Enhancing Context-Based Meta-Reinforcement Learning Algorithms via An Efficient Task Encoder (Student Abstract)

No. 18: AAAI-21 Student Papers and Demonstrations

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 35

AAAI Student Abstract and Poster Program

Meta-Reinforcement Learning (meta-RL) algorithms enable agents to adapt to new tasks from small amounts of exploration, based on the experience of similar tasks. Recent studies have pointed out that a good representation of a task is key to the success of off-policy context-based meta-RL. Inspired by contrastive methods in unsupervised representation learning, we propose a new method to learn the task representation based on the mutual information between transition tuples in a trajectory and the task embedding. We also propose a new estimation for task similarity based on Q-function, which can be used to form a constraint on the distribution of the encoded task variables, making the task encoder encode the task variables more effective on new tasks. Experiments on meta-RL tasks show that the newly proposed method outperforms existing meta-RL algorithms.

10.1609/aaai.v35i18.17965