ATAL
2005
Springer
14 years 16 days ago
2005
Springer
Temporal difference (TD) learning methods [22] have become popular reinforcement learning techniques in recent years. TD methods have had some experimental successes and have been...
ISDA
2009
IEEE
14 years 1 months ago
2009
IEEE
This paper presents postponed updates, a new strategy for TD methods that can improve sample efficiency without incurring the computational and space requirements of model-based ...