多模态生成与编辑

Latent world models are trained to predict what happens next, so nothing in their objective separates what an action caused from what merely co-occurred with it. Object-masking models such as C-JEPA intervene on what the predictor can see; we intervene on what physically happens. From one saved simulator state we run the dynamics under an action $a$ and under a reference action $a_{\varnothing}$, and train the model to predict the difference $Δz=z^{a}-z^{a_{\varnothing}}$ between the two latent futures. The resulting objective, Do-JEPA, has an effect loss, a support loss (where the action enters), a propagation loss (where its effect travels) and invariance losses (what must not change). In a synthetic system with object-aligned variables, support supervision finds the directly intervened object in 99.95% of test cases, where a sparse action mask sends the action to a nuisance slot in every case, and response-onset supervision recovers the ring-shaped propagation graph (edge AUROC 0.975 vs. 0.624). From pixels, the effect loss beats a control trained on exactly the same data: it lowers latent effect error by 28.4% on an end-to-end LeWM model and physical effect error by 13.5% when trained and tested on natural action sequences, and on three independently generated CausalWorld benchmarks it lowers responsive effect error by about 20% under physics shifts and the latent context sensitivity of predicted effects by 66%. Trained from scratch it costs factual accuracy; fine-tuning an existing model with it removes this cost. Together, these results show that intervening on the world, rather than on what the model sees, helps latent world models predict what their actions cause.

Pretrained visual representations support image generation, but may not fully preserve the fine-grained details needed for faithful reconstruction. Meanwhile, intermediate encoder layers contain complementary visual details, but learning to fuse them for reconstruction can produce a latent distribution that is difficult to model. Existing fusion methods require empirical tuning of layer selection or staged optimization of fusion and decoding, increasing configuration effort or training complexity. We introduce HiRAE (Hierarchical Representation Autoencoder), which learns a hierarchical fusion framework over the full encoder hierarchy to improve reconstruction fidelity while maintaining compatibility with generative modeling. HiRAE groups encoder layers by depth and learns residual corrections to the deepest representation. Group-wise norm caps bound these corrections relative to the deep anchor, with tighter budgets for shallower groups. Our HiRAE-24 preserves the latent token count and channel dimension. On ImageNet-256, HiRAE-24 reduces reconstruction FID from 0.299 to 0.209 relative to RAEv2 while maintaining competitive guided generation quality. For text-to-image generation, HiRAE-24 improves alignment over RAEv2 on GenEval, DPG-Bench, and GenAI-Bench both before and after supervised fine-tuning. Under the same generator-training and evaluation protocol, post-fine-tuning GenEval increases from 84.86 to 87.70.

Interactive panoramic video generation aims to synthesize immersive 360\textdegree{} videos that remain visually coherent while following user-specified camera trajectories during exploration. However, progress is limited by a coupled data-and-model gap: existing panoramic video datasets are often short, weakly annotated, or lack camera trajectories, while existing camera-controlled video generation models are designed for perspective videos and do not directly support panoramic geometry. In this paper, we introduce MUGEN and Wan360 to address these limitations. MUGEN is a large-scale real-world panoramic video dataset tailored to interactive 360-degree world exploration, comprising over 1,300 hours of at least 4K panoramic videos with rich semantic and geometric annotations. Built on MUGEN, we further present Wan360, a camera-controllable interactive panoramic video generation model. Panoramic videos are commonly represented by EquiRectangular Projection (ERP), which unfolds a spherical 360-degree view into a rectangular frame with cyclic longitude seams and pole distortions. To this end, Wan360 introduces three parameter-free ERP-aware components: periodic longitude RoPE for seam-consistent positional encoding, ERP-aware padding for reducing boundary artifacts, and random roll yaw for consistent learning. For camera control, Wan360 uses a panoramic Plücker embedding that represents camera motion with ERP rays rather than perspective pinhole rays. Experiments show that MUGEN serves as a data foundation for panoramic world exploration, and that Wan360 enables high-quality, temporally coherent, camera-controllable 360-degree video generation.

