Practical Issues in Temporal Difference Learning

  title={Practical Issues in Temporal Difference Learning},
  author={G. Tesauro},
  journal={Machine Learning},
This paper examines whether temporal difference methods for training connectionist networks, such as Sutton's TD(λ) algorithm, can be successfully applied to complex real-world problems. A number of important practical issues are identified and discussed from a general theoretical perspective. These practical issues are then examined in the context of a case study in which TD(λ) is applied to learning the game of backgammon from the outcome of self-play. This is apparently the first application… Expand
427 Citations
Truncating Temporal Differences: On the Efficient Implementation of TD(lambda) for Reinforcement Learning
  • 41
  • Highly Influenced
  • PDF
Temporal difference learning with eligibility traces for the game connect four
  • 16
  • PDF
On learning with imperfect representations
  • 6
  • PDF
Learning to Play Chess Using Temporal Differences
  • 120
  • PDF
Evolution versus Temporal Difference Learning for learning to play Ms. Pac-Man
  • P. Burrow, S. Lucas
  • Computer Science
  • 2009 IEEE Symposium on Computational Intelligence and Games
  • 2009
  • 39
  • PDF
Learning to Play Hearts
  • 6
  • PDF
TD Learning of Game Evaluation Functions with Hierarchies Neural Architectures
  • 17
Gradient temporal-difference learning algorithms
  • 124
  • PDF