzing how video MLLMs acquire evidence, preserve context, and produce grounded outputs. We introduce a formulation that characterizes video understanding systems by their perceptual representations, memory states, reasoning traces, and final predictions. Based on this formulation, we identify challenges in spatio-temporal perception, efficient long-video processing, memory modeling, streaming understanding, and faithful reasoning. Representative methods are organized by their roles in video MLLM systems. Watching covers fine-grained, comprehensive, audio-visual, and efficient perception. Remembering includes offline and streaming memory, while reasoning covers text-only reasoning and thinking with videos. We further examine application domains such as egocentric, sports, instructional, medical, and narrative videos, and cover training datasets and evaluation benchmarks across task types, supervision formats, modalities, and capability dimensions. Finally, we outline open problems and future directions for scalable, memory-aware, and evidence-grounded video intelligence. Related works will be continuously traced at https://github.com/marinero4972/Awesome-HumanView-VideoUnderstanding.

cs.CV↗

arXiv2026

Benchmarking and Evolving Reason-Reflect-Rectify for Reflective Visual Generation

Text-to-Image (T2I) models and Unified Multimodal Models (UMMs) have achieved remarkable progress in visual generation. However, their reliance on a single-pass generation paradigm limits their ability to handle complex prompts requiring iterative refinement. To enable multi-round Reflective Visual Generation (RVG), we formalize the Reason-Reflect-Rectify (R^3) loop as a core framework and introduce R^3-Bench, a benchmark of over 600 expert-annotated instances that quantifies iterative reasoning and rectification capabilities. Evaluation on R^3-Bench reveals a critical gap: while state-of-the-art models can identify generation errors, they fail to generate actionable rectification instructions. To bridge this gap, we propose R^3-Refiner, a dual-stage framework leveraging Group Relative Policy Optimization (GRPO) and a Hierarchical Reward Mechanism (HRM) to better align rectification with reflective reasoning. Experiments show that R^3-Refiner achieves significant improvements on R^3-Bench (+12.0% in Reflective Verdict Score, +9.0% in Rectification Score), and can be seamlessly integrated with various MLLMs to enhance the generation quality of different T2I models on GenEval++ and T2I-CompBench. Code is available at https://github.com/xiaomoguhz/R3-Bench.

cs.CV↗

arXiv2026

Semantic Generative Tuning for Unified Multimodal Models

Unified multimodal models (UMMs) strive to consolidate visual understanding and visual generation within a single architecture. However, prevailing training paradigms independently optimize understanding via sparse text signals and generation through dense pixel objectives. Such a decoupled strategy yields misaligned representation spaces, isolating visual understanding from generation and hindering their mutual reinforcement. This work presents the first systematic investigation into generative post-training, where we formulate hierarchical visual tasks as generative proxies to bridge the isolation in UMMs. Our empirical investigation reveals that high-level semantic tasks, particularly image segmentation, serve as optimal proxies. Unlike low-level tasks that distract models with texture details, segmentation provides structural semantics that significantly enhance both vision-centric perception and generative layout fidelity. Building upon these insights, we introduce Semantic Generative Tuning (SGT), a novel paradigm that leverages segmentation as a generative proxy to align and synergize multimodal capabilities. Mechanistic analyses further demonstrate that SGT fundamentally improves feature linear separability and optimizes visual-textual attention allocation pattern. Extensive evaluations show that SGT consistently improves both multimodal comprehension and generative fidelity across mainstream benchmarks. Our code is available on the https://song2yu.github.io/SGT/.

