多模态生成与编辑

Action-conditioned video world models predict future observations from an initial observation and an action signal. In robotics, actions influence future observations through two distinct processes: they are first realized into robot motion by the robot body and controller, and the scene then responds through contact and object motion. Conditioning directly on action commands asks the world model to learn the realization process itself, while conditioning on logged future states leaks the interaction outcomes it is meant to predict. We propose robot-factored world models, which move two robot-specific factors outside the world model. First, action realization: each command is rolled through the robot's own controller and kinematics into a deployment-available nominal trajectory, a middle signal that avoids both action-realization learning and future-state leakage. Second, robot rendering: this nominal trajectory is rendered through the robot URDF, factoring the robot's geometry, kinematics, and appearance out of the model and into explicit rendered robot geometry. To resolve depth ambiguity, we pair end-effector depth with scene depth, giving geometric cues for contact and occlusion beyond image-plane overlap. Together, camera-aware static RGB/depth context and rendered robot geometry form a shared visual world-model interface that stays consistent across viewpoints and robot embodiments, so the model sees the action only as visible robot geometry and learns how objects respond to it. Our experiments show that the rendered interface outperforms vector-conditioned baselines and generalizes to unseen robot embodiments at inference. We further demonstrate that our model generates robot manipulation videos from human demonstrations by retargeting and rendering the hand motion as robot geometry.

Recent advances in video diffusion models have spurred interest in human-object interaction (HOI) video generation, which demands fine-grained control over interaction logic beyond single-subject animation. However, existing HOI methods rely heavily on explicit motion control, limiting scalability and generalization across diverse objects and interactions. In this study, we propose AgentHOI, a text-driven HOI video generation following a thinking-before-generation framework that bridges the gap between high-level textual intent and physical execution through multi-agent reasoning over perception, interaction, and motion planning. Building upon the generated interaction plans, we further strengthen text-driven motion understanding. We introduce an implicit text-motion alignment strategy that distills text-to-motion priors into the video diffusion model, enabling robust HOI synthesis without explicit motion inputs at inference. Experiments show that AgentHOI significantly improves interaction naturalness, object appearance preservation, and adherence to complex textual instructions across challenging object-centric scenarios such as wearing and riding. The code is available at https://github.com/bone-11/agenthoi.

Talking-face (TF) deepfake generation synthesizes photore- alistic facial video from a static source image and an au- dio signal, producing forgeries that current image-based detectors consistently fail to identify. Unlike face-swap ma- nipulation, TF synthesis has no underlying real video from which to inherit physiological characteristics, making re- mote photoplethysmography (rPPG) a uniquely motivated detection modality for this forgery category. We propose a detection framework that extracts per-video rPPG wave- forms via RhythmFormer and trains a suite of lightweight classifiers to distinguish real from synthesized physiologi- cal signals. Evaluated on the TF subset of Celeb-DF++ un- der a strict subject-independent protocol, where test identi- ties are completely separated from training identities, our 1D ResNet achieves an AUC of 0.806 and EER of 27.8%, placing it within 2.4 points of the best published general- purpose detector (Effort, ICML 2025) while operating ex- clusively on the physiological channel. We document a con- trolled reproduction study of DeepFakesON-Phys, the rep- resentative prior rPPG detector, demonstrating degrada- tion from AUC 0.999 on legacy face-swap data to 0.622 on the TF subset of Celeb-DF++. We further show that detec- tion difficulty is strongly method-dependent: AUC ranges from 0.985 (Real3DPortrait) to 0.690 (IP-LAP) across the seven TF generators, with the ranking remaining perfectly stable across all evaluation protocols. This spread reflects an interpretable physiological property of each generator rather than evaluation noise, and constitutes the primary theoretical contribution of the work.

We present Oxygen-TryOn, a unified foundation model for any-item virtual try-on. Rather than repurposing a general-purpose image editor, Oxygen-TryOn is fashion-native, built for try-on through a dedicated data engine and try-on-specific training. Given one or more reference items (clean product shots or in-the-wild worn-on photos) and a single target subject image, it synthesizes a photorealistic image of the subject wearing the items across virtually any fashion category. Prior systems handle a single garment category in a studio setting, and recent multi-reference methods remain garment-centric; in contrast, Oxygen-TryOn supports diverse items and scenarios, including full- and half-body views, a variable number of references, and free multi-item composition, while faithfully preserving both subject identity and item appearance. Instead of mask-based inpainting, we reformulate try-on as a multi-reference, understanding-driven generation task. We build a data engine that collects, manufactures, annotates, and filters high-quality try-on data at scale, and design a three-stage recipe of continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning (RL). The RL stage uses a hybrid reward combining an in-house try-on reward model with a proprietary, rubric-guided general-purpose model, jointly supervising fine-grained consistency and instruction-level quality. It also follows general editing instructions (e.g., pose changes) in the same pass. Across public benchmarks and our in-house Oxygen-TryOn Bench, it achieves state-of-the-art consistency and realism on single-item try-on and leads on multi-item try-on, matching or surpassing both leading proprietary systems (Nano Banana Pro, GPT-Image-2, Seedream5 Lite) and open-source models (FLUX.2).