Adding or removing a sounding object requires coordinated changes to visual content and sound while preserving the surrounding scene. Yet paired supervision for localized non-speech audiovisual editing remains limited, as visual and acoustic edits must target the same object and isolate its sound from overlapping sources. To address this gap, we introduce AVIOBench, a dataset comprising 37.9 hours of paired audiovisual examples spanning 1{,}878 target-object names. AVIOBench links the visual presence and acoustic contribution of each target object through a shared identity and visual mask. Our automated pipeline uses visual grounding and cross-modal consistency to select target-sound removal candidates, then jointly refines the audiovisual pairs to improve perceptual quality and cross-modal consistency. Building on this dataset, we propose AVIO, which adapts a pretrained text-to audiovisual generation model through source-conditioned feature modulation to jointly learn object addition and removal. A reference-frame curriculum gradually reduces reference conditioning during training, enabling one model to perform instruction-only editing with optional visual guidance. Quantitative and qualitative evaluations demonstrate effective audiovisual object removal and addition, with optional reference guidance providing appearance and placement control for addition.

Interpretability methods such as probes, activation patches and learned editors are designed to reveal or modify a model's current computation. World models pose a harder requirement: because their predictions become inputs to later predictions, a useful internal correction must survive after editing stops. We therefore propose RolloutFaith, a framework that measures semantic improvement both in the prediction produced at intervention time and over later autonomous predictions under fixed events, actions, noise, and information budgets. We evaluate ten fitted editors on three world models across Crafter, Cartpole, and CoinRun. We also use Reference Activation Patching, which replaces a model activation with the paired activation computed from the real observation, to measure the correction available at the chosen interface. This reference intervention improves later predictions in all nine model and task combinations and outperforms the best fitted editor in eight, yet its sustained gain decreases with horizon in five of nine combinations. Current fitted editors recover only limited and inconsistent long term effects. By restoring individual state components to their untouched values, we find that persistent effects travel through the newest generated frame in DIAMOND, recurrent memory in DreamerV3, and both in STORM. These findings suggest that training should reward future consequences. To test this hypothesis, we propose Delayed LoReFT, which optimizes the same low rank intervention through four frozen future transitions and improves sustained intervention effects to some extent.

Autoregressive (AR) video diffusion enables low-latency, streamable video generation, but prediction errors often accumulate over long rollouts. Training the generator on its own rollouts exposes it to these imperfect histories. However, existing video-level distribution matching distillation (DMD) scores the whole rollout jointly. Because a chunk is evaluated together with its past and future, its correction can favor matching artifacts in the surrounding context merely to preserve temporal consistency. To provide a clearer visual-quality signal, we introduce Rollout-Marginal Distillation (RMD). RMD retains the generated history for AR prediction but scores each chunk independently against a chunk teacher, ensuring its quality correction is not compromised by an imperfect temporal context. To compensate for the lack of temporal context in independent chunk scoring, RMD subsequently applies video-level DMD to restore temporal coherence. Extensive experiments demonstrate that RMD maintains high visual quality far beyond its training horizon and outperforms video-level DMD baselines. Code and video results are available at https://cjeen.github.io/RMD

World simulation is inherently multisensory, demanding synchronized visual and acoustic dynamics in real time. Yet prevailing interactive world models remain strictly silent, focusing exclusively on visual rendering and control while overlooking the acoustic dimension. We present HelixWorld, a real-time interactive audio-visual world model where visual scenes and camera-grounded spatial stereo sound co-evolve natively under user interaction. We curate a high-fidelity spatial audio-visual dataset with true stereo acoustics and metric camera poses, upon which we pre-train a bidirectional teacher conditioned on 6-DoF camera trajectories and user actions. To enable low-latency causal interaction, we distill the teacher into a few-step streaming student via an online trajectory distillation loss, sustaining drift-free joint audio-visual rollouts at 24 FPS on a single GPU. Furthermore, we formalize spatial-acoustic consistency and introduce HelixBench to evaluate whether synthesized sound fields faithfully track dynamic viewpoint motion. Extensive experiments demonstrate that HelixWorld matches state-of-the-art silent world models in visual fidelity and responsiveness, while significantly surpassing existing baselines in camera-aligned spatial-acoustic immersion.

Co-speech gesture generation aims to synthesize natural gestures that are both temporally synchronized with speech and semantically consistent with the spoken content. Although recent methods can generate rhythmically plausible motions, they often rely heavily on acoustic prosody while underutilizing textual semantics, especially when semantic annotations are incomplete, noisy, or unavailable. Consequently, the generated gestures may follow speech rhythm while failing to express the intended semantics. To address this problem, we propose a reliability-aware semantic-rhythm control framework for co-speech gesture generation. We first learn a discrete motion prior that represents continuous gestures in a compact and structured motion-code space. We then introduce a dual-branch semantic contribution estimation mechanism consisting of a full multimodal branch and an audio-only branch. Their distributional discrepancy is formulated as conditional information gain to quantify how much textual semantics changes the predicted motion. Based on this estimate, a controllable semantic-rhythm objective selectively strengthens semantic guidance in content-relevant segments while limiting unnecessary semantic intervention in rhythm-dominant segments. Furthermore, we treat background noise as an acoustic reliability condition and introduce noise-conditioned feature modulation together with beneficial latent perturbation to improve generation robustness under realistic acoustic environments. Experiments on benchmark datasets demonstrate that the proposed framework achieves a favorable balance among semantic expressiveness, rhythmic synchronization, motion diversity, and robustness, enabling reliable and controllable co-speech gesture generation.

