3 search results - page 1 / 1 » Efficient Reinforcement Learning with Multiple Reward Functi... |
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We introduce new, efficient algorithms for value iteration with multiple reward functions and continuous state. We also give an algorithm for finding the set of all nondominated a...
EUROCAST
2007
Springer
13 years 11 months ago
2007
Springer
Abstract a paradigm of modern Machine Learning (ML) which uses rewards and punishments to guide the learning process. One of the central ideas of RL is learning by “direct-online...
NIPS
2001
13 years 6 months ago
2001
We propose a new approach to reinforcement learning which combines least squares function approximation with policy iteration. Our method is model-free and completely off policy. ...