Generative world models are increasingly driven as simulators: a planner forks a state, rolls out futures, backtracks, and returns to a visited viewpoint. Recent benchmarks establish that current video world models fail this usage, and attribute it to the model, prescribing new architectures and training objectives. We show this attribution is incomplete, and for an important class of models simply wrong. Snapshotting the state the runtime already holds — an observation plus RNG state, a memory bank, or a windowed KV context, by architecture — and restoring it after a genuine excursion reproduces the never-left continuation byte-identically on all three; corrupting only the RNG degrades it. The capability was never missing: request-centric serving discarded it, inheriting from language-model serving the assumption that runtime state is recomputable — but world-model state carries a non-recomputable kernel. We define Persistent Computational State (PCS), the minimal non-recomputable state that must survive across requests, show it can be discovered by measurement, and build a session-centric runtime over it. Checkpoint and restore cost 0.012 ms against a 1.85 s generation step; resident sessions become host- rather than device-bounded (measured to 1,024); and world memory must be evicted by relevance to the return, not recency — the inverse of LLM practice.

Integrating heterogeneous biomedical data, including clinical metadata, histopathology images, and molecular profiles, is crucial for comprehensive disease understanding. However, gene expression data acquisition remains constrained by high costs and privacy concerns, limiting its use in multimodal research and AI-driven applications. We present MultiModal Molecular Generation (M$^3$-Gen), a novel framework for the generation of gene expression profiles by conditioning a Generative Adversarial Network on histopathology images and clinical metadata. M$^3$-Gen learns a unified latent representation from the clinical variables and the images, leveraging contrastive learning, and exploits the embeddings of the two modalities to guide a generative model in producing biologically coherent gene expression profiles. Evaluations on the TCGA dataset demonstrate that M$^3$-Gen generates realistic and functionally meaningful gene expression data. Importantly, by integrating multiple modalities in an attention-based mechanism, M$^3$-Gen provides intrinsic explainability: it allows the identification of which regions of the histopathology images most strongly influenced the generation of specific gene expression profiles, making the model's decisions interpretable by design.

Structured understanding of satellite video is essential for advancing dynamic geospatial scene analysis from low-level perception to high-level cognition. To move beyond object-centric perception, this paper introduces spatio-temporal panoptic scene graph generation (TPSG) in satellite video as a new benchmark task. TPSG aims to generate a structured graph composed of a set of triplets <subject, relationship, object> with explicit temporal spans, thereby describing dynamic geospatial scenes by jointly modeling identity-consistent instance masks and spatio-temporal relationships among panoptic scene elements. However, there is still no dedicated dataset for TPSG in satellite video. Moreover, TPSG in satellite video is intrinsically challenging, as objects are often small and weakly textured, cross-frame association is easily disrupted by occlusion and background clutter, and relationship semantics are highly coupled with spatial structure and temporal evolution. Consequently, TPSG models developed for natural videos are not directly applicable to satellite video. This paper presents T-STAR, a large-scale benchmark dataset for TPSG in satellite video, comprising over 1.1 million instance masks and over 3.8 million spatio-temporal triplets across 39 fine-grained object categories and 70 fine-grained relationship categories. To enable TPSG in satellite video, we propose a unified framework to enhance cross-frame instance consistency and spatio-temporal relationship prediction. Extensive experiments demonstrate the significance of T-STAR and the effectiveness of the proposed framework, establishing a strong benchmark for future research on structured satellite video understanding. The dataset and code are available at https://github.com/linlin-dev/T-STAR.

