Beyond lowest-warping cost action selection in trajectory transfer.
Dylan Hadfield-MenellAlex X. LeeChelsea FinnEric TzengSandy H. HuangPieter AbbeelPublished in: ICRA (2015)
Keyphrases
- action selection
- robot soccer
- reinforcement learning
- basal ganglia
- decision making
- temporal difference
- action selection mechanism
- cost reduction
- high cost
- total cost
- continuous state and action spaces
- knowledge transfer
- human robot
- transfer learning
- minimum cost
- vision system
- expected cost
- cost sensitive
- total reward
- markov chain