Recently I read this paper Replacing Rewards with Examples: Example-Based Policy Search via Recursive Classification by Eysenbach, et al. The blog post provides an excellent description of the paper (and code) which is a lot more mathematical. Is there an rllib implementation for this? Thx.