Diffusion models have demonstrated exceptional visual quality in video generation, making them promising for autonomous driving world modeling. However, existing video diffusion-based world models struggle with flexible-length, long-horizon predictions and integrating trajectory planning. This is because conventional video diffusion models rely on global joint distribution modeling of fixed-length frame sequences rather than sequentially constructing localized distributions at each timestep. In this work, we propose Epona, an autoregressive diffusion world model that enables localized spatiotemporal distribution modeling through two key innovations: 1) Decoupled spatiotemporal factorization that separates temporal dynamics modeling from fine-grained future world generation, and 2) Modular trajectory and video prediction that seamlessly integrate motion planning with visual modeling in an end-to-end framework. Our architecture enables high-resolution, long-duration generation while introducing a novel chain-of-forward training strategy to address error accumulation in autoregressive loops. Experimental results demonstrate state-of-the-art performance with 7.4\% FVD improvement and minutes longer prediction duration compared to prior works. The learned world model further serves as a real-time motion planner, outperforming strong end-to-end planners on NAVSIM benchmarks. Code will be publicly available at \href{https://github.com/Kevin-thu/Epona/}{https://github.com/Kevin-thu/Epona/}.
Video generation models (VGMs) offer a promising pathway for unified world modeling in robotics by integrating simulation, prediction, and manipulation. However, their practical application remains limited due to (1) slowgeneration speed, which limits real-time interaction, and (2) poor consistency between imagined videos and executable actions. To address these challenges, we propose Manipulate in Dream (MinD), a hierarchical diffusion-based world model framework that employs a dual-system design for vision-language manipulation. MinD executes VGM at low frequencies to extract video prediction features, while leveraging a high-frequency diffusion policy for real-time interaction. This architecture enables low-latency, closed-loop control in manipulation with coherent visual guidance. To better coordinate the two systems, we introduce a video-action diffusion matching module (DiffMatcher), with a novel co-training strategy that uses separate schedulers for each diffusion model. Specifically, we introduce a diffusion-forcing mechanism to DiffMatcher that aligns their intermediate representations during training, helping the fast action model better understand video-based predictions. Beyond manipulation, MinD also functions as a world simulator, reliably predicting task success or failure in latent space before execution. Trustworthy analysis further shows that VGMs can preemptively evaluate task feasibility and mitigate risks. Extensive experiments across multiple benchmarks demonstrate that MinD achieves state-of-the-art manipulation (63%+) in RL-Bench, advancing the frontier of unified world modeling in robotics.
We train models to Predict Ego-centric Video from human Actions (PEVA), given the past video and an action represented by the relative 3D body pose. By conditioning on kinematic pose trajectories, structured by the joint hierarchy of the body, our model learns to simulate how physical human actions shape the environment from a first-person point of view. We train an auto-regressive conditional diffusion transformer on Nymeria, a large-scale dataset of real-world egocentric video and body pose capture. We further design a hierarchical evaluation protocol with increasingly challenging tasks, enabling a comprehensive analysis of the model's embodied prediction and control abilities. Our work represents an initial attempt to tackle the challenges of modeling complex real-world environments and embodied agent behaviors with video prediction from the perspective of a human.
Action-labeled data for robotics is scarce and expensive, limiting the generalization of learned policies. In contrast, vast amounts of action-free video data are readily available, but translating these observations into effective policies remains a challenge. We introduce AMPLIFY, a novel framework that leverages large-scale video data by encoding visual dynamics into compact, discrete motion tokens derived from keypoint trajectories. Our modular approach separates visual motion prediction from action inference, decoupling the challenges of learning what motion defines a task from how robots can perform it. We train a forward dynamics model on abundant action-free videos and an inverse dynamics model on a limited set of action-labeled examples, allowing for independent scaling. Extensive evaluations demonstrate that the learned dynamics are both accurate, achieving up to 3.7x better MSE and over 2.5x better pixel prediction accuracy compared to prior approaches, and broadly useful. In downstream policy learning, our dynamics predictions enable a 1.2-2.2x improvement in low-data regimes, a 1.4x average improvement by learning from action-free human videos, and the first generalization to LIBERO tasks from zero in-distribution action data. Beyond robotic control, we find the dynamics learned by AMPLIFY to be a versatile latent world model, enhancing video prediction quality. Our results present a novel paradigm leveraging heterogeneous data sources to build efficient, generalizable world models. More information can be found at https://amplify-robotics.github.io/.
