로그인
통찰 - Meta-Reinforcement Learning Methods
SplAgger: Split Aggregation for Meta-Reinforcement Learning by Beck et al.
The authors present SplAgger, a method that combines permutation variant and invariant components to enhance meta-reinforcement learning performance.
SplAgger: Split Aggregation for Meta-Reinforcement Learning
Task inference sequence models are beneficial in meta-RL, even without task inference objectives.
SplAgger: Split Aggregation for Meta-Reinforcement Learning
Task inference sequence models are beneficial in meta-RL, even without task inference objectives.