Vision-Language-Action (VLA) models augmented with world modeling represent a promising paradigm for end-to-end autonomous driving. While pixel-level future prediction enables fine-grained spatiotemporal reasoning, it compromises robustness in noisy driving scenarios. Conversely, latent-based world models alleviate this sensitivity but often incur limited interpretability and representational degradation due to absent pixel-level grounding. To reconcile this trade-off, we propose HyWorldVLA, a hybrid world-VLA framework that unifies pixel-level supervision and latent representation learning. In the pre-training stage, HyWorldVLA predicts video latents encoded by a pre-trained video VAE, while simultaneously reconstructing video frames to provide precise pixel-level grounding. During the subsequent co-fine-tuning phase, the model exclusively predicts latent features, which are fed into an action expert to generate trajectories. Extensive experiments on NAVSIM v1 and v2 benchmarks demonstrate that HyWorldVLA significantly outperforms both pixel-based and latent-based world model baselines. Notably, we present the first comprehensive qualitative and quantitative analysis of world model noise robustness in autonomous driving, establishing a new benchmark for evaluating future architectures.

Replacing an object with one that differs in category or shape requires complete source removal, natural target formation unconstrained by the source silhouette, and preservation of unrelated content. Existing training-free editors either localize edits from terminal predictions under source and target prompts or preserve unrelated content through spatially unselective source-feature reuse without explicit region discovery. Before reaching the terminal predictions, prompt-induced semantic differences undergo additional network transformations that may obscure their spatial localization, reducing localization precision. Spatially unselective feature reuse forces a trade-off between edit completeness and background preservation. Therefore, we propose PC-Edit, a prompt-contrastive framework for training-free MM-DiT editing. PC-Edit contrasts the image-token attention outputs under the source and target prompts, capturing prompt-induced semantic differences directly where text-conditioned information is delivered to image tokens. The same contrast identifies a source-erasure region during inversion and a target-emergence region during denoising. Their union suppresses source remnants while allowing the target object to form naturally. PC-Edit further couples region discovery and background preservation within each sampling step by estimating the current edit region from preceding attention blocks and immediately injecting cached source K/V features outside it in subsequent blocks, thereby protecting unrelated content before the latent update. Experiments on PIE-Bench and our EditRegion-Bench, with human-verified edit-region annotations for single- and multi-object addition and replacement, show that PC-Edit achieves the best editing quality and background preservation among methods without user-specified edit regions.

Streaming video diffusion models have made substantial progress toward interactive and dynamic world simulation, but the nested autoregressive and denoising loops of conventional next-frame generation hinder real-time deployment. Recent rolling-window methods pipeline denoising across multiple consecutive frames at different noise levels, improving throughput and long-horizon stability. However, they tokenize every state at the same fine spatial granularity, leaving substantial noise-dependent redundancy in the joint denoising window. We propose Ms.Forcing, an efficient streaming video generation paradigm that adapts spatial granularity to each state's noise level. Its Multi-Scale Patchification (MSP) assigns coarser patches to noisier states, reducing the active-window token count by 45%, while Multi-Scale Self-Attention (MSSA) matches the density of visible non-sink keys and values to each query scale to further reduce attention cost. Because both schedules are fixed by window position, Ms.Forcing retains a static, hardware-friendly computation graph. We further introduce Homogeneous-Noise-Level DMD (H-DMD), which assembles each fake video from clean predictions sharing the same source noise level, thereby reducing the mismatch between DMD training sequences and inference-time rollouts. The multi-scale design helps offset the additional training cost of backpropagating through overlapping windows. We include both quantitative and qualitative experiments to show that Ms.Forcing reaches 22.84 FPS on a single H200 GPU, 39.6% faster than Rolling Forcing, while significantly improving VBench scores in both short video and long video generation setting.

Video face swapping has no natural paired supervision: no real footage exists of one person's face performing another person's video. The strongest current answer, DreamID-V's SyncID-Pipe, mints pairs by replacing the identity in exactly two frames of a real clip — the first and the last — and regenerating the rest from a pose sequence alone. Pose carries no appearance evidence of the swapped-in identity, so over long clips, occlusions, and extreme pose excursions the synthesized identity has a long unanchored span on which to drift; no published ablation examines anchor count or placement. We propose Adaptive Identity Anchoring (AIA): (i) generalize the synthesizer to arbitrary anchor sets, architecturally natural for diffusion-forcing-style transformers where conditioning on a frame is clamping its tokens to zero noise; (ii) place anchors by a closed feedback loop that scores every generated frame against the real reference identity and inserts an image-face-swapped anchor at the worst-scoring frame until the pair passes a threshold or exhausts a budget; (iii) reuse the loop's verdict as an automatic data filter. A second pathology, the beauty-filter look of over-smoothed skin, has the same root cause: micro-texture, like identity, is priced by none of the pipeline's objectives. We therefore pair AIA with Reality-Referenced Texture Restoration: matched re-graining from each real frame's non-face regions, band-split transfer of sub-identity micro-texture from the real footage, and a second, spectral acceptance channel refereed by the footage's own spectrum. Identity-anchor density, we argue, is a controllable quality dial, and we specify falsifiable experiments — drift-versus-gap curves, uniform-versus-adaptive placement at matched budgets, student training on AIA-minted data, and texture ablations with a human beauty-filter study — that would validate or refute the proposal.

