Searcharxiv⌕ Search

arXiv subjects

Jingdong Wang

Publications and source records attributed to Jingdong Wang.

At least 19 recordsLinked to original sources
arXiv2026

SlotMemory: Object-Centric KV Memory for Streaming Long-Video Generation

Streaming video generation models typically rely on temporal-centric memory, organizing historical context as latent queues, compressed context tokens, or unclustered token sets. This often causes identity drift and semantic inconsistency when entities exit the frame or during interactive prompt transitions. To address these limitations, we propose SlotMemory, an object-centric Key-Value (KV) memory mechanism for streaming video diffusion. Our approach shifts memory abstraction from ``when'' an event occurred to ``what'' is represented by decomposing the transformer's KV manifold into discrete, reusable semantic slots. Because the slots serve only as routing addresses and the stored content remains the original transformer KV tokens, the mechanism adds entity-level persistence and prompt-aware retrieval across long horizons. Evaluated on 60-second interactive narratives using the Wan2.1-T2V-1.3B backbone, SlotMemory achieves the highest Dynamic Degree among evaluated streaming methods and the second-highest aggregate Quality Score, preserving prompt adherence rather than trading it away for consistency. At the 30-second horizon, SlotMemory leads in Total, Quality, and Dynamic Degree, improving Dynamic Degree by 31.2% over MemFlow, the strongest memory-based baseline. Our results support treating structured semantic representation, rather than raw temporal capacity, as a critical primitive for persistent long-form video synthesis.

cs.CV↗
arXiv2026

Unveiling the Secret of AdaLN-Zero in Diffusion Transformer

Diffusion transformer (DiT), a rapidly emerging architecture for image generation, has gained much attention. However, despite ongoing efforts to improve its performance, the understanding of DiT remains superficial. In this work, we delve into and investigate a critical conditioning mechanism within DiT, adaLN-Zero, which achieves superior performance compared to adaLN. Our work studies three potential elements driving this performance, including an SE-like structure, zero-initialization, and a "gradual" update order, among which zero-initialization is proved to be the most influential. Building on this understanding, we propose an analysis-guided initialization strategy, termed adaLN-Gaussian, which serves both as an empirical validation of our analysis and as a practical initialization method that consistently improves optimization efficiency. On the other hand, inspired by the SE-like structure, we introduce an improved conditioning mechanism called SE-adaLN-Zero. Extensive experiments following DiT on four datasets, especially on ImageNet1K, demonstrate the effectiveness and generalization of adaLN-Gaussian and SE-adaLN-Zero. Beyond class-to-image generation, we also evaluate the generalization of the two improved methods on text-to-image generation.

cs.CV↗
arXiv2026

MultiWorldBench: Do Independently Controlled Views Describe One Shared World?

Multiplayer world models must ensure that independently controlled views remain consistent with a single shared and persistent world. We introduce MultiWorldBench, a diagnostic Minecraft benchmark comprising 495 case configurations across seven task suites and ten capabilities, including independent control, cross-view motion, shared-state synchronization, persistence, structural reasoning, concurrent interaction, and delayed revisit. Using Engine GT as a reference, we evaluate two multiplayer world models, Solaris (MP) and Gamma-World, together with a single-player generation variant, Solaris (SP). Among the generated systems, Gamma-World achieves the highest ten-capability average at 21.39, followed by Solaris (MP) at 20.88 and Solaris (SP) at 10.06, while Engine GT reaches 91.69. Gamma-World performs better on several control, shared-state, and revisit capabilities, whereas Solaris (MP) leads in cross-view motion and race-condition consistency. Nevertheless, all generated systems score at most 8.00 on state persistence and 1.33 on structural consistency, and none succeeds in spatial reasoning or building-identity preservation. Human pairwise evaluation yields the same overall system ranking and shows strong alignment with the automatic evaluation, with a mean dimension-level Spearman correlation of 0.95. These results show that current multiplayer world models still exhibit a substantial gap relative to Engine GT in maintaining a coherent shared world.

cs.AI↗
arXiv2026

Plausible Future, Wrong Choice: Steering World-Action Models toward Intended Future