cs.CV↗

arXiv2026

Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond

As AI systems move from generating text to accomplishing goals through sustained interaction, the ability to model environment dynamics becomes a central bottleneck. Agents that manipulate objects, navigate software, coordinate with others, or design experiments require predictive environment models, yet the term world model carries different meanings across research communities. We introduce a "levels x laws" taxonomy organized along two axes. The first defines three capability levels: L1 Predictor, which learns one-step local transition operators; L2 Simulator, which composes them into multi-step, action-conditioned rollouts that respect domain laws; and L3 Evolver, which autonomously revises its own model when predictions fail against new evidence. The second identifies four governing-law regimes: physical, digital, social, and scientific. These regimes determine what constraints a world model must satisfy and where it is most likely to fail. Using this framework, we synthesize over 400 works and summarize more than 100 representative systems spanning model-based reinforcement learning, video generation, web and GUI agents, multi-agent social simulation, and AI-driven scientific discovery. We analyze methods, failure modes, and evaluation practices across level-regime pairs, propose decision-centric evaluation principles and a minimal reproducible evaluation package, and outline architectural guidance, open problems, and governance challenges. The resulting roadmap connects previously isolated communities and charts a path from passive next-step prediction toward world models that can simulate, and ultimately reshape, the environments in which agents operate. Code and resources are available at: https://github.com/matrix-agent/awesome-agentic-world-modeling.

cs.AI↗

arXiv2026

High-rate quantum digital signatures over 250 km of optical fiber

Quantum digital signatures (QDS) offer information-theoretic security for message integrity, authenticity, and non-repudiation, and constitute a fundamental cryptographic primitive for future quantum networks. Despite significant progress, the practical deployment of QDS has been severely constrained by limited signature rates and poor tolerance to channel loss, particularly in long-distance and metropolitan-scale networks. Here, we report a high-rate, loss-resilient QDS system that overcomes these two key bottlenecks simultaneously. Our implementation combines intrinsically phase-stable polarization modulation based on a Sagnac interferometer with gigahertz-rate quantum state encoding and low-timing-jitter superconducting nanowire single-photon detectors, enabling robust and continuous operation at high repetition frequencies. By integrating this hardware platform with a one-time universal hashing-based QDS protocol, we achieve a signature rate improvement of more than two orders of magnitude compared with existing QDS implementations under comparable channel-loss conditions. Notably, the system maintains a non-zero effective signature rate of approximately 1.25 times per second at a total channel loss of up to 49.05 dB, representing the highest loss tolerance reported for QDS to date. These results establish a practical and scalable technological pathway for deploying QDS in real-world quantum communication networks.

quant-ph↗

arXiv2026

One Loss to Rule Them All: Marked Time-to-Event for Structured EHR Foundation Models

Clinical events captured in Electronic Health Records (EHR) are irregularly sampled and may consist of a mixture of discrete events and numerical measurements, such as laboratory values or treatment dosages. The sequential nature of EHR, analogous to natural language, has motivated the use of next-token prediction to train prior EHR Foundation Models (FMs) over events. However, this training fails to capture the full structure of EHR. When a given event occurs must be captured, but the event value (abnormal lab) also modulates the likelihood of other clinical events. Most existing EHR FMs do not jointly model this likelihood and are unable to capture the full observation process, impacting downstream capabilities. We propose ORA, a marked time-to-event pretraining objective that jointly models event timing and associated measurements. Across multiple datasets, downstream tasks, and model backbones, this objective consistently yields more generalizable representations than next-token prediction and pretraining losses that ignore continuous measurements. Importantly, the proposed objective yields improvements beyond traditional classification evaluation, including better regression and time-to-event prediction. Beyond introducing a new family of FMs, our ablations suggest a broader takeaway: pretraining objectives that account for EHR structure are critical for expanding downstream capabilities and generalizability.

cs.LG↗

arXiv2025

Visual Reasoning Tracer: Object-Level Grounded Reasoning Benchmark

Recent advances in Multimodal Large Language Models (MLLMs) have significantly improved performance on tasks such as visual grounding and visual question answering. However, the reasoning processes of these models remain largely opaque; they typically output only final predictions without revealing the intermediate steps or fine-grained evidence (e.g., pixels, locations) that lead to the result. This contrasts with human intelligence, which naturally operates through a chain of visual reasoning. To address this limitation, we introduce the Visual Reasoning Tracer (VRT) task, which requires models to not only localize the target object but also explicitly predict the intermediate objects that form the reasoning path. To advance research in this area, we contribute: (1) VRT-Bench, a human-annotated benchmark for evaluating visual reasoning; (2) a new metric for assessing the quality of reasoning traces; and (3) VRT-80k, a large-scale dataset for reasoning model training. Our experiments reveal that while existing models often produce the correct final output, they struggle to ground their intermediate reasoning. In contrast, models trained on VRT-80k achieve substantial improvements in tracing the reasoning path.

