多模态生成与编辑

Recent advances in Multimodal Large Language Models (MLLMs) have extended Image Aesthetic Assessment (IAA) beyond scalar scores toward interpretable critique and guidance. Yet existing benchmarks mainly assess intrinsic visual quality or fixed domain criteria, leaving open whether an appealing image is appropriate for a specific purpose, audience, cultural setting, or domain convention. We introduce AesCanvas, a unified suite with two complementary components: CritiqueCanvas with 519,136 instruction-response pairs from 54,300 images supports long-form, multi-dimensional critique across photography, painting, and virtual imagery, whereas ContextCanvas with 301 expert-reviewed use scenarios evaluates contextual aesthetic suitability in realistic use scenarios. Under a unified protocol, we evaluate closed-source frontier, open-weight general, and aesthetic-specific MLLMs. Results reveal a clear separation between critique generation and context-sensitive judgment: reference-based lexical and semantic metrics only partially capture critique quality, while aesthetic specialists remain competitive on selected critique metrics yet substantially lag strong general-purpose MLLMs on ContextCanvas. Further analyses show that aesthetic specialization does not reliably transfer to contextual suitability and that model decisions may fail to track or ground themselves in decisive contextual visual cues. These findings establish culturally situated, evidence-grounded suitability as a distinct objective for aesthetic modeling.

Many multimodal tasks depend on how visual elements are ordered and composed, not only on recognizing them in isolation. Internet memes are a compact case of this problem: their punchline often depends on a constrained reading order and cross-panel visual--textual cues. While large vision-language models (LVLMs) show strong performance on single-image understanding, it remains unclear whether they can perform sequence-aware reasoning over structured meme layouts, especially in Chinese social media. We introduce CMPM, a Chinese Multi-Panel Meme benchmark with 1,214 annotated samples covering five structural types, ordering dependency, panel-order constraints, and optional comment context. We formulate a two-layer evaluation: Task1 probes structure typing and order-sensitive panel sequencing (with a context ablation setting), and Task2 evaluates Chinese meme explanation generation with human ratings on five 1-3 Likert dimensions (visual, panel, humor, context, and faithfulness). We benchmark five representative LVLMs under a unified protocol. Results indicate that canonical-display accuracy is not by itself evidence of order understanding: the primary shuffled condition produces a sharp accuracy drop, revealing a persistent gap in order-sensitive multimodal reasoning. Task2 preferences place Gemini 3.1 Pro and GPT-5.5 above the open models, while comment context yields only a small and mixed Core4 gain. Code and data will be released upon acceptance.

Visual storytelling requires generating images that follow a narrative while preserving consistent character identities across frames. In free-form story generation, a character is fully described only when first introduced and is later referred to by a type-level mention or pronoun. Although this setting better reflects natural storytelling, later prompts may omit important identity-related semantics, making character consistency more difficult to maintain. We propose Sidecar, a plug-and-play semantic augmentation module that preserves entity-level information from the initial description and injects the missing semantics into later prompt embeddings. Sidecar requires no additional training and does not modify the architecture of the base diffusion model. Experiments on FreeStoryBench show that Sidecar consistently improves prompt-image alignment and character consistency across multiple SDXL- and FLUX-based baselines, with negligible computational overhead.

Vision generative artificial intelligence (AI) has emerged as one of the most rapidly advancing areas of deep learning. The explosion of multimodal models has made them widely associated with text-to-image applications running on large datacentres. However, vision generative models are equally needed in applications that operate under strict hardware constraints at the edge, including autonomous vehicles, agricultural sensors, and mobile devices. In this Perspective, we argue that progress in vision generative AI has been driven by output quality, with hardware evolving reactively to accommodate growing model demands. We quantify the parameter cost and energy efficiency of these models across a range of accelerator platforms, and map four generative model families against seven real-world application domains. Finally, we advocate a software-hardware co-design approach, where deployment constraints are considered from the start of the design process, ensuring that the "right model" runs on the "right hardware" to serve the "right application", making generative AI deployment sustainable and accessible across a much broader range of platforms.