Modern image editors excel at semantic manipulation and visual synthesis, yet remain limited in precise spatial control, motivating the development of drag-based editing. However, existing drag-based methods often struggle to balance drag accuracy with natural, plausible, and intent-aligned generation. We propose MoRe-Drag, a motion-grounded drag-based editing method. Our key insight is to treat pixel-space warping as coarse motion evidence, and to inject this evidence into the generative sampling trajectory. Specifically, MoRe-Drag performs region-aware latent recomposition over refinement, inpainting, and anchor regions, coupled with stage-adaptive conditioning that progressively shifts from motion-grounded structure formation to semantic refinement. We further support an instruction-free interface by adapting the MLLM-based text encoder for drag-aware instruction inference. Experiments on DragBench-SR and DragBench-DR show that MoRe-Drag substantially improves drag precision over strong base editors and achieves superior drag accuracy among SOTA drag-based methods, while delivering strong semantic consistency and visually realistic results. Code and dataset will be publicly released.

World-Action Models (WAMs) couple action generation with predictions of how physical interactions unfold. However, current post-deployment learning paradigms typically improve behavior without requiring better world predictions. Especially in dexterous manipulation, small execution errors can compound in high-dimensional action spaces, hindering policy improvement and pushing interactions beyond the world model's training distribution. Motivated by this, we propose Direct Experience World-Model Optimization (DEWO), a post-deployment learning paradigm for WAMs that, alongside action imitation, refines world representations through visual experience to better condition action generation. Specifically, it identifies interaction turning points and learns from successful and failed futures to support classifier-free guidance. An additional value head estimates task progress from video representations and activates guidance when progress stalls during inference. Across five DexJoCo tasks, DEWO improves average success across all three WAM formulations. Ablations show that visual supervision from successful and failed continuations improves both prediction and control beyond action supervision alone. On four real-world tasks across Wuji and Sharpa, 3 x 3 grid evaluations show that two rounds of deployment learning increase success from 51.0% to 71.7% in cells with at least one initial success, a gain of 20.7 percentage points. These findings support continued predictive learning for improving control through deployment experience, making world modeling an active part of WAM adaptation.

We introduce FracGen, a fracture-aware video generation model that produces plausible, controllable fracture dynamics from a single image of an intact object, conditioned on physics signals. To train FracGen, we build FracSim, a fracture-aware simulation framework that augments material point method (MPM) simulation with a continuum damage model, producing paired fracture videos and dense, pixel-aligned physical fields at no additional cost beyond standard rendering. FracGen leverages these maps in two ways: it is trained to jointly predict them alongside RGB video, encouraging the model to capture physical state rather than surface appearance; and it is supervised with physics-informed losses that encourage consistency among the predicted maps. As a result, FracGen captures distinct material-specific fracture behavior without expensive test-time simulation or per-scene tuning, while offering fine-grained control over where an object tears, how fast the crack propagates, and how much deformation precedes failure. We further introduce a benchmark for evaluating the physical plausibility of generated fracture video, and show through extensive experiments that FracGen outperforms existing video generation baselines in both physical and visual fidelity. Results are best viewed in our project website: https://fracgen.github.io/.

