Singh, S; Dayan, P; (1997) Analytical mean squared error curves in temporal difference learning. In: Mozer, MC and Jordan, MI and Petsche, T, (eds.) ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 9. (pp. 1054 - 1060). M I T PRESS
Full text not available from this repository.
We have calculated analytical expressions for how the bias and variance of the estimators provided by various temporal difference value estimation algorithms change with offline updates over trials in absorbing Markov chains using lookup table representations. We illustrate classes of learning curve behavior in various chains, and show the manner in which TD is sensitive to the choice of its step-size and eligibility trace parameters.
|Title:||Analytical mean squared error curves in temporal difference learning|
|Event:||10th Annual Conference on Neural Information Processing Systems (NIPS)|
|Dates:||1996-12-02 - 1996-12-05|
|UCL classification:||UCL > School of Life and Medical Sciences > Faculty of Life Sciences > Gatsby Computational Neuroscience Unit|
Archive Staff Only: edit this record