Robot learning from human demonstrations with inconsistent contexts

Zhifeng Qian, Mingyu You, Hongjun Zhou, Xuanhui Xu, Bin HeRobotics Auton. Syst 2023 · OpenReview

Abstract

Highlights•A framework is proposed to learn from human videos with inconsistent contexts.•The viewpoint transformation network can transform novel viewpoints by meta-learning.•The robot2human alignment network aligns the features between human and robots.•We collect a multi-view multi-context dataset for three tasks in the real world.
BibTeX
@inproceedings{qian2023robot,
  author = {Zhifeng Qian and Mingyu You and Hongjun Zhou and Xuanhui Xu and Bin He},
  title = {{Robot learning from human demonstrations with inconsistent contexts}},
  booktitle = {Robotics Auton. Syst},
  year = {2023},
  url = {https://openreview.net/forum?id=6XssUFmYIw},
}

Loading related content...

Figure