Generative text-to-image (T2I) systems are increasingly adopted in creative practice, yet their normative behaviors remain underexplored from the perspective of creative practitioners. In this workshop paper, we present a within-subject study with 14 creative practitioners using Stable Diffusion to create illustration from two tasks of differing specificity. We investigate whether and how practitioners perceive normative behavior in a T2I system and how it impacts their creative process depending on task specificity. Our findings show that participants perceived normative behavior through invariant patterns, stereotypical output, and unsolicited omission or addition of details. These experiences led to feelings of disempowerment and creative compromise. We discuss implications for XAIxArts, including prompt transparency and artist empowerment in creative contexts.

State-of-the-art action-conditioned video models are typically restricted to a single robot embodiment, preventing them from leveraging the vast corpus of heterogeneous video data that contains rich signals for learning generalizable physics. To bridge this gap, we introduce CLAP, a framework for cross-embodiment action-conditioned video generation capable of being trained on diverse, internet-scale videos across human and robotic agents. CLAP is grounded in the insight that universal physical laws govern spatiotemporal dynamics regardless of the actor. However, cross-embodiment learning is non-trivial because action representations vary sharply across robot platforms and are typically absent in human videos. CLAP addresses this fundamental challenge through the following core contributions. First, CLAP reconciles disparate action spaces using end-effector poses, language instructions, and latent actions. Second, to resolve their individual limitations, CLAP introduces a curriculum-based cross-embodiment learning recipe that first learns foundational physical priors across unlabeled video data using latent actions and subsequently grounds them in end-effector action spaces for zero-shot deployment to real-world tasks. Crucially, CLAP approaches or surpasses state-of-the-art single-embodiment video models in challenging environments like DROID. These performance advantages compound via few-shot adaptation to establish a novel paradigm for training single-embodiment video world models. Ultimately, CLAP delivers the most comprehensive suite of action-conditioned video world models to date - spanning diverse action-conditioning spaces (end-effector, language, and latent) and robot morphologies (including cross-embodiment, DROID, Bridge, bimanual YAM robots, and G1 humanoids). We open-source all code and models. Project Website at https://omni-clap.github.io .

Streaming autoregressive video models generate long videos chunk by chunk, using historical memory to maintain consistency. Existing methods typically expose subject and scene queries to history through similar policies. This stabilizes the subject, but can also lock backgrounds, viewpoints, and scene structure to previously generated states even when local motion continues. We call this failure memory-anchored scene under-progression; consistency and motion metrics alone can miss it. We introduce TetherMem, a training-free, query-aware spatiotemporal memory router for frozen video generators. TetherMem separates subject and scene queries and modulates historical access with region- and age-conditioned priors: subject queries retain identity-bearing history, while scene queries reduce reliance on subject history and stale backgrounds. Across 2,400 blinded pairwise judgments from 10 annotators, TetherMem achieves the highest estimated expected preference among eight streaming long-video baselines for overall quality (0.780) and scene progression (0.769). On complete 30-second videos, it sustains changes in background, viewpoint, and scene state while preserving subject recognizability and temporal continuity.

Visual intelligence seeks to perceive, interpret, and synthesize the visual world and is central to modern computer vision. Human-centered visual intelligence is especially demanding because it studies people as expressive, socially situated subjects whose meaning is rarely conveyed by appearance alone. It couples vision with audio and language across four representative tasks: human emotion recognition, human video generation, human voice cloning, and human video matting. Yet existing resources remain task-specific, providing modalities and annotations for individual problems rather than a shared foundation coordinating understanding and generation. This limits multimodal signal use and broader research. We address this gap with HUG-VIS, a unified benchmark for Human-centered Understanding and Generation in Visual Intelligence. It contains 8,400 seated half-body videos of 30 professional actors, each performing the same 280 emotion-action-prompt assignments under a controlled Mandarin studio protocol, with synchronized video, audio, text, and alpha mattes. We evaluate diverse open- and closed-source models across the four tasks under a unified zero-shot protocol using automatic metrics, criterion-specific mean opinion scores, and multiple cross-task analyses. Results show that (i) linguistic content dominates current emotion recognition, while purely visual affect recognition is weakest; (ii) in video generation and voice cloning, automatic metrics and human judgment agree overall but differ in their top rankings, requiring joint reporting; (iii) boundary fidelity under motion is the main remaining obstacle for human matting; and (iv) task difficulty varies across emotions, models, and metrics, with notable cross-task correlations. The dataset and results are available at https://github.com/GML-MMGroup/HUG-VIS.