Creating dynamic and physically realistic 4D worlds from natural language descriptions is both fascinating and challenging. Traditional computer graphics methods rely on manual creation, requiring extensive human effort to fine-tune materials, motions, and visual fidelity. Recent advances in generative foundation models have sparked interest in learning to generate such 4D worlds from large-scale data; however, existing methods still struggle to ensure physical plausibility and controllability. In this work, we take a different path by leveraging foundation models to construct an agentic system that emulates how humans traditionally create 4D worlds, yet automates the entire process. We present GS-Agent, an end-to-end multi-agent framework that integrates physics engines in the loop to generate realistic, dynamic, and controllable 4D physical worlds from natural language. Inspired by how humans build 4D worlds, GS-Agent decomposes the task into entity management, covering 3D asset curation, material tuning, placement, and motion control, and rendering configuration, including camera and lighting manipulation. Multiple agents with distinct expertise interact with the physics engine via code, seek multimodal feedback, and collaborate to iteratively construct 4D worlds that align with the given descriptions. Experimental results show that GS-Agent effectively converts natural language into diverse and physically plausible 4D worlds exhibiting rich interactions among liquids, deformable objects, and rigid bodies, while achieving cinematic camera and lighting control. We envision GS-Agent as a foundation for a new paradigm in 4D world generation, empowering creative content creation and physical AI. Project page at https://umass-embodied-agi.github.io/gs-agent/

Rectified-flow-based diffusion transformers, particularly FLUX, have demonstrated outstanding performance in high-quality image generation. However, achieving fast and accurate inversion--transforming images back to latent noise for faithful reconstruction and editing--remains a challenging bottleneck due to the discretization errors of linear solvers. This paper introduces SlerpFlow, a straightforward yet highly effective zero-shot approach that unlocks the full potential of FLUX for high-fidelity inversion and editing. Unlike existing approaches (e.g., RF-Solver) that rely on complex numerical approximations such as high-order Taylor expansions to correct trajectory errors, we present a geometric view based on the Manifold Hypothesis: the empirically observed trajectory curvature is not a numerical artifact, but rather serves as a necessary "centripetal force" that constrains the flow to remain on the data manifold. Guided by this insight, SlerpFlow integrates Spherical Linear Interpolation (Slerp) to rectify flow velocity directions on the hypersphere, strictly adhering to the intrinsic curvature of the latent space. Crucially, by caching the corrected velocity for subsequent steps, SlerpFlow achieves high-precision inversion while maintaining the computational efficiency of a first-order Euler solver. Extensive experiments on FLUX-based reconstruction and editing tasks demonstrate that SlerpFlow improves reconstruction fidelity and achieves stronger semantic alignment in editing without requiring additional training. Code is available at https://github.com/0answer0/SlerpFlow.

Recent one-step text-to-image (T2I) models enable efficient image synthesis and provide new opportunities for real-time image editing. However, existing one-step editing methods primarily rely on text conditioning for semantic transformation, lacking explicit spatial control over where to edit. More importantly, even when spatial constraints are introduced, these methods often struggle to achieve strong and stable semantic modifications within the target regions. In this work, we revisit one-step image editing from a spatially controlled perspective and identify two key challenges: discovering editable regions and achieving effective localized semantic transformation. We reveal that existing methods perform global semantic transport, which limits high-intensity local editing under the one-step setting. To address this issue, we propose WhereEdit, a framework that reformulates one-step editing as localized adaptive editing. WhereEdit automatically identifies semantically relevant regions from internal model features and applies adaptive local modulation to enhance target-region editing while preserving non-target areas and structural consistency. Experiments on the PIE-Bench benchmark demonstrate that WhereEdit consistently outperforms existing one-step image editing methods, achieving superior editing quality while maintaining the efficiency of one-step generation. Additional experiments with region-level supervision further highlight the importance of explicit spatial reasoning for high-quality one-step image editing.

