Rating-Based Reinforcement Learning

Devin White, Mingkang Wu, Ellen Novoseller, Vernon Lawhern, Nicholas R Waytowich, Yongcan Cao

ICMLW 2023

/icmlw/2023/white2023icmlw-ratingbased/

Abstract

This paper develops a novel rating-based reinforcement learning approach that uses human ratings to obtain human guidance in reinforcement learning. Different from the existing preference-based and ranking-based reinforcement learning paradigms, based on human relative preferences over sample pairs, the proposed rating-based reinforcement learning approach is based on human evaluation of individual trajectories without relative comparisons between sample pairs. The rating-based reinforcement learning approach builds on a new prediction model for human ratings and a novel multi-class loss function. We conduct several experimental studies based on synthetic ratings and real human ratings to evaluate the effectiveness and benefits of the new rating-based reinforcement learning approach.

PDF ICMLW OpenReview Semantic Scholar

Cite

Text

White et al. "Rating-Based Reinforcement Learning." ICML 2023 Workshops: MFPL, 2023.

Markdown

[White et al. "Rating-Based Reinforcement Learning." ICML 2023 Workshops: MFPL, 2023.](https://mlanthology.org/icmlw/2023/white2023icmlw-ratingbased/)

BibTeX

@inproceedings{white2023icmlw-ratingbased,
  title     = {{Rating-Based Reinforcement Learning}},
  author    = {White, Devin and Wu, Mingkang and Novoseller, Ellen and Lawhern, Vernon and Waytowich, Nicholas R and Cao, Yongcan},
  booktitle = {ICML 2023 Workshops: MFPL},
  year      = {2023},
  url       = {https://mlanthology.org/icmlw/2023/white2023icmlw-ratingbased/}
}