Audio-video generation is rapidly moving from prompt-driven synthesis toward multimodal conditioning, where text, images, audio, and video can jointly shape the generated output. This shift changes the nature of safety evaluation: harmful intent may no longer reside in any single input, but instead emerge from how otherwise benign or weakly harmful conditions interact across modalities and time. Existing safety benchmarks, however, remain largely prompt-centric or tied to fixed conditioning interfaces, leaving such compositional risks difficult to study systematically. To bridge this gap, we introduce Multi2AV-Safety, the first safety benchmark, to the best of our knowledge, to cover all 11 non-singleton T/I/A/V conditioning configurations for audio-video generation, comprising 11,024 attack instances. Evaluation on Multi2AV-Safety reveals systematic weaknesses in representative multimodal safety guards across attack mechanisms and harm-evidence structures. Our evaluation reveals two complementary failure modes: harmful semantics can emerge from the combination of individually benign inputs, while explicit harmful cues can become harder to detect when mixed with benign multimodal context. Together, these results identify compositional risk perception as a central capability gap in safeguarding multimodal-conditioned audio-video generation: current safety guards fail to reliably integrate safety evidence across modalities and time, even when all conditioning inputs are observable. The dataset will be publicly released in October 2026.

Recent video generation models are increasingly framed as world models. Many physical processes can unfold in more than one valid way. Therefore, a world model should reproduce not only a plausible trajectory, but also the distribution of possible behaviors under the same initial observation and action. We call this distribution-level requirement probabilistic alignment. However, existing evaluations largely assess individual-video plausibility and do not test whether repeated generations recover the correct distribution. This raises a central question: how far are current video generators from probabilistically aligned world modeling? To answer it, we formalize probabilistic alignment as a distributional criterion for world models and introduce PAWBench, a benchmark for evaluating video generators as stochastic samplers of world dynamics. We further introduce PAWEval, an outcome-level protocol that converts repeated video rollouts into empirical distributions over possible physical behaviors. Across 50 scenarios and eleven current systems, no model consistently matches the reference probabilities while recovering the range of valid behaviors. Having established this gap, we test whether language prompts, initial noise sampling, or model training can reshape the model's predictive distribution. We believe our work can serve as a foundation for future efforts to move towards probabilistically aligned world modeling.

High similarity between first-visit and return frames does not necessarily show that a video world model remembered the scene; the intervening rollout may simply have changed very little. This ambiguity makes absolute revisit scores sensitive to rendering stability, repetitive content, and failed motion. We introduce R2M-Bench (Relative Revisit Memory Benchmark), a benchmark of observable revisit-selective consistency. For every detected return, R2M-Bench compares the revisit pair with two controls from the same rollout: a gap-matched non-revisit pair that measures generic temporal stability and a short-range pair that estimates short-horizon consistency. These comparisons produce MemoryGain (MG), the revisit advantage over the temporal baseline, and the Normalized Memory Ratio (NMR), which normalizes this advantage by the short-to-baseline dynamic range. R2M-Bench combines 100 reference scenes with three leave-and-return trajectories to form 300 instances and evaluates appearance fidelity, scene and object identity, local geometry, and persistent state. Across seven action-conditioned video world models, Overall NMR correlates with human consistency judgments at Spearman's $ρ=0.547$ (95% CI $[0.45,0.63]$). Its within-model correlation magnitude with generated motion is $0.072$, compared with $0.207$ for raw revisit similarity, indicating that relative calibration substantially reduces the slow-motion shortcut. DreamX-World-Memo achieves the highest Overall NMR among the evaluated video models. Together, these results support same-rollout relative calibration as a practical way to distinguish revisit-specific consistency from generic temporal stability.

Zero-shot image restoration methods with text-to-image latent diffusion models have achieved great success in universal image restoration tasks without training. However, applying them to video restoration will result in severe temporal flickering. In this paper, we propose a novel framework for zero-shot video restoration and enhancement which uses a text-to-image latent diffusion model and multi-modal references. Through the proposed dual prompt tuning inversion and sampling, the inference time can be reduced to nearly 1/3 of the original. The performance and temporal consistency can be also significantly stregthened. By using the proposed texture-aware video token merging, the temporal correlation between frames can be further utilized to improve the temporal consistency. We futher propose the referenced self-attention and referenced token merging to support image reference. Experimental results demonstrate the superiority of the proposed method in restoring and enhancing temporally consistent videos.

