Exploring the Edges of Latent State Clusters for Goal-Conditioned Reinforcement Learning

Exploring unknown environments efficiently is a fundamental challenge in unsupervised goal-conditioned reinforcement learning.