cs.CV↗

arXiv2025

Visual Spatial Tuning

Capturing spatial relationships from visual inputs is a cornerstone of human-like general intelligence. Several previous studies have tried to enhance the spatial awareness of Vision-Language Models (VLMs) by adding extra expert encoders, which brings extra overhead and usually harms general capabilities. To enhance the spatial ability in general architectures, we introduce Visual Spatial Tuning (VST), a comprehensive framework to cultivate VLMs with human-like visuospatial abilities, from spatial perception to reasoning. We first attempt to enhance spatial perception in VLMs by constructing a large-scale dataset termed VST-P, which comprises 4.1 million samples spanning 19 skills across single views, multiple images, and videos. Then, we present VST-R, a curated dataset with 135K samples that instruct models to reason in space. In particular, we adopt a progressive training pipeline: supervised fine-tuning to build foundational spatial knowledge, followed by reinforcement learning to further improve spatial reasoning abilities. Without the side-effect to general capabilities, the proposed VST consistently achieves state-of-the-art results on several spatial benchmarks, including $34.8\%$ on MMSI-Bench and $61.2\%$ on VSIBench. It turns out that the Vision-Language-Action models can be significantly enhanced with the proposed spatial tuning paradigm, paving the way for more physically grounded AI.

cs.CV↗

arXiv2025

Grasp Any Region: Towards Precise, Contextual Pixel Understanding for Multimodal LLMs

While Multimodal Large Language Models (MLLMs) excel at holistic understanding, they struggle in capturing the dense world with complex scenes, requiring fine-grained analysis of intricate details and object inter-relationships. Region-level MLLMs have been a promising step. However, previous attempts are generally optimized to understand given regions in isolation, neglecting crucial global contexts. To address this, we introduce Grasp Any Region (GAR) for comprehen- sive region-level visual understanding. Empowered by an effective RoI-aligned feature replay technique, GAR supports (1) precise perception by leveraging necessary global contexts, and (2) modeling interactions between multiple prompts. Together, it then naturally achieves (3) advanced compositional reasoning to answer specific free-form questions about any region, shifting the paradigm from passive description to active dialogue. Moreover, we construct GAR-Bench, which not only provides a more accurate evaluation of single-region comprehension, but also, more importantly, measures interactions and complex reasoning across multiple regions. Extensive experiments have demonstrated that GAR-1B not only maintains the state-of-the-art captioning capabilities, e.g., outperforming DAM-3B +4.5 on DLC-Bench, but also excels at modeling relationships between multiple prompts with advanced comprehension capabilities, even surpassing InternVL3-78B on GAR-Bench-VQA. More importantly, our zero-shot GAR-8B even outperforms in-domain VideoRefer-7B on VideoRefer-BenchQ, indicating its strong capabilities can be easily transferred to videos.

cs.CV↗

arXiv2025

Field-Trial Quantum Key Distribution with Qubit-Based Frame Synchronization

Quantum key distribution (QKD) is a cryptographic technique that uses quantum mechanical principles to enable secure key exchange. Practical deployment of QKD requires robust, cost-effective systems that can operate in challenging field environments. A major challenge is achieving reliable clock synchronization without adding hardware complexity. Conventional approaches often use separate classical light signals, which increase costs and introduce noise that degrades quantum channel performance. To address this limitation, we demonstrate a QKD system incorporating a recently proposed qubit-based distributed frame synchronization method, deployed over a metropolitan fiber network in Nanning, China. Using the polarization-encoded one-decoy-state BB84 protocol and the recently proposed qubit-based distributed frame synchronization method, our system achieves synchronization directly from the quantum signal, eliminating the need for dedicated synchronization hardware. Furthermore, to counteract dynamic polarization disturbances in urban fibers, the system integrates qubit-based polarization feedback control, enabling real-time polarization compensation through an automated polarization controller using data recovered from the qubit-based synchronization signals. During 12 hours of continuous operation, the system maintained a low average quantum bit error rate (QBER) of 1.12/%, achieving a secure key rate of 26.6 kbit/s under 18 dB channel loss. Even under a high channel loss of 40 dB, a finite-key secure rate of 115 bit/s was achieved. This study represents the first successful long-term validation of a frame-synchronization based QKD scheme in a real urban environment, demonstrating exceptional stability and high-loss tolerance, and offering an alternative for building practical, scalable, and cost-efficient quantum-secure communication networks.