Reward models play an essential role in aligning visual generative models, yet most existing visual reward models use a single scalar score or rely on fixed criteria that cannot adapt to different instructions. This limits both interpretability and task sensitivity, especially for text-to-image generation and instruction-based image editing, where different inputs require different evaluation dimensions. We propose RubricRM, a pairwise generative reward modeling framework that first produces an input-specific rubric with evaluation dimensions, weights, and scoring criteria, and then applies the rubric to score candidate images. We train dedicated RubricRM models for text-to-image generation and image editing using a two-stage training pipeline: supervised fine-tuning teaches the model the rubric-based scoring paradigm, while GRPO further improves scoring through fine-grained dimension-level rewards. Experiments on multiple generation and editing benchmarks show that RubricRM outperforms existing specialized reward models and remains competitive with strong proprietary MLLM judges despite using smaller backbones. Our models, data, and code are available at https://github.com/zijiankan/RubricRM.

Explorable image-to-scene generation is essential for applications in gaming, robotics, and virtual reality. Existing methods based on video diffusion model (VDM) commonly rely on incomplete conditioning signals such as sparse point clouds or 2D panoramas, leading to stochastic hallucinations, long-term drifts and suboptimal 3D consistency. We present SpatialCrafter, a novel two-stage framework that addresses these issues by introducing a global 3D proxy for high-fidelity image-to-scene generation. Specifically, we decompose the generation process into global proxy generation and appearance refinement. For proxy generation, we propose a Point-anchored Sparse Structure~(PaSS) Flow module that predicts a spatially aligned and geometrically consistent 3D proxy. For appearance refinement, we re-frame the VDM as a Generative Deferred Refiner which synthesizes high-frequency photorealistic details upon proxy-defined scene geometry. To better integrate the proxy with the pre-trained VDM, we introduce Parallel Geometry Injection and Proxy-Aware Corruption training strategies, which improve robustness to proxy artifacts without disrupting the pretrained generative manifold. Furthermore, as no suitable dataset exists for this explorable scene generation task, we construct a new large-scale dataset of 115K scenes. To the best of our knowledge, it is the first hybrid dataset for image-to-scene generation. Extensive experiments on both synthetic and real-world datasets show that SpatialCrafter outperforms state-of-the-art methods, mitigates long-term drift, and remains robust and consistent under rapid camera motion and extreme viewpoint changes. Our project page: \href{https://fangchuan.github.io/SpatialCrafter/}{fangchuan.github.io/SpatialCrafter/}

Conventional video editing primarily focuses on scene-level content, whereas live streaming places greater emphasis on the human subject. However, directly applying existing video-editing methods to human-centric live streaming remains challenging, as they may introduce facial-expression inconsistencies and typically depend on multiple offline inference steps, making them unsuitable for real-time interaction. We propose EditaLive, a novel framework for real-time streaming character video editing. In detail, we start from a pretrained image animation model (Wan-Animate), which naturally decouples appearance from motion, and repurpose it as the base model for instruction-based human-centric video editing by reference frame editing and video reconstruction via the collected CharEdit-50K dataset. Besides, we adapt the model from offline bidirectional to causal streaming generation, and design an aligned self-rollout distillation strategy that compresses the model into a two-step sampler, where fixed RoPE and align forcing reduce training--inference discrepancies, and first-frame preserved sparse attention filters redundant historical information to mitigate appearance drift. Extensive experiments demonstrate that EditaLive delivers state-of-the-art editing performance with faithful preservation of facial expressions and low-latency real-time streaming inference.

In recent years, image-to-video (I2V) generation models have made remarkable progress in subject consistency and temporal coherence, enabling high quality video synthesis. However, these advances also introduce new safety risks. Existing studies mainly focus on jailbreak attacks involving single frame violations, while largely overlooking the temporal dimension unique to video generation models. In this paper, we investigate three attack scenarios and uncover a temporal vulnerability in I2V systems: unsafe semantics may emerge not from a single frame, but from semantic composition over time. We further identify two key challenges in such attacks: temporal abstraction and semantic camouflage. To address these issues, we propose TempJail, a novel temporal jailbreak framework for I2V systems. For temporal abstraction, we decompose a target malicious caption into an initial frame visual condition and a temporal text instruction. For semantic camouflage, on the image side we model semantic injection as controlled latent perturbation in diffusion sampling and introduce gradient guidance from pretrained encoders. On the text side, we rewrite the caption into an innocuous ``subject-action-scene'' template that bypasses safety filters while preserving temporal guidance. In the black-box inference phase, these two modalities jointly enable malicious semantics to be gradually triggered over time. Experiments on closed-source commercial models, including Kling, Seedance, Veo and PixVerse, show that TempJail improves attack success rate over prior state-of-the-art methods by 23.3% under GPT-5.2 evaluation and 22.0% under human evaluation. Our codes are available at \href{https://github.com/luqi-glory/TempJail}{GitHub}.