World-Action Models (WAMs) have emerged as a promising paradigm for embodied generalist policies by jointly modeling future world evolution and robot actions across diverse manipulation tasks. However, representing diverse futures is only the first step: a capable WAM must also generate futures that match the requirements of the current task and execution progress. In long-horizon manipulation where similar observation-action segments occur across different tasks and stages, existing approaches may generate locally plausible yet misdirected rollouts to drift toward wrong task stage, i.e., a phenomenon we term Stage Confusion. To address this problem, we propose TaskWAM, a closed-loop, stage-aware framework that represents fine-grained task requirements with Mode Tokens and tracks execution progress with State Tokens, forming a structured task state to steer world--action generation toward intended future. We further introduce StageBench, a controlled multi-stage manipulation benchmark with parameterized stage variations and automated evaluation.Extensive experiments on both simulation benchmarks and real-world tasks demonstrate that TaskWAM effectively reduces confusion and achieves competitive performance with state-of-the-art methods. Moreover, TaskWAM provides strong plug-and-play transfer across heterogeneous WAM architectures and robot embodiments, including single-arm robots, dual-arm grippers, dexterous hands, and humanoid robots.

cs.RO↗
arXiv2026

4LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling

For video generation in dynamic world modeling, observed static and dynamic content should be efficiently reused to provide conditions for repeated generation at different times and viewpoints. We introduce 4LC, a 4D latent cache with composite visual representations and a flexible spatiotemporal access interface. As a latent cache, 4LC directly reuses visual representations of static and dynamic content to provide generation conditions, avoiding per-request RGB rendering and re-encoding. As a composite cache, 4LC binds visual representations from multiple visual encoders (e.g., VAE, DINO, V-JEPA, VGGT) to shared geometry and motion, allowing the composite representations to describe the dynamic world with rich information. Besides, we design a flexible spatiotemporal interface, which directly reads the cache at independently specified times and viewpoints and combines readouts across observations, providing a common conditioning format for diverse downstream tasks. Building on 4LC, we develop 4LC-Video to address multiple video generation tasks through different configurations of source observations, spatiotemporal requests, and historical reuse, including camera control, temporal control, novel-view synthesis, scene revisits, and long-term generation. Our experiments show more efficient condition acquisition, effective spatiotemporal controllability, and scene consistency across tasks using a single 4LC-Video model without task-specific fine-tuning.

cs.CV↗
arXiv2026

LeWAM: Latent evolving World Action Model

World Action Models (WAMs) jointly model action generation and environment dynamics and are mostly built on pretrained Video Diffusion Models (VDMs). In VDM-based WAMs, observations are first encoded by a VAE, and the resulting compressed latents are then processed by large video diffusion backbones to extract effective features for action generation. However, this paradigm ties WAM performance and training cost to large-scale video generation pretraining, limiting WAM efficiency and scalability. In this paper, we theoretically and empirically investigate how visual representations affect action generation in WAMs. Our results show that predictive embeddings from Joint-Embedding Predictive Architecture (JEPA) encoders better support action generation than compressed VAE latents, with I-JEPA performing best in our encoder comparison. Based on these findings, we propose LeWAM, which conditions action generation on JEPA embeddings and models environment evolution by predicting future embeddings in the same space, without relying on a video diffusion backbone. We further find that imitation learning matches demonstrated actions but does not distinguish better actions from worse ones, even though small action deviations can greatly affect task success. To address this limitation without additional environment interaction or the human oversight required for resets and safety, we introduce Demonstration-Guided DPO (DemoDPO), an offline preference refinement stage that derives preference supervision directly from demonstrations. With only 0.4B trainable parameters, LeWAM achieves an average success rate of 92.28\% on RoboTwin 2.0, comparable to that of state-of-the-art VLAs and WAMs, and maintains practical effectiveness on real-world manipulation tasks.

cs.CV↗
arXiv2026

World-Ego Modeling for Embodied Video Generation in Long-Horizon Navigation-Manipulation Tasks

Embodied video world models typically capture both scene evolution and the robot's behavior, which we refer to as the \emph{world} and the \emph{ego}, respectively. The world and the ego exhibit different underlying dynamics: world prediction relies primarily on visual history and emphasizes scene stability, whereas ego prediction relies more strongly on the current instruction and emphasizes accurate instruction following. Modeling both components within a single generation stream can entangle these different dependencies, making it difficult to specialize the prediction of either component. Consequently, it becomes difficult to simultaneously maintain scene consistency and accurate instruction following, particularly in long-horizon navigation-manipulation tasks. In this paper, we propose to decompose an embodied video into the world and the ego and disentangle their generation processes. Specifically, we define the world as the background and currently unmanipulated objects, and the ego as the robot and currently manipulated objects. Based on this definition, we develop the \emph{World-Ego Model} (WEM), which combines a vision-language state predictor using role-conditioned attention (RCA) and asymmetric query budgets with a semantic-routed mixture-of-experts (SR-MoE) diffusion generator. To enable rigorous evaluation, we further construct HTEWorld, a dataset and benchmark for long-horizon embodied video generation with hybrid navigation-manipulation tasks, providing 125K training video clips comprising over 4.5M frames with fine-grained instructions, together with 300 multi-turn evaluation trajectories covering over 2K instructions. Extensive experiments show that WEM achieves state-of-the-art performance on HTEWorld while remaining competitive on existing manipulation-oriented evaluations.

