Neural Temporal-Difference Learning Converges to Global Optima
Different from existing consensus-type TD algorithms, the ap- proach here develops a simple decentralized TD tracker by wedding TD learning with gradient ...
Target-Based Temporal-Difference LearningIn this work, we introduce a new family of target-based temporal difference (TD) learning algorithms that main- tain two separate learning parameters ? the ... Incremental Least-Squares Temporal Difference Learning - AAAIThe least-squares TD algorithm (LSTD) is a recent alter- native proposed by Bradtke and Barto (1996) and extended by Boyan (1999; 2002) and Xu et al. (2002). An Analysis Of Temporal-difference Learning With Function ... - MITTemporal-difference learning, originally proposed by Sutton. [2], is a method for approximating long-term future cost as a function of current state. The ...
Autres Cours:
