| title | Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos |
|---|---|
| authors | Lin Long, Tianren Ma, Zhong Chu, Di Wu, Maosong Cao, Qixiang Ye, Junbo Zhao, Shanghang Zhang |
| date | 2026-11-10 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2611.13685 |
| affiliation | Ant Group |
| upvotes | 5 (Hugging Face Daily Papers, 2026-11-19) |
| links | abs pdf hf code |
| cite | |
| related | SearchOS-V1: Towards Robust Open-Domain Information-Seeking Agent Collaboration UI-Venus-2 Technical Report SWE-Trail: History-Guided Trajectory Collection for Coding Agent Training SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories CausalCine: Real-Time Autoregressive Generation for Multi-Shot Video Narratives |
| abstract | In this paper, we investigate how human egocentric videos can benefit robot policy learning without action supervision. Building on a pretrained VLM, we propose Shared Prediction on Action and Representation (SPAR), an MoT architecture in which future visual state prediction can provide independent and complementary learning signals alongside action prediction. We find that representations learned from egocentric videos improve downstream performance on out-of-distribution robot tasks, with further gains as video data scales. We also study the effects of objective choices and data curation, providing insights into how to leverage action-free videos for scalable robot learning. |
bibtex
@misc{long2026ego,
title={Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos},
author={Lin Long and Tianren Ma and Zhong Chu and Di Wu and Maosong Cao and Qixiang Ye and Junbo Zhao and Shanghang Zhang},
year={2026},
eprint={2611.13685},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2611.13685}
}