ICML
2006
IEEE
14 years 5 months ago
2006
IEEE
We introduce the use of learned shaping rewards in reinforcement learning tasks, where an agent uses prior experience on a sequence of tasks to learn a portable predictor that est...