| title | JEPA4R: Embodied JEPA World Models with 4D Point Planning |
|---|---|
| authors | Jiarui Dong, Zhiheng Liang, Yanjun Li, Zeyu Zhang, Heejeong Nam, Haoyu Chen, Hao Tang |
| date | 2026-11-15 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2611.19771 |
| affiliation | Peking University |
| upvotes | 5 (Hugging Face Daily Papers, 2026-11-19) |
| links | abs pdf hf code |
| cite | |
| primer | World models |
| related | SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes? DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines OmniEdu: Open Foundation Models for Learning and Teaching DataFlex-RL: An Evaluation Platform for RLVR Data Policies K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs |
| abstract | Joint-embedding predictive architecture (JEPA) world models predict future embeddings instead of pixels and plan in latent space. Two challenges remain: 2D images leave the metric geometry that manipulation acts on implicit and entangled with appearance, and what these models imagine stays hidden, as they are mostly judged in simulation by short-horizon latent planning. We present JEPA4R, a JEPA world model that observes single-depth-camera point clouds, predicts and plans in their latent space, and acts through a policy head on its frozen latents; we call this 4D point planning. A decoder on the frozen world model shows what it imagines as 4D point clouds, with a mean point error within 1.1 cm over 200 steps. Under the policy-learning protocols of OGBench and CortexBench, JEPA4R leads pixel JEPA models on OGBench Cube-Single, Cube-Double, Cube-Triple and Scene, and ranks first on MetaWorld and Adroit Relocate with 98.4% and 88.0%, above the runner-up, R3M, at 96.0% and 65.3%. Trained only in a digital twin, JEPA4R and its policy heads run zero-shot on the real arm with both a gripper and a dexterous hand. |
bibtex
@misc{dong2026jepa,
title={JEPA4R: Embodied JEPA World Models with 4D Point Planning},
author={Jiarui Dong and Zhiheng Liang and Yanjun Li and Zeyu Zhang and Heejeong Nam and Haoyu Chen and Hao Tang},
year={2026},
eprint={2611.19771},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2611.19771}
}