With the rapid advancement of text-to-image (T2I) generation, robust evaluation becomes critical yet challenging, as traditional metrics fail to capture fine-grained alignment and generative artifacts. While large multimodal models (LMMs) are increasingly adopted as evaluators, existing benchmarks typically study semantic understanding, quality perception, and authenticity identification in isolation, while largely neglecting responsibility detection. This leaves a gap in unified and comprehensive validation. To bridge this gap, we introduce SQUARE-Bench, a comprehensive benchmark that systematically evaluates LMM capabilities as evaluators of AI-generated images across four aspects: Semantics, Quality, Authenticity, and Responsibility. SQUARE-Bench introduces a granular taxonomy of 38 sub-dimensions to evaluate nearly 10K AI-generated images sampled from 22 diverse models, ranging from legacy to state-of-the-art generators, complemented by over 3K real-world images. The images are annotated with curated question-answering pairs. Extensive experiments on 23 LMMs reveal that top proprietary models, such as Gemini-3-Pro, already outperform the individual human expert baseline. However, the performance gap between models remains significant, exhibiting notable disparities in fine-grained inference and domain-specific robustness. Beyond benchmarking, we conduct a proof-of-concept study of LMM-guided iterative editing, in which dimension-specific LMMs provide diagnostic feedback to fixed image editors. The resulting guided system yields selective improvements in semantics, authenticity, and responsibility, while exhibiting a consistent visual-quality trade-off. SQUARE-Bench can serve as both a diagnostic tool for characterizing LMM evaluator capabilities and studying their use in T2I generation refinement. The benchmark and dataset will be released upon publication.

Video world models require persistent scene memory to maintain consistency during long-horizon video generation. Existing spatial memories accumulate RGB observations or latent features, increasing storage requirements as generation proceeds. We introduce Honeycomb, a video world model built on HexMemory, our proposed low-rank representation for storing scene features in a fixed-size memory with a total of six spatial and spatiotemporal planes. A feed-forward writer maps each generated chunk into new plane features. As the spatial coverage or temporal range expands, we warp the previous planes while preserving their dimensions, then fuse them with the new features through confidence-weighted pooling and a learned residual correction. A reader retrieves latents from HexMemory to condition subsequent video generation. The writer processes only observations from the new chunk, avoiding per-scene optimization and repeated processing of the full history. Experiments on WorldScore and RealEstate10K demonstrate strong video generation quality and robust revisit consistency while keeping HexMemory feature storage constant throughout generation. Code and additional visualizations are available on our project page at https://jackswl.github.io/honeycomb/.

World-action models (WAMs) couple future visual-state prediction with action generation. By adapting video generators or image-editing models pretrained at scale, a prominent line of recent WAMs inherits both predictive knowledge and the models in which it was learned. We ask whether a predictive visual latent space induced by large-scale predictive pretraining can instead provide a sufficient foundation for effective WAM learning without inheriting a complete pretrained visual generative model. To answer this question, we introduce V-JEPA Policy, a simple framework that builds a WAM on the latent space of a frozen V-JEPA 2.1 encoder. An instruction-conditioned future-latent predictor and a flow-matching action expert are jointly learned from scratch in a single downstream stage, with the predictor's future-informed context key--value states conditioning action generation. With 0.9B total parameters, of which 0.6B are trainable, V-JEPA Policy achieves competitive performance with representative WAM and vision-language-action baselines across LIBERO, LIBERO-Plus, and RoboCasa-GR1. Comparing visual foundations under the same downstream framework and training budget identifies V-JEPA latents as more effective than the discriminative, reconstructive, and video-understanding-oriented alternatives, particularly under distribution shifts. Beyond task-specific learning, pretraining the predictor on DROID video--instruction pairs without action labels and adapting it into a WAM yields substantial gains in downstream control and out-of-distribution generalization. Together, these findings establish predictive visual latents as a foundation for effective WAM learning from task-specific demonstrations and for transferring future-modeling knowledge acquired from broader in-the-wild videos. Our code is available at https://github.com/breez3young/VJEPA-Policy.

We present Waypoint 1.5, a real-time diffusion world model for interactive video generation on consumer-grade hardware. Unlike general video diffusion models, interactive world models (iWMs) must respond to dense user controls under strict latency and throughput constraints. Waypoint 1.5 is pre-trained on 100,000 hours of diverse, control-aligned video game data across hundreds of games, and generates playable video conditioned on full keyboard and mouse input. The model includes two resolution variants that run across a wide spectrum of consumer hardware. To characterize this unique setting, we distinguish rendered FPS, latent FPS, and control rate. We describe the data pipeline, architecture, training methodology, and runtime system behind Waypoint 1.5. We evaluate interactivity through latency and throughput. Finally, we discuss the safety and ethics considerations unique to iWMs.

