Fetching the paper…

The Value-Improvement Path: Towards Better Representations for Reinforcement Learning · Around