cs.CV↗
arXiv2026

Online Video Agent Harness for Long Video Understanding

Long video understanding often behaves like a visual needle-in-a-haystack problem: query-relevant evidence is sparsely distributed across long temporal spans, while packing dense frames into a single VLM context incurs \textit{context rot} and high cost. Existing video agents often rely on query-agnostic offline preprocessing or ad hoc tool sets, which can miss query-specific details and waste computation. In this work, we present VideoXAgent, a purely online video-agent harness for long video understanding that starts from the given video file and user query, plans and decomposes the task, invokes specialized expert tools on demand, and aggregates multimodal evidence to produce a final answer while resolving conflicts among observations. To support this on-demand invocation, we design a suite of heterogeneous expert tools guided by a data-driven taxonomy of atomic capabilities, spanning scripts, VLMs, and domain models (e.g., detection, OCR, ASR, face recognition). The harness further enforces objective evidence prompting and budget-aware control to curb hallucination and non-termination. Across Video-MME-Long, LongVideoBench-Long, LVBench, and MINERVA, VideoXAgent is competitive with frontier LMMs and video agents under a smaller context footprint---about 50k tokens of agent context per sample, even on hour-long videos. In particular, on complex video-reasoning benchmarks such as MINERVA, it matches this level while using only about 15\% of the context of a 1,024-frame dense-packing baseline. Notably, the harness remains effective with a visually weak or even text-only orchestrator, suggesting that strong long-video understanding can emerge from progressive agentic evidence seeking rather than from packing the full video into a single context. Project page: https://go-agent-x.github.io/video_agent_harness/

cs.CV↗
arXiv2026

Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams

Multimodal Large Language Models (MLLMs) have been growing the capability for scientific writing and collaboration. For example, OpenAI Prism is a free workspace for scientific writing and collaboration. One important feature in Prism is turning scientific diagrams directly into LaTeX TikZ code. In this paper, we build a benchmark, Diagram-MMU, a multi-modal benchmark designed to assess MLLMs' ability for scientific diagram parsing and understanding. Diagram-MMU features 3.7k curated diagrams and 18.3k human-validated questions across six domains. It evaluates MLLMs on three tasks common in vibe writing workspaces: diagram-to-code parsing, diagram-to-code editing, and diagram question answering, alongside agentic settings per task. The evaluation of 12 MLLMs reveals that diagram-to-code tasks are more challenging than diagram question answering: models can reason well over diagrams but struggle to parse and edit them, underscoring the need for methods to enhance MLLMs' capability in diagram-to-code generation. Under agentic settings, most models improve parsing and editing performance but degrade on question answering, while Claude-4.6 Opus consistently improves across all three tasks. Project Page: https://vi-ocean.github.io/projects/diagram-mmu.

cs.CV↗
arXiv2026

MVHOI: Bridge Multi-view Condition to Complex Human-Object Interaction Video Reenactment via 3D Foundation Model

Human-Object Interaction (HOI) video reenactment aims to transfer the interaction dynamics of a source video to a novel target object while preserving realistic hand-object coordination. Existing methods typically rely on sparse 2D motion controls and monocular references, which are insufficient for complex out-of-plane motion and large viewpoint changes. We present MVHOI, a two-stage framework combining implicit motion extraction, 3D-aware multi-view reasoning, and video generation. In the first stage, a motion extractor encodes object dynamics into implicit motion descriptors. Conditioned on these descriptors, our Motion-Driven Object Prior (MDOP) module queries a 3D foundation model over multi-view references of the target object and autoregressively predicts coarse object anchors, a sequence of images that track the object's evolving orientation and appearance under the source motion without any explicit pose estimation. In the second stage, a DiT-based video generation model uses these anchors as structural guidance and the multi-view references as appearance guidance. We further reuse cross-view attention from MDOP as a soft attention bias to reduce reference-view confusion. For long videos, a cross-iterative inference strategy refreshes subsequent object priors using refined video outputs. Experiments demonstrate consistent improvements over state-of-the-art methods in object fidelity, motion consistency, visual quality, and interaction realism.

cs.CV↗
arXiv2026

Prompt Reinjection: Alleviating Prompt Forgetting in Multimodal Diffusion Transformers for Text-to-Image Generation