Makeup-transfer applies the reference makeup to the source face while preserving the source identity. Despite advances in full-face editing by diffusion-based methods, strong regional controllability, makeup fidelity, and identity preservation remain challenging. The reasons are (i) pixel-to-attention misalignment that causes spillover into non-target areas and weakens regional control; (ii) unclear transfer/preservation concept separation under two-image conditioning, leading to coupling between makeup attributes and identity; and (iii) the lack of a high-resolution dataset that is identity-consistent and region-labeled for fine-grained supervision. In this paper, we propose MagicMakeup, a diffusion transformer-based framework for region-controllable and high-fidelity makeup transfer, built on spatial constraints and concept disentanglement. To enable precise region-specific editing while preserving identity, we propose Token-Aligned Region Gating, which aligns pixel masks with attention and applies region-specific logit gating. To clarify the concepts of transfer and preservation, we further introduce Cross-Modal Perception Guidance, which aligns text and image features to enhance cross-modal concept perception. We also design a pipeline for the generation of 1024 x 1024 data pairs through region-specific makeup removal and establish a unified benchmark in synthetic and real settings. Extensive quantitative and qualitative experiments show that MagicMakeup improves regional controllability, makeup fidelity, and identity preservation, with strong robustness across styles, races, and poses.

Controllable video generation remains challenging due to the difficulty of specifying precise multi-object interactions using text prompts or motion-control inputs that primarily constrain pixel movement. In practice, trajectory-based control often requires users to draw accurate tracks for multiple objects, which scales poorly with scene complexity and becomes ambiguous under occlusion or overlap. To enable flexible yet precise multi-subject control, we introduce $GraphVid$, a graph-conditioned image-to-video generation model that enables interactive control through structured interaction graphs. We further curate $GraphVid-Bench$, a large-scale interaction-centric video dataset with structured relational annotations to enable training of interaction-aware video generation models. Despite using substantially less training data and fewer trainable parameters than prior motion-control methods, GraphVid delivers strong controllability and video quality. Compared with Motion-I2V, GraphVid reduces FID by up to 39.9% and FVD by 37.6%, while improving PSNR (9.87=>15.98) and SSIM (0.38=>0.61). Our results highlight the potential of structured semantic interfaces as a powerful paradigm for controllable video generation.

Current identity customized video generation methodologies are predominantly limited to single-identity scenarios, as the lack of explicit identity separation mechanisms often leads to identity confusion in multi-identity settings. Existing multi-identity approaches, which directly extend single-identity frameworks by concatenating face images as input conditions, frequently result in unnatural facial expressions and motions, manifesting as the "copy-paste" phenomenon. To overcome these limitations, we introduce GroupVideo, a novel framework that leverages multiple individual photographs to generate identitycustomized video. Built upon Video Diffusion Transformers, GroupVideo incorporates multimodal identity alignment: visual alignment jointly encodes multiple face images to provide robust identity references, while semantic alignment introduces a semantic perceiver to enhance the naturalness of motions. An ID localization module with spatial guidance is introduced to address identity blending and enhance identity fidelity, along with bounding box constraints and mask regularization loss, to focus on facial regions and improve training efficiency. In response to the shortage of multi-ID video datasets, we have curated a comprehensive high-quality dataset of 20,000 videos, thereby establishing a crucial resource to advance future research in multi-ID video generation. Extensive experiments demonstrate that GroupVideo outperforms existing methods in generating multi-character videos with consistent identities and natural motions.

Flow-based generative models have enabled remarkable progress in fast and controllable generation across continuous and discrete state spaces, yet existing parameterizations are constrained to fixed dimensions or fixed sequence lengths. Here, we introduce Expanding Generative Flows (EFlows), which define flows between distributions of increasing dimensionality along an expanding interpolant that grows the state by augmenting it with conditional noise. Building on this construction, we propose Expanding Flow Maps (EFMs), a new class of flow maps that distill the expanding interpolant into efficient few-step generative models. Each EFM factors the map between any two timesteps into two learnable operations: an expand operator, which augments the state space with new coordinates or tokens conditioned on the current state, and a transport map, which pushes the expanded state forward along the interpolant. Composing these operators yields a single map that jointly expands and denoises the state, recovering existing fixed-canvas flows and flow maps as the special case in which the expand operator is the identity. We further extend the framework to the discrete simplex, enabling variable-size graph generation and variable-length sequence generation. Across both continuous and discrete modalities, we establish EFlows and EFMs as a principled framework for settings in which output size is itself a learned, controllable degree of freedom.

