Skip to content

Official implementation of the paper: LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations

Notifications You must be signed in to change notification settings

Yutang-Lin/LessMimic

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

teaser

arXiv paper github demo

LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations

Yutang Lin*, Jieming Cui*, Yixuan Li, Baoxiong Jia†, Yixin Zhu†, Siyuan Huang

* Equal contribution   † Corresponding author

Abstract

Humanoid robots that autonomously interact with physical environments over extended horizons represent a central goal of embodied intelligence. Existing approaches rely on reference motions or task-specific rewards, tightly coupling policies to particular object geometries and precluding multi-skill generalization within a single framework. Here we show that Distance Field (DF) provides such a representation: LessMimic conditions a single whole-body policy on DF-derived geometric cues — surface distances, gradients, and velocity decompositions — removing the need for motion references, with interaction latents encoded via a VAE and post-trained using AIP-derived RL. Through DAgger-style distillation that aligns DF latents with egocentric depth features, LessMimic further transfers seamlessly to vision-only deployment without motion capture infrastructure. A single LessMimic policy achieves 80-100% success across object scales from 0.4x to 1.6x on PickUp and SitStand where baselines degrade sharply, attains 62.1% success on 5-task trajectories, and remains viable up to 40 sequentially composed tasks. By grounding interaction in local geometry rather than demonstrations, LessMimic offers a scalable path toward humanoid robots that generalize, compose skills, and recover from failures in unstructured environments.

Code

Code will be released. Stay tuned.

Citation

@article{lin2026lessmimic,
  title={LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations},
  author={Lin, Yutang and Cui, Jieming and Li, Yixuan and Jia, Baoxiong and Zhu, Yixin and Huang, Siyuan},
  journal={arXiv preprint arXiv:2602.21723},
  year={2026},
  url={https://arxiv.org/abs/2602.21723}
}

About

Official implementation of the paper: LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published