Multimodal Diffusion Transformers (MMDiTs) for text-to-image generation maintain separate text and image branches, with bidirectional information flow between text tokens and visual latents throughout denoising. In this setting, we observe a prompt forgetting phenomenon: the semantics of the prompt representation in the text branch is progressively forgotten as depth increases. We further verify this effect on three representative MMDiTs--SD3, SD3.5, and FLUX.1 by probing linguistic attributes of the representations over the layers in the text branch. Motivated by these findings, we introduce a training-free approach, prompt reinjection, which reinjects prompt representations from early layers into later layers to alleviate this forgetting. Experiments on GenEval, DPG, and T2I-CompBench++ show consistent gains in instruction-following capability, along with improvements on metrics capturing preference, aesthetics, and overall text--image generation quality.

cs.CV↗
arXiv2026

GroundShot: Visually Consistent Multi-Shot Long Video Generation via Entity-Grounded Shot Scheduling

Generating visually consistent multi-shot videos remains an open challenge. As videos span more shots, inconsistencies can accumulate across shots, causing entities that reappear across shots -- characters, objects, and locations -- to drift away from how they first appear. We observe that viewers judge consistency by comparing each later appearance of an entity with its first clear appearance; the visual quality of this initial appearance sets the consistency ceiling for all that follows. Motivated by this, we present \textbf{GroundShot}, a training-free, model-agnostic agentic framework for entity-grounded multi-shot generation. GroundShot builds an entity-level visual memory online from accepted generated shots: it schedules shots' generation order by their expected usefulness as entity references, grounds entities from generated videos, verifies their reliability before adding them to memory, and retrieves suitable entity references from memory before each shot is generated. To evaluate this entity-centered view of consistency, we further introduce \textbf{GroundBench}, a diagnostic benchmark that measures consistency at the entity level while isolating controlled challenge dimensions. Experiments show that GroundShot improves multi-shot consistency over existing methods while requiring no additional training or model modification.

cs.CV↗
arXiv2026

From Draft to Draft-Free: One-Step Video Object Removal via Privileged Distillation and Fast Planting

Video object removal is a fundamental yet challenging task in video editing. Despite recent progress, existing methods typically fall into two categories. Traditional approaches based on optical flow or attention mechanisms often introduce noticeable artifacts and yield unnatural results. In contrast, diffusion-based methods improve visual realism but demand multiple denoising steps, limiting their practicality. To address these issues, we propose From-Draft-to-Draft-Free (D2DF), a framework that distills the ability of transforming coarse drafts into refined videos into a one-step video generation model. Within D2DF, a teacher model is trained to refine low-quality removal results ("drafts") into high-fidelity videos by multiple steps. Then, through Prior-Privileged Consistency Distillation (PPCD), we distill this capability into a student model that performs one-step removal conditioned on the draft. To eliminate draft dependency, we introduce a Self-Guided Fast Planting (SGFP) module based on our Temporal Masked Transformer that autonomously generates scene-consistent pseudo-drafts in latent space, enabling a fully draft-free one-step model. Extensive experiments show that both draft-conditioned and draft-free versions achieve state-of-the-art performance on multiple metrics, surpassing traditional and multi-step generative methods in both quality and efficiency. The denoising process for a single video takes only about 1 second.

cs.CV↗
arXiv2026

MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture

This paper studies the training-testing discrepancy (a.k.a. exposure bias) problem for improving the diffusion models. During training, the input of a prediction network at one training timestep is the corresponding ground-truth noisy data that is an interpolation of the noise and the data, and during testing, the input is the generated noisy data. We present a novel training approach, named MixFlow, for improving the performance. Our approach is motivated by the Slow Flow phenomenon: the ground-truth interpolation that is the nearest to the generated noisy data at a given sampling timestep is observed to correspond to a higher-noise timestep (termed slowed timestep), i.e., the corresponding ground-truth timestep is slower than the sampling timestep. MixFlow leverages the interpolations at the slowed timesteps, named slowed interpolation mixture, for post-training the prediction network for each training timestep. Experiments over class-conditional image generation (including SiT, REPA, and RAE) and text-to-image generation validate the effectiveness of our approach. Our approach MixFlow over the RAE models achieve strong generation results on ImageNet: 1.43 FID (without guidance) and 1.10 (with guidance) at 256 x 256, and 1.55 FID (without guidance) and 1.10 (with guidance) at 512 x 512.

cs.CV↗
arXiv2026

From SRA to Self-Flow: Data Augmentation or Self-Supervision?