Test-Time Tuning (TTT) on pretrained diffusion models has emerged as a powerful paradigm for video editing. However, there exists a foundational mismatch between the distribution-mapping nature of generative models and the single-point optimization of standard TTT. In this paper, we demonstrate that this mismatch triggers Prior Collapse, a degenerate state where the model discards the text conditions and spatial latents, collapsing generations to the source video, or entangling the features of distinct regions. To resolve this, we propose ElasticTTT, a novel framework that preserves the prior generative distribution and rescues generative elasticity. Specifically, we propose Target Distribution Regularization to prevent sharp memorization minima, Contrastive CFG to guide inference away from source biases, and Asynchronous Noise Schedule to preserve unedited regions. Extensive evaluations, supported by theoretical analysis, demonstrate that ElasticTTT successfully preserves the generative prior of the base model, achieving state-of-the-art performance on one-shot video editing.

Multi-agent interactive world models should not only generate consistent observations, but also maintain world states that persist across agents and evolve across views. Existing autoregressive video diffusion pipelines carry forward observation history as conditioning context, which makes shared state difficult to maintain in multi-agent and multi-view settings. We present WorldWeaver (W^2), a streaming multi-agent video diffusion model that augments rollout with cross-agent world state registers: learnable tokens that store shared world information, track individual agent status, and are dynamically updated after each generated chunk. We ground these registers with supervision signals spanning individual agent status, global state views including bird's-eye views, and scene text. We further improve the architecture with a Mixture-of-Transformers design that uses separate weights for world state modeling and visual frame modeling. Extensive experiments in two-agent Minecraft video generation show that explicit world-state modeling improves logical consistency and generation quality.

Diffusion and flow-matching models dominate conditional image generation, yet inference-time scaling for these models is far less developed than for autoregressive language models. Because final quality is highly sensitive to the initial noise seed, many approaches spend extra compute on seed search or resampling under a black-box reward, but typically maintaining a constant memory footprint throughout inference. We show that relaxing this constraint enables an underexplored inference-time scaling axis: by front-loading exploration, evaluating many seeds early, and pruning aggressively, we can use a fixed compute budget more effectively. Progressive Seed Pruning (\PSP) scores intermediate denoised estimates and progressively narrows the candidate set so that only promising trajectories are fully denoised, while keeping the total number of model evaluations fixed. Across diffusion and flow-matching backbones, \PSP \ consistently improves reward-guided selection and achieves higher GenEval scores (automated) and better human evaluation on prompt-alignment than best-of-$N$, importance-sampling, and tree-search baselines at matched compute. Project page: https://www.vision.caltech.edu/psp. Code: https://github.com/rogerioagjr/psp.

Post-training quantization (PTQ) of diffusion transformers (DiTs) to W4A4 severely degrades output quality, because activations entering each linear layer contain outliers that 4-bit formats cannot represent. The standard fix applies an invertible linear transform to the activations and its inverse to the weights before quantizing both. Normalization layers between blocks force this transform to run online at every denoising step, making its inference computation cost the binding design constraint. Existing options trade quantization quality for inference cost: per-channel scaling (SmoothQuant) is computationally cheap but impacts the magnitude of the channels, which can harm quantization accuracy; fixed Hadamard transforms yield better quantization accuracy but require large block sizes that incur a high online cost; learned full-$d$ invertible transforms calibrate best but entail a prohibitive dense $d \times d$ matrix multiplication (GEMM) per layer per step. We propose KroQuant, a PTQ method that applies a learned Kronecker-structured invertible transform to each 32-element block of the activation, storing less than half the parameters of per-channel scaling. The block-local structure runs as small tensor-core GEMMs, and on an MI350 GPU the KroQuant quantizer kernel is up to $14%$ faster than the SmoothQuant kernel. Offline LoRaQ weight calibration then absorbs the residual per-weight quantization error. On PixArt-$Σ$, SANA, and FLUX.1-schnell at W4A4 (MXFP4e2), KroQuant produces outputs closer to the FP reference than SVDQuant and LoRaQ on MJHQ-30K and SDCI, while preserving or improving image quality.

