Hugging Face Daily Papers · · 4 min read

ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow

Mirrored from Hugging Face Daily Papers for archival readability. Support the source by reading on the original site.

ODEWorld explores continuous-time world modeling by learning a latent ODE velocity field instead of discrete-time transitions. The approach provides elegant properties such as arbitrary temporal resolution and backward prediction, and shows promising results for robotic planning and video prediction.</p>\n","updatedAt":"2026-08-03T07:55:38.839Z","author":{"_id":"677259dfd3a8d679c9683f71","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/w1ee-KLiJBuZvnXmGi86G.png","fullname":"Dstate","name":"ldxxx","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8510968685150146},"editors":["ldxxx"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/w1ee-KLiJBuZvnXmGi86G.png"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.27924","authors":[{"_id":"6a7048da02c90f968f48a13e","name":"Dongxiu Liu","hidden":false},{"_id":"6a7048da02c90f968f48a13f","name":"Haoyi Niu","hidden":false},{"_id":"6a7048da02c90f968f48a140","name":"Peng Cheng","hidden":false},{"_id":"6a7048da02c90f968f48a141","name":"Yuan Gao","hidden":false},{"_id":"6a7048da02c90f968f48a142","name":"Xirui Kang","hidden":false},{"_id":"6a7048da02c90f968f48a143","name":"Sangli Teng","hidden":false},{"_id":"6a7048da02c90f968f48a144","name":"Koushil Sreenath","hidden":false},{"_id":"6a7048da02c90f968f48a145","name":"Xianyuan Zhan","hidden":false}],"publishedAt":"2026-07-30T00:00:00.000Z","submittedOnDailyAt":"2026-08-03T00:00:00.000Z","title":"ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow","submittedOnDailyBy":{"_id":"677259dfd3a8d679c9683f71","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/w1ee-KLiJBuZvnXmGi86G.png","isPro":false,"fullname":"Dstate","user":"ldxxx","type":"user","name":"ldxxx"},"summary":"In the physical world we inhabit, space and time are fundamentally continuous. However, existing machine learning paradigms for world modeling are largely confined to discrete-time prediction, thereby exhibiting significant inefficiency in capturing the dynamics of physical world. We introduce Physical-Time Flow (PT-Flow), a novel approach that learns a continuous latent velocity field operating in physical time. Crucially, the underlying dynamics of sequential data are parameterized by an ordinary differential equation (ODE) embedded in a well-structured representation space. Under this paradigm, the prediction of future can be recast as temporal integration via an ODE solver in the compressed latent space. Building upon PT-Flow, we construct ODEWorld, a continuous-time latent world model that is both efficient and versatile. By extracting time-variant features and enforcing ODE properties on both the dynamical representation space and the latent velocity field, ODEWorld effectively addresses the long-standing representation collapse issue in latent world model literature. This also enables high-quality image reconstruction even after long-horizon prediction. Moreover, its continuous nature allows for arbitrary temporal resolution and even backward prediction, which is impossible for most discrete-time models. Lastly, ODEWorld can provide rich planning-oriented information to facilitate downstream policy learning. Comprehensive experiments demonstrate that ODEWorld successfully reconciles planning-conducive dynamics abstraction with visual realism, excelling in both video generation and robotic control. https://dstate.github.io/odeworld_website/{Project Website}.","upvotes":1,"discussionId":"6a7048db02c90f968f48a146","projectPage":"https://dstate.github.io/odeworld_website/","githubRepo":"https://github.com/Dstate/ODEWorld","githubRepoAddedBy":"user","githubStars":11},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"677259dfd3a8d679c9683f71","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/w1ee-KLiJBuZvnXmGi86G.png","isPro":false,"fullname":"Dstate","user":"ldxxx","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2607/2607.27924.md","query":{}}">
Papers
arxiv:2607.27924

ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow

Published on Jul 30
· Submitted by
Dstate
on Aug 3
Authors:
,

Abstract

In the physical world we inhabit, space and time are fundamentally continuous. However, existing machine learning paradigms for world modeling are largely confined to discrete-time prediction, thereby exhibiting significant inefficiency in capturing the dynamics of physical world. We introduce Physical-Time Flow (PT-Flow), a novel approach that learns a continuous latent velocity field operating in physical time. Crucially, the underlying dynamics of sequential data are parameterized by an ordinary differential equation (ODE) embedded in a well-structured representation space. Under this paradigm, the prediction of future can be recast as temporal integration via an ODE solver in the compressed latent space. Building upon PT-Flow, we construct ODEWorld, a continuous-time latent world model that is both efficient and versatile. By extracting time-variant features and enforcing ODE properties on both the dynamical representation space and the latent velocity field, ODEWorld effectively addresses the long-standing representation collapse issue in latent world model literature. This also enables high-quality image reconstruction even after long-horizon prediction. Moreover, its continuous nature allows for arbitrary temporal resolution and even backward prediction, which is impossible for most discrete-time models. Lastly, ODEWorld can provide rich planning-oriented information to facilitate downstream policy learning. Comprehensive experiments demonstrate that ODEWorld successfully reconciles planning-conducive dynamics abstraction with visual realism, excelling in both video generation and robotic control. https://dstate.github.io/odeworld_website/{Project Website}.

Community

Paper submitter about 5 hours ago

ODEWorld explores continuous-time world modeling by learning a latent ODE velocity field instead of discrete-time transitions. The approach provides elegant properties such as arbitrary temporal resolution and backward prediction, and shows promising results for robotic planning and video prediction.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images

· Sign up or log in to comment

Get this paper in your agent:

hf papers read 2607.27924
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper

No model linking this paper

Cite arxiv.org/abs/2607.27924 in a model README.md to link it from this page.

Datasets citing this paper

No dataset linking this paper

Cite arxiv.org/abs/2607.27924 in a dataset README.md to link it from this page.

Spaces citing this paper

No Space linking this paper

Cite arxiv.org/abs/2607.27924 in a Space README.md to link it from this page.

Collections including this paper

No Collection including this paper

Add this paper to a collection to link it from this page.

Discussion (0)

Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.

Sign in →

No comments yet. Sign in and be the first to say something.

More from Hugging Face Daily Papers