An important property for lifelong-learning agents is the ability to combine existing skills to solve new unseen tasks. In general, however, it is unclear how to compose existing skills in a principled manner. We show that optimal value function composition can be achieved in entropy-regularised reinforcement learning (RL), and then extend this result to the standard RL setting. Composition is demonstrated in a high-dimensional video game environment, where an agent with an existing library of skills is immediately able to solve new tasks without the need for further learning.
Benjamin van Niekerk (University of the Witwatersrand)
Steve James (University of the Witwatersrand)
Adam Earle (University of the Witwatersrand)
Benjamin Rosman (Council for Scientific and Industrial Research)
Related Events (a corresponding poster, oral, or spotlight)
2019 Poster: Composing Value Functions in Reinforcement Learning »
Wed Jun 12th 06:30 -- 09:00 PM Room Pacific Ballroom