Representation alignment has become an effective way to accelerate diffusion transformer training and improve generation quality. Recent self-alignment methods, such as SRA and Self-Flow, further remove the dependency on external pretrained encoders by constructing alignment within the diffusion model itself. However, the mechanism behind the improvement from SRA to Self-Flow, dual-time scheduling, remains under-examined: Self-Flow attributes its gain to interactions between tokens at different noise levels, where cleaner tokens help infer noisier ones. In this work, we revisit this explanation and ask whether the gain instead comes from data augmentation along the noise dimension. To disentangle these factors, we introduce Attention Separation, which preserves the same dual-timestep input as Self-Flow while blocking attention between tokens assigned to different noise levels. Surprisingly, removing such interaction does not degrade performance and can even improve it, suggesting that the improvement from SRA to Self-Flow mainly comes from data augmentation. Furthermore,We show that Attention Separation itself provides an augmentation effect by splitting a single image into multiple effective training parts to expand the training data. Based on these observations, we combine self-representation alignment with dual-timestep and attention-separation augmentation, and demonstrate the effectiveness of this design on ImageNet.

cs.CV↗
arXiv2026

RefAlign: Representation Alignment for Reference-to-Video Generation

Reference-to-video (R2V) generation is a controllable video synthesis paradigm that constrains the generation process using both text prompts and reference images, enabling applications such as personalized advertising and virtual try-on. In practice, existing R2V methods typically introduce additional high-level semantic or cross-modal features alongside the VAE latent representation of the reference image and jointly feed them into the diffusion Transformer (DiT). These auxiliary representations provide semantic guidance and act as implicit alignment signals, which can partially alleviate pixel-level information leakage in the VAE latent space. However, they may still struggle to address copy--paste artifacts and multi-subject confusion caused by modality mismatch across heterogeneous encoder features. In this paper, we propose RefAlign, a representation alignment framework that explicitly aligns DiT reference-branch features to the semantic space of a visual foundation model (VFM). The core of RefAlign is a reference alignment loss that pulls the reference features and VFM features of the same subject closer to improve identity consistency, while pushing apart the corresponding features of different subjects to enhance semantic discriminability. This simple yet effective strategy is applied only during training, incurring no inference-time overhead, and achieves a better balance between text controllability and reference fidelity. Extensive experiments on the OpenS2V-Eval benchmark demonstrate that RefAlign outperforms current state-of-the-art methods in TotalScore, validating the effectiveness of explicit reference alignment for R2V tasks.

cs.CV↗
arXiv2026

Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling

Recent visual generation models have made major progress in photorealism, typography, instruction following, and interactive editing, yet they still struggle with spatial reasoning, persistent state, long-horizon consistency, and causal understanding. We argue that the field should move beyond appearance synthesis toward intelligent visual generation: plausible visuals grounded in structure, dynamics, domain knowledge, and causal relations. To frame this shift, we introduce a five-level taxonomy: Atomic Generation, Conditional Generation, In-Context Generation, Agentic Generation, and World-Modeling Generation, progressing from passive renderers to interactive, agentic, world-aware generators. We analyze key technical drivers, including flow matching, unified understanding-and-generation models, improved visual representations, post-training, reward modeling, data curation, synthetic data distillation, and sampling acceleration. We further show that current evaluations often overestimate progress by emphasizing perceptual quality while missing structural, temporal, and causal failures. By combining benchmark review, in-the-wild stress tests, and expert-constrained case studies, this roadmap offers a capability-centered lens for understanding, evaluating, and advancing the next generation of intelligent visual generation systems.

cs.CV↗
arXiv2026

GeoLoom: High-quality Geometric Diagram Generation from Textual Input

High-quality geometric diagram generation presents both a challenge and an opportunity: it demands strict spatial accuracy while offering well-defined constraints to guide generation. Inspired by recent advances in geometry problem solving that employ formal languages and symbolic solvers for enhanced correctness and interpretability, we propose GeoLoom, a novel framework for text-to-diagram generation in geometric domains. GeoLoom comprises two core components: an autoformalization module that translates natural language into a specifically designed generation-oriented formal language GeoLingua, and a coordinate solver that maps formal constraints to precise coordinates using the efficient Monte Carlo optimization. To support this framework, we introduce GeoNF, a dataset aligning natural language geometric descriptions with formal GeoLingua descriptions. We further propose a constraint-based evaluation metric that quantifies structural deviation, offering mathematically grounded supervision for iterative refinement. Empirical results demonstrate that GeoLoom significantly outperforms state-of-the-art baselines in structural fidelity, providing a principled foundation for interpretable and scalable diagram generation.

cs.CV↗