Low-bit quantization offers a promising avenue for reducing the computational and memory demands of Multimodal Large Language Models (MLLMs). Recent hardware support for low-precision formats, ranging from MXFP8 to ultra-low-bit formats such as MXFP4 and HiF4, has accelerated research into efficient MLLM training and deployment. In this work, we present a systematic study of these quantization schemes in representative MLLMs that span both video generation and reasoning tasks. Our analysis shows that MXFP8 achieves near-lossless performance, whereas aggressive 4-bit quantization leads to significant degradation. Through extensive ablations, we identify activation quantization as the primary source of this performance loss, contributing substantially more than weight quantization. Motivated by this observation, we propose Residual Fallback Quantization (RFQ), a lightweight activation reconstruction framework that supplements the primary ulta-low-bit activation representation with an auxiliary quantized residual pathway. By explicitly modeling and compensating for quantization errors, RFQ improves activation fidelity while preserving the efficiency advantages of ultra-low-bit computation. RFQ requires no architectural modifications and incurs negligible computational overhead. Extensive experiments on Wan2.2 and Qwen3-VL demonstrate that RFQ consistently recovers a substantial portion of the performance lost under the quantization of MXFP4 and HiF4, significantly narrowing the gap to BF16 baselines across both generation and 4 reasoning benchmarks. Our findings establish activation quantization as the dominant bottleneck in ultra-low-bit MLLMs and highlight residual-based activation reconstruction as an effective and practical strategy for robust 4-bit deployment.

Long autoregressive video generation faces a fundamental memory challenge: with a finite attention window, a model must decide which information from an ever-expanding history to retain. Existing methods organize memory temporally, preserving recent frames while compressing or discarding older ones. We instead propose RECAP-Forcing, organizing memory by appearance novelty. A long video is not merely a sequence of frames, but an evolving cast of subjects, objects, and scenes whose identities must remain consistent over time. We organize memory by retaining the KV cache associated with newly appearing content--such as entering subjects, disoccluded regions, and newly introduced scenes--at the moment it first becomes visible, prioritizing novelty over recency. Memory should scale with the amount of newly introduced content, rather than with video length. This appearance-indexed memory makes long-range consistency an explicit property of the memory structure. Our framework unifies two mechanisms under this single principle. At the beginning of a video, when all visible content is novel, an attention sink preserves the initial scene. As the video evolves, an optical-flow-based novelty bank extends the same principle by selectively retaining newly revealed content. As a training-free inference method with no additional learnable parameters, RECAP-Forcing consistently improves visual quality and semantic fidelity across multiple strong baselines and outperforms existing memory methods.

Diffusion multimodal large language models (dMLLMs) have recently emerged as a new decoding paradigm for multimodal generation. Starting from a fully masked sequence, dMLLMs progressively decode the sequence by unmasking a subset of the remaining masked positions at each step. Since the selected tokens serve as the prediction context for subsequent steps, deciding which tokens to decode is crucial to the quality of the final output. The most common strategy prioritizes tokens based on a certainty measure that tends to favor tokens frequently observed in the training data. Recent approaches instead order tokens according to their influence on subsequent predictions, but do not explicitly account for the input image. We propose the Visual Information-Guided Sampler (VIG-Sampler), which prioritizes tokens based on their attention to image tokens. We further impose a constraint that penalizes candidate tokens whose image-attention distributions are similar to those of previously selected tokens, thereby increasing the information gain of the decoded subset. Extensive experiments on 7 captioning and VQA benchmarks with 3 open-source dMLLMs demonstrate the effectiveness of VIG-Sampler, which outperforms the Info-Gain Sampler by an average of 19.3 CIDEr points across the captioning benchmarks and surpasses it on COCO Caption while using only half as many decoding steps.