quant-ph↗

arXiv2025

How Far are VLMs from Visual Spatial Intelligence? A Benchmark-Driven Perspective

Visual Spatial Reasoning (VSR) is a core human cognitive ability and a critical requirement for advancing embodied intelligence and autonomous systems. Despite recent progress in Vision-Language Models (VLMs), achieving human-level VSR remains highly challenging due to the complexity of representing and reasoning over three-dimensional space. In this paper, we present a systematic investigation of VSR in VLMs, encompassing a review of existing methodologies across input modalities, model architectures, training strategies, and reasoning mechanisms. Furthermore, we categorize spatial intelligence into three levels of capability, ie, basic perception, spatial understanding, spatial planning, and curate SIBench, a spatial intelligence benchmark encompassing nearly 20 open-source datasets across 23 task settings. Experiments with state-of-the-art VLMs reveal a pronounced gap between perception and reasoning, as models show competence in basic perceptual tasks but consistently underperform in understanding and planning tasks, particularly in numerical estimation, multi-view reasoning, temporal dynamics, and spatial imagination. These findings underscore the substantial challenges that remain in achieving spatial intelligence, while providing both a systematic roadmap and a comprehensive benchmark to drive future research in the field. The related resources of this study are accessible at https://sibench.github.io/Awesome-Visual-Spatial-Reasoning/.

cs.AI↗

arXiv2025

Aligning Effective Tokens with Video Anomaly in Large Language Models

Understanding abnormal events in videos is a vital and challenging task that has garnered significant attention in a wide range of applications. Although current video understanding Multi-modal Large Language Models (MLLMs) are capable of analyzing general videos, they often struggle to handle anomalies due to the spatial and temporal sparsity of abnormal events, where the redundant information always leads to suboptimal outcomes. To address these challenges, exploiting the representation and generalization capabilities of Vison Language Models (VLMs) and Large Language Models (LLMs), we propose VA-GPT, a novel MLLM designed for summarizing and localizing abnormal events in various videos. Our approach efficiently aligns effective tokens between visual encoders and LLMs through two key proposed modules: Spatial Effective Token Selection (SETS) and Temporal Effective Token Generation (TETG). These modules enable our model to effectively capture and analyze both spatial and temporal information associated with abnormal events, resulting in more accurate responses and interactions. Furthermore, we construct an instruction-following dataset specifically for fine-tuning video-anomaly-aware MLLMs, and introduce a cross-domain evaluation benchmark based on XD-Violence dataset. Our proposed method outperforms existing state-of-the-art methods on various benchmarks.

cs.CV↗

arXiv2025

DenseWorld-1M: Towards Detailed Dense Grounded Caption in the Real World

Multimodal Large Language Models (MLLMs) demonstrate a complex understanding of scenes, benefiting from large-scale and high-quality datasets. Most existing caption datasets lack the ground locations and relations for visual entities. Several grounded caption datasets face the problems of missing detailed descriptions, relations, and massive object descriptions on high-resolution images. To fill this gap for the community, we present DenseWorld-1M, the first massive, detailed, dense grounded caption dataset in the real world. We design a three-stage labeling pipeline, containing open-world perception, detailed object caption generation, and dense caption merging. The first stage obtains entity-level masks and labels. The second stage generates the object-level, detailed captions with the guidance of masks and labels from the first stage. The final stage merges object captions and masks into spatial and relational dense captions. To accelerate the labeling process and improve caption quality, we present two VLM models: the Detailed Region Caption model and the Spatial Caption Merging model. Extensive experiments on various settings, including vision-language understanding, visual grounding, and region caption generation, demonstrate the effectiveness of our DenseWorld-1M dataset and labeling models.

