Dynamical Distance Learning for Semi-Supervised and Unsupervised Skill Discovery - ICLR 2020 Poster

Reinforcement learning requires manual specification of a reward function to learn a task. While in principle this reward function only needs to sp...