Abstract
In reinforcement learning problems, a learning agent has the task of
learning a good or optimal strategy from interaction with his environment. At the
start of the learning task, the agent usually has very little information. Therefore,
when faced with complex problems that have a large state space, learning a good
strategy might be infeasible or too slow to work in practice. One way to overcome
this problem, is the use of guidance to supply the agent with traces of “reasonable
policies”. However, in a lot of cases it will be hard for the user to supply such a
policy. In this paper, we will investigate the use of transfer learning for Relational
Reinforcement Learning problems. The goal of transfer learning is to accelerate
learning on a target task after training on a different, but related, source task. More
specifically, we introduce an extension of the options framework to the relational
setting and show how one can learn skills that can be transferred across similar,
but different domains. We present some preliminary experiments showing the
possible advantages of using relational options for transfer learning.
Description
Relational Options learning framework
- Contains good overview of options-transfer papers
- Introduces a relational option concept via decision lists
- Rules in lists have both state and state-action precondition
- Both all preconditions (state & state-action) must fire
- state-action preconditions can parameterize final action choice
- Options are learned by:
1 first learning ploicy,
2 categorizing state-actions as either on or off policy,
3 learn a relational decision tree from those examples
4 flatten tree and extract all on-policy rules
Links and resources
Tags
community