Explicit visual intermediates can help multimodal large language models (MLLMs) externalize spatial evidence and updated visual states, but their utility depends on whether an image editor can faithfully realize the required transformation. We introduce Aphanta, an automated task-discovery and closed-loop diagnostic framework for the MLLM -> image editor -> MLLM pipeline. Aphanta evaluates three conditions---direct reasoning, reasoning with an editor-generated intermediate, and reasoning with an idealized reference intermediate---to separate potential visual headroom from the practical utility of current editors. Across 20 candidate tasks and multiple editor--MLLM combinations, we find that utility is strongly task-conditioned. Gains concentrate in visual cue injection, grounding, and counterfactual state realization, whereas intermediates requiring symbol-sensitive construction or structural extrapolation are substantially less reliable. On the selected positive-task subset, our consolidated Qwen pipeline improves the mean task score from 0.343 to 0.445 ($+10.2$ points; $+29.7%$ relative), while the full study also retains filtered and unsuccessful tasks to expose the boundary. These results position image editing as a specialized visual workspace rather than a universal reasoning mechanism, and establish Aphanta as a reusable protocol for measuring task--representation alignment, editor realization, and downstream pipeline utility.

Generative world models provide robots with predictive models of how the world evolves under interaction, with growing potential for simulation, planning, policy evaluation, and robot learning. Beyond clip-level future prediction, a unified generative formulation should relate actions to consequences, support flexible horizons and continuous interaction, and enable reward-driven optimization. We introduce WALL-SS, a world model that generates visual futures through Scale-wise autoregressive Scaling, enabling action-controllable and long-horizon robotic simulation. WALL-SS represents embodied trajectories as causal sequences of temporally interleaved observations and actions, making action-dependent state transitions explicit while naturally supporting variable-length generation, streaming extension through reusable causal states, and direct optimization through sequence probabilities. To make this formulation effective over long horizons, we generate each future observation in a coarse-to-fine manner and develop three complementary components within the same hierarchy. Action-conditioned next-scale prediction injects scale-aligned action representations to improve action-future coupling and model both successful and failed behaviors. Scale-compressed long-horizon memory retains recent interactions at fine resolution while compressing distant observations and actions, with scale-wise dream forcing enhancing robustness to self-generated context. Finally, on-policy alignment optimizes autoregressive visual dynamics with action-following and long-term consistency rewards while preserving the pretrained visual distribution. Experiments show that WALL-SS improves action following and trajectory accuracy, supports coherent minute-long streaming rollout under bounded memory, and consistently benefits from on-policy alignment in reducing action drift and long-horizon inconsistency.

Frontier language models are increasingly marketed as omni systems that can perceive and respond across modalities. Existing evaluation frameworks, however, focus almost exclusively on bimodal understanding, typically text plus one other modality. We propose the Modality Maturity Index (MMI), a benchmark designed to evaluate the multimodal capabilities of large language models across five modalities (text, image, audio, video and document) and combinations of up to three modalities in both inputs and outputs. MMI consists of 893 questions, each carefully crafted to require the model to demonstrate its understanding of multiple input modalities and to generate responses that incorporate various output formats. The questions are designed to be self-contained, with clear expectations for the correct modality or mix of modalities required for an accurate response. Every MMI prompt carries human-authored rubric criteria for each output modality expected in the response; a model's MMI Value expresses the average of the per-modality scores for each prompt. Because low scores can reflect either failure to generate a modality (lack of presence) or failure to generate correct content, we introduce also a supplementary Modality Presence Score (MPS), a per-prompt F1 over the expected output modalities. Applying MMI to five frontier multimodal models, we find that the MPS ranges from only 15.6 (Claude Opus 4.6) to 34.9 (GPT-5.4). Given the low availability of returned modalities to even grade, we report MPS as our main result pending model improvements. To assess the viability of judging output correctness with LLM judges and rubrics, we run a separate experiment with custom generation tools. On the assets that generates, we find that an LLM judge applying the rubrics agrees with rubric-blind human annotators (who score the outputs directly and never see the criteria) on 70.8% of judgments.

Recent advancements in generative models are pushing video generation toward unbounded streaming audio-video generation for real-time interactive worlds. However, existing benchmarks primarily evaluate completed sequences and struggle to capture streaming properties. To bridge this gap, we introduce StreamAV-Bench, the first comprehensive benchmark tailored for streaming audio-video generation. StreamAV-Bench establishes a unified evaluation framework, including the progressive track for instruction adherence and long-horizon stability, and the interactive track for interactive response and state retention and reuse. With expert-verified evaluation cases across 32 fine-grained dimensions, we conduct an extensive evaluation of 13 representative systems. Our analysis reveals that current models suffer from temporal drift in progressive generation and responsiveness bottlenecks during interactive control. Based on a comprehensive failure analysis, we share insights to advance the development of native joint audio-video streaming models.