Existing long-term video prediction methods often rely on an autoregressive video prediction mechanism. However, this approach suffers from error propagation, particularly in distant future frames. To address this limitation, this paper proposes the first AutoRegression-Free (ARFree) video prediction framework using diffusion models. Different from an autoregressive video prediction mechanism, ARFree directly predicts any future frame tuples from the context frame tuple. The proposed ARFree consists of two key components: 1) a motion prediction module that predicts a future motion using motion feature extracted from the context frame tuple; 2) a training method that improves motion continuity and contextual consistency between adjacent future frame tuples. Our experiments with two benchmark datasets show that the proposed ARFree video prediction framework outperforms several state-of-the-art video prediction methods.
Learning a generalizable bimanual manipulation policy is extremely challenging for embodied agents due to the large action space and the need for coordinated arm movements. Existing approaches rely on Vision-Language-Action (VLA) models to acquire bimanual policies. However, transferring knowledge from single-arm datasets or pre-trained VLA models often fails to generalize effectively, primarily due to the scarcity of bimanual data and the fundamental differences between single-arm and bimanual manipulation. In this paper, we propose a novel bimanual foundation policy by fine-tuning the leading text-to-video models to predict robot trajectories and training a lightweight diffusion policy for action generation. Given the lack of embodied knowledge in text-to-video models, we introduce a two-stage paradigm that fine-tunes independent text-to-flow and flow-to-video models derived from a pre-trained text-to-video model. Specifically, optical flow serves as an intermediate variable, providing a concise representation of subtle movements between images. The text-to-flow model predicts optical flow to concretize the intent of language instructions, and the flow-to-video model leverages this flow for fine-grained video prediction. Our method mitigates the ambiguity of language in single-stage text-to-video prediction and significantly reduces the robot-data requirement by avoiding direct use of low-level actions. In experiments, we collect high-quality manipulation data for real dual-arm robot, and the results of simulation and real-world experiments demonstrate the effectiveness of our method.
Diffusion and flow-based models have enabled significant progress in generation tasks across various modalities and have recently found applications in world modeling. However, unlike typical generation tasks that encourage sample diversity, world models entail different sources of uncertainty and require consistent samples aligned with the ground-truth trajectory, which is a limitation we empirically observe in diffusion models. We argue that a key bottleneck in learning consistent diffusion-based world models lies in the suboptimal predictive ability, which we attribute to the entanglement of condition understanding and target denoising within shared architectures and co-training schemes. To address this, we propose Foresight Diffusion (ForeDiff), a diffusion-based world modeling framework that enhances consistency by decoupling condition understanding from target denoising. ForeDiff incorporates a separate deterministic predictive stream to process conditioning inputs independently of the denoising stream, and further leverages a pretrained predictor to extract informative representations that guide generation. Extensive experiments on robot video prediction and scientific spatiotemporal forecasting show that ForeDiff improves both predictive accuracy and sample consistency over strong baselines, offering a promising direction for diffusion-based world models.
The task of estimating the world model describing the dynamics of a real world process assumes immense importance for anticipating and preparing for future outcomes. For applications such as video surveillance, robotics applications, autonomous driving, etc. this objective entails synthesizing plausible visual futures, given a few frames of a video to set the visual context. Towards this end, we propose ProgGen, which undertakes the task of video frame prediction by representing the dynamics of the video using a set of neuro-symbolic, human-interpretable set of states (one per frame) by leveraging the inductive biases of Large (Vision) Language Models (LLM/VLM). In particular, ProgGen utilizes LLM/VLM to synthesize programs: (i) to estimate the states of the video, given the visual context (i.e. the frames); (ii) to predict the states corresponding to future time steps by estimating the transition dynamics; (iii) to render the predicted states as visual RGB-frames. Empirical evaluations reveal that our proposed method outperforms competing techniques at the task of video frame prediction in two challenging environments: (i) PhyWorld (ii) Cart Pole. Additionally, ProgGen permits counter-factual reasoning and interpretable video generation attesting to its effectiveness and generalizability for video generation tasks.
GPT has shown its remarkable success in natural language processing. However, the language sequence is not sufficient to describe spatial-temporal details in the visual world. Alternatively, the video sequence is good at capturing such details. Motivated by this fact, we propose a concise Video-GPT in this paper by treating video as new language for visual world modeling. By analogy to next token prediction in GPT, we introduce a novel next clip diffusion paradigm for pretraining Video-GPT. Different from the previous works, this distinct paradigm allows Video-GPT to tackle both short-term generation and long-term prediction, by autoregressively denoising the noisy clip according to the clean clips in the history. Extensive experiments show our Video-GPT achieves the state-of-the-art performance on video prediction, which is the key factor towards world modeling (Physics-IQ Benchmark: Video-GPT 34.97 vs. Kling 23.64 vs. Wan 20.89). Moreover, it can be well adapted on 6 mainstream video tasks in both video generation and understanding, showing its great generalization capacity in downstream. The project page is at https://zhuangshaobin.github.io/Video-GPT.github.io/.
