Mixture of Vector Experts

Authors: Matthew Henderson, John Shawe-Taylor, and Janez Žerovnik

Source: Algorithmic Learning Theory, 16th International Conference, ALT 2005, Singapore, October 2005, Proceedings, (Sanjay Jain, Hans Ulrich Simon and Etsuji Tomita, Eds.), Lecture Notes in Artificial Intelligence 3734, pp. 386 - 398, Springer 2005.

Abstract. We describe and analyze an algorithm for predicting a sequence of n-dimensional binary vectors based on a set of experts making vector predictions in [0,1]n. We measure the loss of individual predictions by the 2-norm between the actual outcome vector and the prediction. The loss of an expert is then the sum of the losses experienced on individual trials. We obtain bounds for the loss of our expert algorithm in terms of the loss of the best expert analogous to the well-known results for scalar experts making real-valued predictions of a binary outcome.

©Copyright 2005, Springer