Grasp synthesis is a core task in robotic manipulation, for which the solution typically forms a multimodal distribution rather than a point estimate. Generative robotic grasping aims to learn this distribution with deep generative models such as diffusion and flow-based approaches. The iterative nature of such generative models makes them flexible and generalizable; however, multi-step sampling impedes the time-critical operation required in robotics. We devise an approach to fast generative grasping based on MeanFlow on the product Lie group $\mathcal{G} = \mathrm{SO}(3) \times \mathbb{R}^3$. The training objective couples a purely algebraic semigroup consistency condition with Riemannian Conditional Flow Matching on $\mathcal{G}$ that anchors the average velocity to the data distribution. The resulting Lie Group-constrained MeanFlow formulation samples reliable grasps in $\leq 5$ network evaluations, matching the grasp generation performance of state-of-the-art diffusion and flow-based models on the ACRONYM dataset at millisecond-scale inference latency (up to $39\times$ speed-up). We further demonstrate that the approach directly translates to real-world robotic grasping without additional training or domain adaptation, exhibiting robust grasp synthesis under observation noise.

Autoregressive (AR) image generation has shown strong potential for scalable high-fidelity synthesis by modeling images as discrete token sequences. However, traditional next token prediction (NTP) continues to suffer from sparse and myopic supervision, insufficiently discriminative representations, and high training cost caused by dense computation over the full token sequence. To address these issues, we propose multi-token autoregressive (MTAR), a unified training framework that improves autoregressive image generation from three aspects: prediction objectives, representation regularization, and training efficiency. Specifically, MTAR introduces multi-token prediction (MTP) to alleviate the sparsity and myopia of traditional NTP by imposing joint supervision on multiple future tokens; employs token-level contrastive regularization (TCR) to explicitly enhance the separability of sampled token representations and thereby improve representation discriminability; and incorporates semantic dropping (SD) as a semantics-aware training acceleration strategy to reduce redundant computation on low-information tokens while preserving informative learning signals. All three components are applied only during training and introduce no additional overhead during autoregressive inference. On ImageNet, MTAR achieves a better balance between generation quality and training efficiency. Compared with LlamaGen, MTAR achieves up to 0.95 lower FID and 39% faster training. Moreover, even with only 1/3 of the training iterations, it still attains performance comparable to or better than the baseline, substantially reducing training time.

Action-conditioned world models have become an important foundation for embodied prediction, planning, and synthetic data generation, but their errors under new task and scene distributions are often concentrated in localized spatiotemporal regions such as robot arms, manipulated objects, contact areas, and occluded objects. This paper presents ConfAL-WM, a confidence-guided active learning framework for post-training embodied world models. Built upon EVAC, we attach a lightweight confidence probe to UNet decoder features and predict dense confidence maps in the latent space. These maps are aggregated into task-, frame-, and patch-level scores, enabling both efficient data selection and localized training enhancement. Our pipeline first retrains the confidence probe and warms up EVAC with a small subset of target-domain data, then performs task-level prescreening to allocate sampling budgets, and finally applies selected-data retraining with optional frame or patch weighted data enhancement. Experiments on RoboTwin2.0 show that confidence-guided selection improves post-training efficiency, while dense frame and patch weighting further enhances prediction quality and embodied trajectory consistency compared with scalar reward, progress, and judge-based scoring baselines. A quick visual overview of this work is available at https://ConfAL-WM.github.io.