This paper investigates training better visual world models for robot manipulation, i.e., models that can predict future visual observations by conditioning on past frames and robot actions. Specifically, we consider world models that operate on RGB-D frames (RGB-D world models). As opposed to canonical approaches that handle dynamics prediction mostly implicitly and reconcile it with visual rendering in a single model, we introduce FlowDreamer, which adopts 3D scene flow as explicit motion representations. FlowDreamer first predicts 3D scene flow from past frame and action conditions with a U-Net, and then a diffusion model will predict the future frame utilizing the scene flow. FlowDreamer is trained end-to-end despite its modularized nature. We conduct experiments on 4 different benchmarks, covering both video prediction and visual planning tasks. The results demonstrate that FlowDreamer achieves better performance compared to other baseline RGB-D world models by 7% on semantic similarity, 11% on pixel quality, and 6% on success rate in various robot manipulation domains.
Generating videos in the first-person perspective has broad application prospects in the field of augmented reality and embodied intelligence. In this work, we explore the cross-view video prediction task, where given an exo-centric video, the first frame of the corresponding ego-centric video, and textual instructions, the goal is to generate futur frames of the ego-centric video. Inspired by the notion that hand-object interactions (HOI) in ego-centric videos represent the primary intentions and actions of the current actor, we present EgoExo-Gen that explicitly models the hand-object dynamics for cross-view video prediction. EgoExo-Gen consists of two stages. First, we design a cross-view HOI mask prediction model that anticipates the HOI masks in future ego-frames by modeling the spatio-temporal ego-exo correspondence. Next, we employ a video diffusion model to predict future ego-frames using the first ego-frame and textual instructions, while incorporating the HOI masks as structural guidance to enhance prediction quality. To facilitate training, we develop an automated pipeline to generate pseudo HOI masks for both ego- and exo-videos by exploiting vision foundation models. Extensive experiments demonstrate that our proposed EgoExo-Gen achieves better prediction performance compared to previous video prediction models on the Ego-Exo4D and H2O benchmark datasets, with the HOI masks significantly improving the generation of hands and interactive objects in the ego-centric videos.
Video prediction (VP) generates future frames by leveraging spatial representations and temporal context from past frames. Traditional recurrent neural network (RNN)-based models enhance memory cell structures to capture spatiotemporal states over extended durations but suffer from gradual loss of object appearance details. To address this issue, we propose the strong recollection VP (SRVP) model, which integrates standard attention (SA) and reinforced feature attention (RFA) modules. Both modules employ scaled dot-product attention to extract temporal context and spatial correlations, which are then fused to enhance spatiotemporal representations. Experiments on three benchmark datasets demonstrate that SRVP mitigates image quality degradation in RNN-based models while achieving predictive performance comparable to RNN-free architectures.
We address the challenge of representation learning from a continuous stream of video as input, in a self-supervised manner. This differs from the standard approaches to video learning where videos are chopped and shuffled during training in order to create a non-redundant batch that satisfies the independently and identically distributed (IID) sample assumption expected by conventional training paradigms. When videos are only available as a continuous stream of input, the IID assumption is evidently broken, leading to poor performance. We demonstrate the drop in performance when moving from shuffled to sequential learning on three tasks: the one-video representation learning method DoRA, standard VideoMAE on multi-video datasets, and the task of future video prediction. To address this drop, we propose a geometric modification to standard optimizers, to decorrelate batches by utilising orthogonal gradients during training. The proposed modification can be applied to any optimizer -- we demonstrate it with Stochastic Gradient Descent (SGD) and AdamW. Our proposed orthogonal optimizer allows models trained from streaming videos to alleviate the drop in representation learning performance, as evaluated on downstream tasks. On three scenarios (DoRA, VideoMAE, future prediction), we show our orthogonal optimizer outperforms the strong AdamW in all three scenarios.
Transmission latency significantly affects users' quality of experience in real-time interaction and actuation. As latency is principally inevitable, video prediction can be utilized to mitigate the latency and ultimately enable zero-latency transmission. However, most of the existing video prediction methods are computationally expensive and impractical for real-time applications. In this work, we therefore propose real-time video prediction towards the zero-latency interaction over networks, called IFRVP (Intermediate Feature Refinement Video Prediction). Firstly, we propose three training methods for video prediction that extend frame interpolation models, where we utilize a simple convolution-only frame interpolation network based on IFRNet. Secondly, we introduce ELAN-based residual blocks into the prediction models to improve both inference speed and accuracy. Our evaluations show that our proposed models perform efficiently and achieve the best trade-off between prediction accuracy and computational speed among the existing video prediction methods. A demonstration movie is also provided at http://bit.ly/IFRVPDemo.
