Abstraction Selection in Model-based Reinforcement Learning
Published on Dec 05, 20151945 Views
State abstractions are often used to reduce the complexity of model-based reinforcement learning when only limited quantities of data are available. However, choosing the appropriate level of abstract