Interactive Reinforcement Learning Human Generated Reward Presentation for Summer Camp 2015 May 25 2015
Reinforcement Learning • Trial and error learning • Explore and exploit Sutton and Barto 1988 • Represent, predict and control • Connect actions with rewards • Maximize future reward
Interactive Machine Learning Fails and Olsen Jr. 2003
Human Generated Reward • Humans know more! • Shaping systems to adapt Knox and Stone 2012 • Effectively reward learning • Transfer learning through collaboration • How can RL harness human reward?
Learning from Shaping Learning from Advice Kuhlmann et al. 2004 Blumberg et al. 2002 Learning from Demonstration Left: Argall et al. 2010 Right: Koenemann et al. 2014 Thomaz et al. 2006
Learning from Trial and Error Learning from Refinement Levine et al. 2015 Cakmak et al. 2012
Application • Shared control • Augmented representation • Integrate human and non-human interaction • Autonomous prosthetics
Recommend
More recommend