Existing action quality assessment (AQA) datasets and methods rely primarily on visual inputs such as RGB and pose, overlooking physiological dynamics such as muscle mechanics and often modeling actions as monolithic patterns. These limitations hinder fine-grained, biomechanically grounded feedback. We introduce MyoMechanix, a multimodal ecosystem for weight-loaded actions that aligns motion with muscle activity. Expert-annotated, it contains 7,500+ samples of 20 actions from 38 subjects, with synchronized multiview RGB video, 3D pose, sEMG, and additional physiological signals, forming the largest multimodal AQA benchmark to date. We further construct the Fitness Knowledge Graph (FKG), which organizes expert annotations into structured relationships among actions, phases, key steps, errors, and corrective feedback, enabling compositional scoring and interpretable assessment. Building on these representations, we develop CUBIST (Compositional Ontological Reasoning Engine), which performs decomposition-analysis-recomposition for fine-grained error attribution and feedback generation. We also establish MyoMechanix-AQA, MyoMechanix-VideoQA, and a novel MyoMechanix-Video2EMG task. Experiments show that multimodal sensing and structured representations improve performance, interpretability, and error attribution, with CUBIST achieving state-of-the-art results; VideoQA enhances language-grounded action understanding; and Video2EMG suggests video-based alternatives to costly EMG sensing. MyoMechanix advances skilled activity understanding toward biomechanically grounded, multimodal, and compositional reasoning for Physical AI applications in fitness, rehabilitation, healthcare, and machine learning. Project page: https://haoyin116.github.io/MyoMechanix/

Vision-language-action (VLA) models provide a scalable path toward generalist robotic manipulation by integrating visual perception, language understanding, and continuous action control. However, we reveal a critical limitation of VLA architectures: the action expert has limited access to the 3D geometric and 2D semantic information available in VLM features. This accessibility gap weakens perceptual grounding and limits performance on fine-grained robotic manipulation. To address this issue, we propose V-Link, which explicitly recovers visual representations during the vision-language (VL) to action (A) feature transfer. Specifically, V-Link learns complementary Spatial and Semantic Query representations within the VLM and injects them into Action DiT through asymmetric pathways. Semantic Queries complement the original VLM image tokens, whereas Spatial Queries provide dedicated geometric conditioning for spatially grounded action generation. Across LIBERO, LIBERO-Plus, and RoboTwin 2.0, our V-Link improves the average success rate over base model GR00T N1.6 by +1.9%, +31.2%, and +18.8%, respectively. On the AGIBOT A3 Ultra, V-Link further achieves gains of +20% and +24% on two real-world humanoid tasks.

Bimanual manipulation in cluttered, contact-rich environments remains challenging because it requires coordinated motion generation, interaction-aware planning, and reliable execution under tight kinematic constraints. We present PRISM, a projection-integrated sampling-based Model Predictive Control (MPC) framework that uses a GPU-accelerated physics simulator as an online world model for complex dual-arm manipulation. The main algorithmic contribution is a QP-guided control sampling strategy that decouples trajectory exploration from kinematic feasibility. At each MPC step, sampled joint-velocity trajectories are projected onto the set of motions satisfying joint position, velocity, acceleration, and jerk bounds, together with an initial-velocity boundary condition, before rollout evaluation. This enables broad yet feasible exploration of coordinated bimanual behaviors. To support efficient online execution, we derive a custom ADMM/Bregman-splitting QP solver that exploits joint-wise separability and reusable matrix factorizations. We further use Bayesian optimization to tune task-cost weights offline, reducing manual parameter selection. We evaluate PRISM on challenging variants of PerAct$^{2}$ tasks, including obstacle-constrained ball transport, tray transport, cube handover, and box lifting. Experiments show improved robustness and task success relative to representative sampling-based baselines, while maintaining real-time or near-real-time execution. We also demonstrate successful sim-to-real transfer on dual UR5e manipulators, highlighting the practical potential of physics-based online planning for contact-rich bimanual manipulation. Project details, including code and supplementary videos, are available at \href{https://sites.google.com/view/prismbimanual}{https://sites.google.com/view/prismbimanual}.

A common strategy for scaling world models is to train on more crawled video with more compute. We argue that this strategy is inefficient: scaling world models also requires a recursive data engine that offers grounded reward signals. The success of code agents illustrates why this matters. As code is executable, compilers and runtimes can provide high-quality rewards for Reinforcement Learning (RL) post-training of LLMs. By contrast, spatial generation still relies largely on fuzzy proxies such as CLIP scores. These signals are fuzzy and biased, making them hard to support RL post-training. Compared with these, game development provides a missing reward environment for spatial world models. A scene encoded by a game engine is an executable world specification: the engine can efficiently check collision, physics, navigability and bounded playability, while the developer provides the global verification signal by judging whether the scene should be accepted. Game development also provides real-world long-horizon trajectory data for RL post-training. We therefore propose Reinforcement Learning with Human-Engine Verification (RLHEV), a post-training paradigm that combines dense engine signals with implicit human acceptance feedback from the development process.