Recent multimodal image generation models can take multiple images and textual instructions as input, enabling reference-based generation guided not only by text but also by visual instructions such as layouts, arrows, and pose cues. However, existing benchmarks do not evaluate the joint setting in which multiple references must be composed under multiple and heterogeneous visual-instruction images. To address this gap, we introduce VIF-Bench, a benchmark of 1,241 tasks designed to assess the edge of model capabilities in this joint setting by covering: (i) multi-reference generation (up to 7) under multiple heterogeneous visual instructions (up to 6), (ii) cases where reference images can potentially compete with visual instructions (e.g., a strongly posed subject vs. a target pose), and (iii) controlled comparison of visual instructions with text descriptions at different levels of specificity. Using these capabilities, we uncover three findings: (1) models face an adherence-artifact trade-off: once models reach stronger visual instruction adherence, stronger adherence tends to coincide with more instruction artifacts in generated images, (2) visual instruction adherence tends to be lower on tasks whose reference images carry a salient state of the controlled attribute (e.g., a neon-lit subject under a light-direction instruction), most consistently for light and wind, and (3) for models that can understand visual instructions, it is often better to provide visual constraints directly rather than describe them in text; when using text, a moderate level of detail works better than an exhaustive description. VIF-Bench is released as an open benchmark to establish a basis for fair comparison in controllable multi-reference image generation.

Human image animation aims to transfer motion from a driving video to subjects in a reference image. Despite remarkable progress in video generation, achieving high-fidelity animation of multiple interacting subjects remains a challenge. Many existing approaches rely on explicit motion representations such as 2D skeletons or parametric body meshes and struggle to preserve identity-motion binding under inter-person occlusion. To address this limitation, we propose WeLike2Party, a multi-human animation framework built on direct in-context video conditioning without explicit pose or mesh extraction at inference. We further introduce Reference Asymmetric RoPE Conditioning to preserve fine-grained appearance details, and Identity Binding Supervision to associate each reference identity with its intended motion trajectory. To support cross-identity training, we construct MotionTwin, a large-scale synthetic dataset comprising 14.4K cross-identity video pairs with shared subject and camera motions, totaling 84.3 hours of photorealistic video. We additionally present MotionTwin-Bench, a cross-identity benchmark specifically designed to evaluate subject-level visual fidelity and identity-motion binding. Extensive experiments on MotionTwin-Bench and real-world videos demonstrate that WeLike2Party outperforms recent state-of-the-art methods in subject-level visual fidelity, identity-motion binding, and overall perceptual quality, particularly in multi-person interactions with substantial occlusion.

We present World2Motion, a framework that generates scene-aware 3D human motion and corresponding video from a single image and a text prompt. While existing 3D motion generators learn from motion datasets, their generalization is constrained by limited coverage of environments. In contrast, video world models such as Cosmos 3 offer broader environmental priors but are not designed for full-body motion generation; recovering motion from their generated videos requires costly two-stage inference. To address these, we turn Cosmos 3 into a single-stage 3D motion generator. This adaptation has two challenges: the scarcity of paired video--motion data and temporal instability in the generated motion. First, we construct a training dataset combining synthetic video--motion pairs with real videos paired with estimated 3D motion. Second, we propose a shift-decoupled noise schedule that assigns different noise levels to video and motion through shared denoising progress. This design accommodates the different denoising requirements of the two modalities, reducing motion jitter. Experiments on a multi-source interaction benchmark show that World2Motion has better motion--text alignment and scene interaction compared with the evaluated 3D motion generators. It also matches the interaction success rate of the two-stage baseline while achieving approximately 3.3$\times$ faster inference. Our project page is available at https://fyantu.github.io/World2Motion/.

A video can exhibit convincing motion and photorealism yet fail immediately when visual text collapses. Unlike generic scene content, visual text is unforgiving in video generation: minor stroke corruption, temporal instability, or editing errors instantly break legibility and realism. Existing benchmarks overlook this challenge by treating text as incidental or using static OCR metrics that ignore temporal dynamics. We introduce VidScribe, a unified diagnostic benchmark spanning four generation regimes: writing from language (T2V), transferring text identity from a reference (R2V), sustaining text under dynamics (I2V), and localized text editing (V2V). VidScribe contains 803 human-verified samples across a 12-axis conditionally orthogonal factor space covering Intrinsic Text Properties, Physical Imaging Conditions, and Temporal Behavior. For reliable evaluation, we build a track-grounded, gated suite with 11 shared metrics and 2 task-specific probes under strict measurability conditions. Benchmarking 11 commercial and open-source systems shows that video text capability is non-monolithic, with content recognition decoupled from stroke-level glyph correctness. Performance is highly task-asymmetric: I2V sustains text most reliably, whereas V2V editing is the primary bottleneck. Counter-intuitively, degradation concentrates on a small subset of text-centric structural and temporal factors rather than adverse imaging conditions. Further probes show that visual references improve glyph and typographic fidelity rather than content accuracy, while localized editing fails to isolate target text without corrupting undeclared source text. Beyond evaluation, VidScribe also provides an actionable training signal, where benchmark-aligned preference optimization measurably improves visual text generation. https://huggingface.co/datasets/Vicky0720/VidScribe.