Temporal consistency is critical in video prediction to ensure that outputs are coherent and free of artifacts. Traditional methods, such as temporal attention and 3D convolution, may struggle with significant object motion and may not capture long-range temporal dependencies in dynamic scenes. To address this gap, we propose the Tracktention Layer, a novel architectural component that explicitly integrates motion information using point tracks, i.e., sequences of corresponding points across frames. By incorporating these motion cues, the Tracktention Layer enhances temporal alignment and effectively handles complex object motions, maintaining consistent feature representations over time. Our approach is computationally efficient and can be seamlessly integrated into existing models, such as Vision Transformers, with minimal modification. It can be used to upgrade image-only models to state-of-the-art video ones, sometimes outperforming models natively designed for video prediction. We demonstrate this on video depth prediction and video colorization, where models augmented with the Tracktention Layer exhibit significantly improved temporal consistency compared to baselines.
Predicting future video frames is essential for decision-making systems, yet RGB frames alone often lack the information needed to fully capture the underlying complexities of the real world. To address this limitation, we propose a multi-modal framework for Synchronous Video Prediction (SyncVP) that incorporates complementary data modalities, enhancing the richness and accuracy of future predictions. SyncVP builds on pre-trained modality-specific diffusion models and introduces an efficient spatio-temporal cross-attention module to enable effective information sharing across modalities. We evaluate SyncVP on standard benchmark datasets, such as Cityscapes and BAIR, using depth as an additional modality. We furthermore demonstrate its generalization to other modalities on SYNTHIA with semantic information and ERA5-Land with climate data. Notably, SyncVP achieves state-of-the-art performance, even in scenarios where only one modality is present, demonstrating its robustness and potential for a wide range of applications.
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
Text-video prediction (TVP) is a downstream video generation task that requires a model to produce subsequent video frames given a series of initial video frames and text describing the required motion. In practice TVP methods focus on a particular category of videos depicting manipulations of objects carried out by human beings or robot arms. Previous methods adapt models pre-trained on text-to-image tasks, and thus tend to generate video that lacks the required continuity. A natural progression would be to leverage more recent pre-trained text-to-video (T2V) models. This approach is rendered more challenging by the fact that the most common fine-tuning technique, low-rank adaptation (LoRA), yields undesirable results. In this work, we propose an adaptation-based strategy we label Frame-wise Conditioning Adaptation (FCA). Within the module, we devise a sub-module that produces frame-wise text embeddings from the input text, which acts as an additional text condition to aid generation. We use FCA to fine-tune the T2V model, which incorporates the initial frame(s) as an extra condition. We compare and discuss the more effective strategy for injecting such embeddings into the T2V model. We conduct extensive ablation studies on our design choices with quantitative and qualitative performance analysis. Our approach establishes a new state-of-the-art for the task of TVP. The project page is at https://github.com/Cuberick-Orion/FCA .
Recent years, weather forecasting has gained significant attention. However, accurately predicting weather remains a challenge due to the rapid variability of meteorological data and potential teleconnections. Current spatiotemporal forecasting models primarily rely on convolution operations or sliding windows for feature extraction. These methods are limited by the size of the convolutional kernel or sliding window, making it difficult to capture and identify potential teleconnection features in meteorological data. Additionally, weather data often involve non-rigid bodies, whose motion processes are accompanied by unpredictable deformations, further complicating the forecasting task. In this paper, we propose the GMG model to address these two core challenges. The Global Focus Module, a key component of our model, enhances the global receptive field, while the Motion Guided Module adapts to the growth or dissipation processes of non-rigid bodies. Through extensive evaluations, our method demonstrates competitive performance across various complex tasks, providing a novel approach to improving the predictive accuracy of complex spatiotemporal data.
Training visual reinforcement learning (RL) in practical scenarios presents a significant challenge, $\textit{i.e.,}$ RL agents suffer from low sample efficiency in environments with variations. While various approaches have attempted to alleviate this issue by disentanglement representation learning, these methods usually start learning from scratch without prior knowledge of the world. This paper, in contrast, tries to learn and understand underlying semantic variations from distracting videos via offline-to-online latent distillation and flexible disentanglement constraints. To enable effective cross-domain semantic knowledge transfer, we introduce an interpretable model-based RL framework, dubbed Disentangled World Models (DisWM). Specifically, we pretrain the action-free video prediction model offline with disentanglement regularization to extract semantic knowledge from distracting videos. The disentanglement capability of the pretrained model is then transferred to the world model through latent distillation. For finetuning in the online environment, we exploit the knowledge from the pretrained model and introduce a disentanglement constraint to the world model. During the adaptation phase, the incorporation of actions and rewards from online environment interactions enriches the diversity of the data, which in turn strengthens the disentangled representation learning. Experimental results validate the superiority of our approach on various benchmarks.