Reward Adaptation on Point-Maze Shift (meta-test)
-5.37Average ReturnExpert
Evaluation Results
| Method | Links | |
|---|---|---|
| ExpertType=Oracle/Ground Truth2019.09 | -5.37 | |
| PEMIRLTask Mode=Reward Adaptation2019.09 | -9.04 | |
| AIRLTask Mode=Reward Adaptation2019.09 | -29.07 | |
| Meta-InfoGAILTask Mode=Reward Adaptation2019.09 | -29.72 |