Imitation Learning from Imperfection: Theoretical Justifications and Algorithms

This repository has the code for reproducing experiments in NeurIPS 2023 spotlight paper: Imitation Learning from Imperfection: Theoretical Justifications and Algorithms

[Paper Link]

Usage

Please refer to the subfolders mujoco, atari, and image_classification for detailed instructions.

Bibtex

@inproceedings{li2023imitation,
  title     = {Imitation Learning from Imperfection: Theoretical Justifications and Algorithms},
  author    = {Li, Ziniu and Xu, Tian and Qin, Zeyu and Yu, Yang and Luo, Zhi-Quan},
  booktitle = {Advances in Neural Information Processing Systems 37},
  year      = {2023},
}