We introduce LIFT, a unified image-to-video generation framework that complements camera control with Layout-In-FuTure control, enabling users to specify what should appear in a future view and where it should appear. This addresses a practical need in controllable video generation: given an initial image, users often care not only about how the camera moves, but also about what the scene should look like at key future moments, especially the final frame. Existing camera controls specify viewpoint trajectories, while text prompts provide only coarse semantic guidance; neither precisely determines the content and spatial layout of future views. This limitation becomes particularly pronounced under large viewpoint changes, where the camera reveals regions that are not visible in the first frame. LIFT therefore uses the last-frame layout as an explicit control signal for the desired future scene. Since learning from such sparse layout guidance is substantially more challenging than conditioning on dense per-frame layouts, we introduce on-policy self-distillation (OPSD) to transfer the control capability of a dense-layout teacher to a last-frame-layout student. We further curate LIFT-Vista, a dataset featuring large viewpoint changes with camera and temporally consistent layout annotations. Experiments show that LIFT improves video quality, future-layout controllability, and camera controllability over other methods.

Video world models aim to predict future content from an observed scene while following prescribed camera motion. Real-world scene evolution is determined not only by changes in viewpoint and object dynamics, but also by environmental conditions such as weather, which can substantially alter scene appearance and visibility. Modeling such realistic weather evolution is challenging because the required weather modification depends jointly on the observed and desired weather states. Depending on their relation, the model may need to preserve, introduce, or remove a weather effect. Existing video world models typically leave this weather transition implicit, forcing the generation backbone to infer weather evolution together with scene dynamics and camera motion, which leads to imprecise weather control. To address this limitation, we propose MeteoVerse, a unified weather-controllable video world model that generates future videos from a single sunny or adverse-weather image, conditioned on a weather-free scene description, a target-weather instruction, and a camera trajectory. Rather than conditioning only on the desired weather, MeteoVerse explicitly estimates the observed and target weather states and represents the required weather transition. A transition-aware mixture of weather experts then translates this transition into category-specific residual weather features, unifying weather preservation, introduction, and removal while enabling fine-grained control over introduced weather intensity. We further construct the MeteoVerse dataset with over 50K real-world weather video clips, generated sunny counterparts, disentangled scene and weather descriptions, weather-intensity annotations, and camera trajectories. Extensive experiments demonstrate substantially improved weather controllability while retaining competitive scene consistency and camera-control performance.

Video world models are largely regarded as predictive models of the physical world and are therefore expected to anticipate the consequences of observed events. However, evaluation has mainly focused on reference similarity, physical-law consistency, or judgment plausibility, estimating anticipation only indirectly. We address this directly: when a release or impact has just occurred but its consequence is withheld, can a world model anticipate what should happen next? We introduce an event-anchored evaluation based on 62 controlled real-world free-fall recordings and 124 clips spanning three object types, with fine-grained release and impact annotations and ground-truth trajectories. The protocol separates consequence production, temporal placement, and physical realization. Across six contemporary video generation and world models, Runway and Veo produce release and subsequent impact events at rates above 93% but often initiate them substantially late, whereas Cosmos-Predict-2.5 and MAGI-1 frequently preserve the pre-event state and produce little or no measurable consequence. Among measurable falls, plausible timing does not necessarily imply physically consistent motion. We further conduct a 15-participant, 20-condition human study in which participants describe the expected consequence from a single event-anchored frame and draw its trajectory. Human predictions favor the recorded future in aggregate while revealing genuine ambiguity among plausible continuations. Overall, physical foresight emerges as a sequence of distinct challenges: initiating a consequence, anchoring it in time, and realizing its motion.

Pretrained text-to-image models contain broad visual knowledge, yet they cannot reliably acquire or refine a specific visual identity from only a few references while preserving compositional control. Token-embedding methods are compact but often underfit identity, whereas adapter-based methods improve fidelity through persistent weight updates that can be costly to store and interfere when concepts are composed. We introduce V-Engram, a trigger-indexed external memory mechanism for Stable Diffusion 3.5. Each concept is assigned an explicit trigger that retrieves concept-specific memory, whose gated directions enter frozen text-encoder and MMDiT context states as relative residuals. Separating this memory from backbone adaptation enables prompt-selective and multi-concept access without merging model updates. Experiments show that V-Engram broadly matches DreamBooth-LoRA in overall subject fidelity while showing advantages in settings such as contextual subject preservation. Prompt-matched loading retrieves only matched entries, reducing most additional adaptation-state loading for a single-concept query. Qualitative results further demonstrate paired-trigger composition and same-class separation, while prompts without registered entries retain the frozen model's base behavior. Together, these results establish trigger-indexed memory as a modular interface for adding targeted visual evidence without rewriting the generator.

