Timezone: »

 
Poster
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
Michael Laskin · Aravind Srinivas · Pieter Abbeel

Thu Jul 16 09:00 AM -- 09:45 AM & Thu Jul 16 08:00 PM -- 08:45 PM (PDT) @
in Poster Session 44 »

We present CURL: Contrastive Unsupervised Representations for Reinforcement Learning. CURL extracts high-level features from raw pixels using contrastive learning and performs off-policy control on top of the extracted features. CURL outperforms prior pixel-based methods, both model-based and model-free, on complex tasks in the DeepMind Control Suite and Atari Games showing 1.9x and 1.2x performance gains at the 100K environment and interaction steps benchmarks respectively. On the DeepMind Control Suite, CURL is the first image-based algorithm to nearly match the sample-efficiency of methods that use state-based features. Our code is open-sourced and available at https://www.github.com/MishaLaskin/curl.

Keywords: Algorithms · Deep Reinforcement Learning · Representation Learning · Reinforcement Learning - Deep RL

Author Information

Michael Laskin (UC Berkeley)
Aravind Srinivas (UC Berkeley)
Pieter Abbeel (UC Berkeley & Covariant)

More from the Same Authors