Improving Context-Based Meta-Reinforcement Learning with Self-Supervised Trajectory Contrastive Learning.
Bernie WangSimon XuKurt KeutzerYang GaoBichen WuPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- learning algorithm
- learning process
- function approximation
- learning problems
- state space
- supervised learning
- online learning
- unsupervised learning
- learning agents
- learning tasks
- machine learning
- training data
- active learning
- inductive inference
- learning mechanism
- reinforcement learning algorithms
- learning agent
- decision trees
- relational reinforcement learning