Video diffusion models are increasingly developed into specialized models for diverse downstream tasks, and this development often includes a distillation stage, for example to accelerate sampling or to improve long-video generation. This stage is typically repeated for every specialized model. We introduce LongLive-Plug, a once-for-all distillation framework that learns reusable capabilities as LoRAs on a base model for training-free, plug-and-play deployment to compatible downstream models. These capabilities include single-pass classifier-free guidance, few-step sampling, and long-context error correction for autoregressive generation. The adapters remain reusable even when downstream models add conditioning branches, expand output channels. Despite training at a fixed guidance scale, our dedicated CFG LoRA provides text guidance control through its inference weight. Combining it with a few-step LoRA simultaneously preserves few-step generation and CFG controllability on downstream tasks. We verify training-free deployment on 54 downstream models across three backbone families and eight task categories, including world modeling, robotics, editing, and multimodal generation. The approach may support additional compatible models. Each capability can thus be distilled once per backbone family and reused without per-target retraining.

Few-step streaming audio--video generation requires both causal modeling and step distillation, yet standard training recipes face two context-related challenges. Teacher forcing pairs clean history with a noisy target, but supervises predictive contextual representations only indirectly through velocity prediction. Meanwhile, directly reusing bidirectional score models in causal Distribution Matching Distillation (DMD) creates a mismatch between generation and scoring contexts. We address these challenges with Salt++, a two-stage post-training framework comprising Causal Self-Flow (CSF) and context-aligned autoregressive DMD. CSF exploits contextual information asymmetry by varying the history while keeping the noisy target fixed: a noise-mixed-history student aligns its intermediate representations with those of a clean-history exponential-moving-average teacher. This self-supervised signal encourages the student to extract semantic information and improves cross-modal alignment. Context-aligned AR DMD shares the causal mask and prefix across generator sampling, fake-score training, and real-score evaluation to match generated and reference distributions under a block-conditional KL objective. With calibrated teacher guidance, it performs clean-prefix few-step distillation and then adapts to generated histories without switching objectives or requiring separate consistency distillation. At 480p, Salt++ improves visual and motion quality by 57% and 45% over OmniForcing on JavisBench under the same 4-step causal setting. A separate scale-wise post-training stage extends Salt++ to 4-step $1664\times960$ generation, outperforming bidirectional LTX-2 on six of seven reported metrics. Project page: https://xingtongge.github.io/Saltpp

Quantum Generative Adversarial Networks (QGANs) have emerged as representative generative models in the Noisy Intermediate-Scale Quantum (NISQ) era and have attracted increasing attention in quantum machine learning. However, most existing QGAN methods rely on patch-based decomposition strategies, which weaken the global consistency of generated images and increase quantum resource overhead. In this work, we investigate a simpler approach: pixel-level, end-to-end image generation using a single-quantum-circuit QGAN. By analyzing the structural matching relationship between the quantum prior and the target data distribution in Hilbert space, we provide a new theoretical perspective for understanding the training behavior of naive end-to-end QGANs. Specifically, we introduce the Quantum Fidelity Landscape (QFL), defined as the pairwise-fidelity structure induced by an ensemble of quantum states and preserved under shared unitary transformations of the quantum generation process. We show that, under a fixed Lipschitz readout, this invariant imposes a one-sided bound on decoded sample separation, motivating calibration of the prior-induced QFL before adversarial training. To validate this theoretical insight, we propose BasicQGAN, a QGAN framework incorporating quantum prior calibration. Before adversarial optimization, BasicQGAN aligns the prior-induced QFL with the data-induced QFL. Experimental results on small-scale grayscale image datasets show that BasicQGAN achieves stable and effective end-to-end pixel-level image generation while requiring fewer qubits and trainable parameters than representative patch-based quantum generators. Furthermore, experiments with different initial quantum-state ensembles show that QFL-calibrated ensembles achieve better generative performance.

