Theory of general reinforcement learning
Reinforcement learning is the task of learning to act well in a variety of unknown environments. The traditional approach is to study small classes and construct computationally and data efficient algorithms to minimise some form of loss function such as regret or sample-complexity. The grand dream, however, is to solve the problem where the class of possible environments is sufficiently large to include any challenge that might reasonably be faced by an agent living in this universe. Such a...[Show more]
|Collections||Open Access Theses|
|b35790301-Lattimore_T.pdf||379.96 MB||Adobe PDF|
Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.