We introduce SANA-Video 2.0, a hybrid video diffusion transformer instantiated at 5B and 14B scales under a unified architecture. Designed to generate high-quality video up to 720p on a single GPU, SANA-Video 2.0 matches full-softmax video DiTs in quality while retaining the favorable long-sequence scaling of linear attention. To avoid quadratic attention throughout, Hybrid Linear-Softmax Attention combines gated linear attention for O(N)-dominated mixing with periodic gated-softmax anchors at a 3:1 ratio, restoring the full-rank token interactions that pure linear attention lacks. To propagate these refreshed representations across depth, Block Attention Residuals (AttnRes) route completed block summaries into later linear layers, enabling anchor-feature reuse and boosting deep-layer effective rank by ~12%. Through from-scratch training, SANA-Video 2.0 learns the complete hybrid directly rather than linearizing pretrained models, with reduced-resolution proxy studies establishing 25% softmax as the optimal quality-efficiency trade-off. With 40-step sampling, SANA-Video 2.0 achieves a VBench score of 84.30 in 13.2s at 480p on a single H100, remaining competitive with far larger softmax video DiTs at a fraction of the latency. Its compiled DiT forward pass is 3.2x faster than a matched full-softmax baseline at 720p/60s, a gap that expands with video duration. Furthermore, full-stack Sol-Engine optimization (kernel fusion, caching, and sparse attention) accelerates this hardware-friendly backbone by a further 3.58x, bringing the 5B pipeline to 13.06s at 720p/5s and making it 120x faster than Wan 2.2-A14B on one H100. Overall, our hybrid design recovers softmax-level expressiveness at substantially reduced cost, unlocking scalable long, high resolution video generation.

Predicting how deformable objects evolve under robotic manipulation is a longstanding challenge. Existing approaches typically rely on per-object optimization to fit material parameters, which can be slow and cannot generalize, while end-to-end learned alternatives extrapolate poorly and often violate basic physical structure. We present PhysCoRe, a physics-corrected residual world model that couples a differentiable Material Point Method (MPM) simulator with two feed-forward neural networks. A material refinement module, Material from Motion (MfM), infers per-particle elasticity from visual observations, grounding the simulator in object-specific physics. A residual correction module, Residual from Dynamics (RfD), learns the discrepancy and predicts corrections to the simulator's internal dynamics, absorbing systematic biases that the analytical model cannot capture. This design also supports online material identification on novel objects. MfM adapts from limited interactions, and its predictive uncertainty steers further exploration toward the regions where its estimate is least confident. Experiments on real deformable-object manipulation sequences show that PhysCoRe outperforms state-of-the-art baselines in prediction accuracy, and that its predicted confidence forms a reliable distribution across the object's geometry, providing a natural signal for future confidence-guided exploration.

Recent advances in 3D scene editing have leveraged iterative diffusion models to update input views. However, this process is computationally expensive and struggles to produce sharp details. Meanwhile, ``hallucination drift'' frequently introduces multi-view inconsistencies, leading to structural artifacts when rendering novel viewpoints. To address this problem, we present 3D-GIMP (3D Gaussian Inpainting Meets Patch Matching), a novel hybrid paradigm designed for high-fidelity object removal in 3D Gaussian Splatting. Instead of diffusing every view, 3D-GIMP performs a single generative inpainting on a key reference view, which serves as an appearance prior. We then introduce a 3D-aware PatchMatch algorithm to propagate these reference textures across all remaining views via correspondence matching, effectively bypassing the stochastic nature of frame-by-frame diffusion. By prioritizing reconstructive consistency over iterative generation, 3D-GIMP maintains high-frequency details across arbitrary resolutions while ensuring a mathematically consistent 3D reconstruction. Our experiments demonstrate that 3D-GIMP not only achieves competitive inpainting quality as previous methods using diffusion in multiple views, but also outperforms these methods in rendering speed and view consistency.

Real-world video deblurring remains challenging due to diverse motion patterns, complex degradations, and the scarcity of realistic training data, yet robust restoration is critical for downstream pipelines such as mobile imaging and 3D reconstruction. This work presents RealVDeblur, an efficient generative framework designed to improve in-the-wild robustness under diverse real capture conditions. First, a large-scale, physically grounded blur synthesis pipeline is constructed from scene-level 3D Gaussian Splatting (3DGS) assets and high-frame-rate videos, providing realistic training data covering both camera-induced and object-motion blur. Second, a video diffusion prior is leveraged for restoration; to better accommodate frame-dependent blur variations, temporal compression in the VAE is disabled and a frame-wise encoding scheme is adopted. For practical deployment on long videos, multi-step diffusion sampling is distilled into an efficient one-step generator, and a training-free Temporal Window Mask stabilizes inference beyond the training horizon with constant memory usage. Extensive experiments on diverse real-world benchmarks demonstrate strong perceptual quality, semantic fidelity, and temporal consistency on unseen videos, as well as improved robustness in downstream 3D reconstruction under severe motion blur. Project page: https://rbjin.github.io/RealVDeblur