Text-to-video (T2V) diffusion models can generate realistic depictions of actions such as kicking, stabbing, and shooting, raising safety concerns that motivate targeted concept erasure. Although concept erasure has been extensively studied for static concepts in text-to-image and T2V models, erasing motion concepts remains largely unexplored. We present a systematic study of motion concept erasure in video Diffusion Transformers (DiTs). Through causal interventions, we show that text-conditioning attention carries concept-specific motion information and supports selective intervention, whereas perturbing temporal positional encoding suppresses both target and non-target dynamics. We further find that directly adapting ESD, a representative weight-level image erasure method, to a video DiT yields modest and uneven motion suppression: reducing its erasure training loss does not by itself remove the concept signal from the difference between the conditional and unconditional predictions, which classifier-free guidance (CFG) then scales at every denoising step. From these findings, we derive three requirements for motion concept erasure: concept specificity, spatial selectivity, and temporal naturalness. Each determines one component of MUTE (Motion concept Unlearning in Text-to-video gEneration): at each denoising step, MUTE extracts a concept direction through token neutralization, derives a spatial gate from the direction's intrinsic structure, and subtracts the resulting correction from the velocity output before CFG is applied. MUTE is training-free and requires no weight modification. Experiments on 20 motion concepts show that MUTE outperforms representative prompt-level, weight-level, and inference-time baselines on Wan2.1-T2V, and the same formulation transfers to CogVideoX, supporting its applicability across distinct T2V attention architectures.

Latent Diffusion Models (LDMs) typically adopt a two-stage pipeline: an auto-encoder (AE) is first pre-trained to define a latent space, then a diffusion model is trained to perform denoising within it. Such a two-stage design introduces a representation mismatch, as the latent space is optimized for reconstruction rather than adapting the denoising dynamics. We reveal that the LDM itself is an AE, and consequently present LDM-is-AE, an end-to-end one-stage LDM training framework that eliminates the need for a separately trained tokenizer. Our key observation is that the LDM backbone actually performs a latent-to-feature-to-latent transformation at each denoising step, which can be interpreted as an internal decoding--encoding process. Leveraging this structure, we split the DiT backbone into two reciprocal components, DiT-E (i.e., DiT Encoding) and DiT-D (i.e., DiT Decoding), and impose image-space supervision on the intermediate features across all timesteps. Our model encourages the internal representation to align with the image domain throughout denoising, thereby establishing an explicit latent-to-image-to-latent path. At the zero-noise timestep, our model further performs an image-to-latent-to-image mapping, corresponding to an auto-encoding process. As a result, LDM-is-AE jointly learns latent representations and denoising dynamics in an end-to-end manner, yielding a diffusion-native latent space tailored to the generation process. Experiments demonstrate that LDM-is-AE exhibits highly competitive generation performance, achieving an FID of 1.80 and 1.90 on 256x256 and 512x512 class-conditional image generation, respectively.

We present a method for generating high quality materials for 3D objects entirely in texture space. We finetune a video diffusion transformer for text-guided material generation, multi-view material generation, and material upscaling. Our key insight is to use the known projection from image space to texture space, enabling the diffusion process to generalize across arbitrary geometries and texture parameterizations. This approach also avoids the view consistency issues inherent in video and multi-view diffusion models. Because texture space is two dimensional, we can reuse the strong priors of pretrained video diffusion models. We apply our method to high quality material reconstruction from posed photos captured under unknown lighting, as well as to text- and image guided material generation. Our method can scale to high resolutions (8K), 100+ input views, and neural material representations. In quantitative and qualitative evaluations we show state-of-the-art results for material generation and reconstruction.

Text-to-image users often provide concise and underspecified prompts, whereas generative models benefit from detailed textual conditions for reliable instruction following. Existing systems bridge this gap with Prompt Enhancers (PEs) that rewrite raw prompts at inference time, introducing additional latency and leaving prompt elaboration external to the generator. We instead view enhanced prompts as privileged training information and ask whether their benefits can be internalized. We propose Prompt-Enhanced On-Policy Self-Distillation (PE-OPSD) for text-to-image flow-matching models. During training, a raw-prompt student follows its own generation trajectory, while an enhanced-prompt teacher provides vector-field targets at the states visited by the student. This on-policy supervision distills the behavior induced by enhanced prompts into the raw-prompt student without requiring additional text--image pairs. At inference, both the PE and teacher are removed, and the student generates directly from raw prompts. Across multiple model families, PEs, and benchmarks, PE-OPSD achieves the strongest aggregate prompt fidelity among the evaluated baselines, yields positive aggregate visual appeal gains, and retains the base-model inference efficiency.