Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs

Details Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs

2006-10-30

arxiv.org/abs/cs/0610170v1

Computer Science

Learning

23 pages

Scientific paper

In this paper we propose a method that learns to play Pac-Man. We define a set of high-level observation and action modules. Actions are temporally extended, and multiple action modules may be in effect concurrently. A decision of the agent is represented as a rule-based policy. For learning, we apply the cross-entropy method, a recent global optimization algorithm. The learned policies reached better score than the hand-crafted policy, and neared the score of average human players. We argue that learning is successful mainly because (i) the policy space includes the combination of individual actions and thus it is sufficiently rich, (ii) the search is biased towards low-complexity policies and low complexity solutions can be found quickly if they exist. Based on these principles, we formulate a new theoretical framework, which can be found in the Appendix as supporting material.

Affiliated with

Also associated with

No associations

Rating

Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.

If you have personal experience with Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Low-complexity modular policies: learning to play Pac-Man and a new framework beyond MDPs will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFWR-SCP-O-230355