Recent 3D generative models produce high-quality geometry from a single image using large-scale priors and diffusion architectures. However, they assume complete visibility and single-view inputs, limiting applicability in multi-view, occluded, or editing scenarios. Although prior works address these challenges individually, they lack a unified framework for controllable 3D completion under diverse conditioning signals. We present Axolotl3D, a multi-modal and occlusion-aware 3D generation model that jointly conditions on images, visibility masks, camera parameters, and a partial point cloud. The point cloud serves as a geometric anchor promoting faithful shape completion, while camera parameters ensure consistent multi-view alignment in a shared 3D coordinate system. A unified training strategy synthesizes diverse conditioning regimes from large-scale 3D data, enabling robust cross-modal reasoning. Experiments on Toys4K and OmniObject3D demonstrate state-of-the-art performance under both clean and occluded settings, as well as strong results in real-world reconstruction and geometry-consistent editing.

Thermal-to-visible face translation presents fundamental challenges including geometric discontinuities, semantic attribute mismatches, and identity degradation. We propose MTVDiff, a novel multimodal latent diffusion framework that synergistically integrates depth and textual information to address these limitations while preserving identity characteristics. The MTVDiff framework presents three core technical contributions: (1) a Dual-Branch Cross-Attention Fusion (DBCAF) module for multi-scale thermal-depth feature extraction and fusion; (2) a Gated Text-to-Visual Feature Alignment mechanism for semantically-guided generation; and (3) Spatial Feature Transformations (SFT) for adaptive multimodal prior integration. Extensive experiments on the MCXFace and SpeakingFaces datasets demonstrate that our multimodal approach significantly outperforms existing GAN-based and diffusion-based approaches across multiple metrics, achieving substantial improvements in both image quality and face verification performance, with FID reductions of up to 48.3% and Rank-1 accuracy improvements of up to 8.9%. Our work provides a robust solution for face recognition systems operating under varying illumination conditions and advances the state-of-the-art in cross-spectral facial image translation through effective multimodal integration.

Latent world models improve sample efficiency in continuous control by optimizing policies over imagined latent trajectories, but common neural transitions offer limited direct control over modal persistence and error accumulation in long rollouts. We propose Koopman Dreamer, a Dreamer-style world model with a spectrally constrained deterministic latent dynamics core. Its Koopman-inspired backbone uses two-dimensional rotation--scaling blocks with bounded radii to represent damping, rotation, and near-periodic modes. Linear and low-rank bilinear action terms capture global and state-dependent control effects, while stochastic-state modulation supplies local correction information. To reduce the mismatch between posterior-conditioned training and prior-only imagination, the model combines posterior-conditioned EMA teacher targets with one-step consistency, multi-step rollout, and open-loop observation-prediction objectives. We further derive a multi-step rollout-error bound that separates amplification by the spectral backbone and bilinear interaction from the additive effects of stochastic-state mismatch and modeling residuals, clarifying the trade-off between error attenuation and long-term information retention. Experimental results on proprioceptive continuous-control tasks from the DeepMind Control Suite and UAV-LiDAR autonomous navigation demonstrate that Koopman Dreamer improves the stability of long-horizon latent rollouts and achieves stronger closed-loop control performance on tasks that rely on high-quality multi-step imagination.

Evaluating the physical consistency of embodied world models(EWMs) is a critical open challenge. While closed-loop evaluation via simulator rollouts offers a more faithful assessment of physical plausibility than open-loop alternatives, existing frameworks almost exclusively rely on Inverse Dynamics Models(IDMs) for action extraction. Due to the intricate mapping from 2D pixel space to 3D kinematic space, the learned IDMs can be brittle to data outside their training distribution, resulting in unreliable action extraction from the generated videos with novel objects and scenarios. This creates an unavoidable attribution ambiguity between world model inaccuracies and extractor errors. To reduce this ambiguity, we present KineBench, an IDM-free closed-loop benchmark for EWMs, built upon an explicit kinematic grounding pipeline. Given a generated video, KineBench employs cascaded visual foundation models to directly extract 6D end-effector poses from individual frames, which are then executed in a physics simulator for closed-loop validation. Beyond execution-based task success, KineBench incorporates two classical 3D kinematic metrics--Spectral Arc Length (SPARC) and the Maruyama Manipulability Index--to characterize trajectory smoothness and kinematic feasibility from a robot-centric perspective. Built on 20 diverse manipulation tasks in ManiSkill3, KineBench evaluates EWMs across four progressive suites: basic execution, task transfer, visual out-of-distribution generalization, and complexity-conditioned scaling. Evaluation across frontier models reveals task-complexity-bounded nonlinear scaling in embodied video generation, providing empirical guidance for future data-scaling strategies.