## On the Worst-case Analysis of Temporal-difference Learning AlgorithmsComputer Research Laboratory, [University of California, Santa Cruz, 1994 - Computer algorithms - 19 pages |

### What people are saying - Write a review

We haven't found any reviews in the usual places.

### Common terms and phrases

algorithm TD(A AT&T Bell Laboratories average per-trial loss batch model batch of instances best linear predictor bound in equation Cesa-Bianchi Computer Research Laboratory denote difference learning given in equation given in Lemma goal gradient descent identity matrix initial vector instances received ji)x know bounds learning algorithm learning problem learning rate learning to predict linear model Lº(u Long and Warmuth loss of TD"(1 lower bound Machine Learning Manfred Manfred K matrix Mt method of temporal minimize number of trials on-line learning optimal choice paper parameters prediction function prediction ji prediction rule proof prove real number Santa Cruz Section TD(X temporal difference learning temporal differences Theorem total loss training sequence tth element ulſ update rule upper bounds vector u e vector wi weight vector wt wi and learning Widrow-Hoff algorithm worst-case analysis worst-case approach worst-case loss bounds worst-case performance Xài XCOA)'Z Xèri