CoBERL: Contrastive BERT for Reinforcement Learning
This addresses data inefficiency in reinforcement learning for agents learning from pixels, though it appears incremental as it builds on existing contrastive and transformer methods.
The paper tackled the problem of reinforcement learning agents requiring large amounts of experience by proposing CoBERL, which improved data efficiency and achieved consistent performance gains across Atari, control tasks, and a 3D environment.
Many reinforcement learning (RL) agents require a large amount of experience to solve tasks. We propose Contrastive BERT for RL (CoBERL), an agent that combines a new contrastive loss and a hybrid LSTM-transformer architecture to tackle the challenge of improving data efficiency. CoBERL enables efficient, robust learning from pixels across a wide range of domains. We use bidirectional masked prediction in combination with a generalization of recent contrastive methods to learn better representations for transformers in RL, without the need of hand engineered data augmentations. We find that CoBERL consistently improves performance across the full Atari suite, a set of control tasks and a challenging 3D environment.