Robust Imitation Learning from Noisy Demonstrations

Abstract

Robust learning from noisy demonstrations is a practical but highly challenging problem in imitation learning. In this paper, we first theoretically show that robust imitation learning can be achieved by optimizing a classification risk with a symmetric loss. Based on this theoretical finding, we then propose a new imitation learning method that optimizes the classification risk by effectively combining pseudo-labeling with co-training. Unlike existing methods, our method does not require additional labels or strict assumptions about noise distributions. Experimental results on continuous-control benchmarks show that our method is more robust compared to state-of-the-art methods.

Cite

Text

Tangkaratt et al. "Robust Imitation Learning from Noisy Demonstrations." Artificial Intelligence and Statistics, 2021.

Markdown

[Tangkaratt et al. "Robust Imitation Learning from Noisy Demonstrations." Artificial Intelligence and Statistics, 2021.](https://mlanthology.org/aistats/2021/tangkaratt2021aistats-robust/)

BibTeX

@inproceedings{tangkaratt2021aistats-robust,
  title     = {{Robust Imitation Learning from Noisy Demonstrations}},
  author    = {Tangkaratt, Voot and Charoenphakdee, Nontawat and Sugiyama, Masashi},
  booktitle = {Artificial Intelligence and Statistics},
  year      = {2021},
  pages     = {298-306},
  volume    = {130},
  url       = {https://mlanthology.org/aistats/2021/tangkaratt2021aistats-robust/}
}