Expert policy가 필요한 단점
- Roll out learned policy
- Query expert action at visited states
- Aggregate corrections with existing data
- Update policy
human gated DAgger
Expert intervenes at time when policy makes mistake
DAgger
Creator
Creator
Seonglae ChoCreated
Created
2024 Apr 4 13:56Editor
Editor
Seonglae ChoEdited
Edited
2024 Jun 7 3:38Refs
Refs