Drop-Bottleneck: Learning Discrete Compressed Representation for Noise-Robust Exploration

Open in new window