Learning to Predict by the Methods of Temporal Differences

Machine Learning

Volumn 3, Issue 1, 1988, Pages 9-44

  Sutton, Richard S ^a  

^a VERIZON   (United States)

Author keywords

connectionism; credit assignment; evaluation functions; Incremental learning; prediction

Indexed keywords

EID: 33847202724     PISSN: 08856125     EISSN: 15730565     Source Type: Journal    
DOI: 10.1023/A:1022633531479     Document Type: Article

References (30)

1. A learning algorithm for Boltzmann machines
2. Doctoral dissertation
3. Strategy learning with multilayer connectionist representations
4. Learning by statistical cooperation of self-interested neuron-like computing elements
5. Neuronlike elements that can solve difficult learning control problems
6. Doctoral dissertation
7. Learning static evaluation functions by linear regression
8. A unified theory of heuristic evaluation functions and its application to learning
9. Learning to predict sequences
10. The logic of Limaxlearning
11. Doctoral dissertation
12. Disjunctive models of boolean category learning
13. Escaping brittleness:The possibilities of general purpose learning algorithms applied to parallel rule-based systems
14. Temporal primacy overrides prior training in serial compound conditioning of the rabbit's nicti-tating membrane response
15. Simulation of the classically conditioned nictitating membrane response by a neuron-like adaptive element:Response topography,neuronal firing and interstimulus intervals
16. Some studies in machine learning using the game of checkers
17. Toward a modern theory of adaptive networks:Expectation and prediction
18. An adaptive network that constructs and uses an internal model of its environment
19. A temporal-difference model of classical conditioning
20. The learning of world models by connectionist network
21. An adaptive optimal controller for discrete-time Markov environments