cs.CV↗

arXiv2025

Mixed-R1: Unified Reward Perspective For Reasoning Capability in Multimodal Large Language Models

Recent works on large language models (LLMs) have successfully demonstrated the emergence of reasoning capabilities via reinforcement learning (RL). Although recent efforts leverage group relative policy optimization (GRPO) for MLLMs post-training, they constantly explore one specific aspect, such as grounding tasks, math problems, or chart analysis. There are no works that can leverage multi-source MLLM tasks for stable reinforcement learning. In this work, we present a unified perspective to solve this problem. We present Mixed-R1, a unified yet straightforward framework that contains a mixed reward function design (Mixed-Reward) and a mixed post-training dataset (Mixed-45K). We first design a data engine to select high-quality examples to build the Mixed-45K post-training dataset. Then, we present a Mixed-Reward design, which contains various reward functions for various MLLM tasks. In particular, it has four different reward functions: matching reward for binary answer or multiple-choice problems, chart reward for chart-aware datasets, IoU reward for grounding problems, and open-ended reward for long-form text responses such as caption datasets. To handle the various long-form text content, we propose a new open-ended reward named Bidirectional Max-Average Similarity (BMAS) by leveraging tokenizer embedding matching between the generated response and the ground truth. Extensive experiments show the effectiveness of our proposed method on various MLLMs, including Qwen2.5-VL and Intern-VL on various sizes. Our dataset and model are available at https://github.com/xushilin1/mixed-r1.

cs.CL↗

arXiv2025

FoMoH: A clinically meaningful foundation model evaluation for structured electronic health records

Foundation models (FMs) promise to address core limitations of traditional supervised machine learning: (i) reliance on large amounts of labeled data, (ii) task specificity, and (iii) poor transportability. Despite methodological advances in structured electronic health record (EHR) foundation models, no systematic benchmark has validated whether these models meaningfully deliver on these promises. We introduce a benchmark of 14 clinically meaningful prediction tasks spanning patient prognosis and early diagnosis of acute and chronic conditions. We benchmark 6 state-of-the-art EHR FMs beyond population-level discrimination, emphasizing the need for evaluating their calibration and fairness, with rigorous controls for data contamination and reproducibility across more than 6 million patients from Columbia University Irving Medical Center and MIMIC-IV. Our benchmark identifies that FMs deliver on some of their promises. In particular, top-performing FMs outperform traditional baselines on discriminative performance, especially under limited labeled data, and exhibit more equitable performance across socio-medical groups. However, these models may underperform in low-prevalence settings, as pretraining losses may discard discriminative information about such conditions, and present lower calibration under limited labeled data. Further, cross-institutional transportability remains a challenge for structured EHR FMs. Together, these findings advance our understanding of EHR FMs' potential for clinical utility, highlight critical gaps that remain to be addressed, and provide a reproducible framework to track progress.

cs.LG↗

arXiv2025

Seed1.5-VL Technical Report

We present Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning. Seed1.5-VL is composed with a 532M-parameter vision encoder and a Mixture-of-Experts (MoE) LLM of 20B active parameters. Despite its relatively compact architecture, it delivers strong performance across a wide spectrum of public VLM benchmarks and internal evaluation suites, achieving the state-of-the-art performance on 38 out of 60 public benchmarks. Moreover, in agent-centric tasks such as GUI control and gameplay, Seed1.5-VL outperforms leading multimodal systems, including OpenAI CUA and Claude 3.7. Beyond visual and video understanding, it also demonstrates strong reasoning abilities, making it particularly effective for multimodal reasoning challenges such as visual puzzles. We believe these capabilities will empower broader applications across diverse tasks. In this report, we mainly provide a comprehensive review of our experiences in building Seed1.5-VL across model design, data construction, and training at various stages, hoping that this report can inspire further research. Seed1.5-VL is now accessible at https://www.volcengine.com/ (Volcano Engine Model ID: doubao-1-5-thinking-vision-pro-250428)

cs.CV↗