AISAT-CQL.pdf (397.98 kB)
Reducing the Time Complexity of Goal-Independent Reinforcement Learning
conference contribution
posted on 2023-05-26, 07:32 authored by Ollington, R, Vamplew, PConcurrent Q-Learning (CQL) is a goal independent\ reinforcement learning technique that learns the action\ values to all states simultaneously. These action values\ may then be used in a similar way to eligibility traces to\ allow many action values to be updated at each time\ step. CQL learns faster than conventional Q-learning\ techniques with the added benefit of being able to apply\ all experiences gained performing one task to any new\ task within the problem domain. Unfortunately the\ update time complexity of CQL is O(|S|2x|A|). This\ paper presents a technique for reducing the update\ complexity of CQL to O(|A|) with little impact on\ performance.
History
Pagination
132-137Publication status
- Published
Event title
AISAT2004: International Conference on Artificial Intelligence in Science and TechnologyEvent Venue
Hobart, Tasmania, AustraliaDate of Event (Start Date)
2004-11-21Date of Event (End Date)
2004-11-25Repository Status
- Open