Regret bounds for prediction problems
Conference Paper, Proceedings of 12th Annual Conference on Computational Learning Theory (COLT '99), pp. 29 - 40, July, 1999
Abstract
We present a unified framework for reasoning about worst-case regret bounds for learning algorithms. This framework is based on the theory of duality of convex functions. It brings together results from computational learning theory and Bayesian statistics, allowing us to derive new proofs of known theorems, new theorems about known algorithms, and new algorithms.
BibTeX
@conference{Gordon-1999-16711,author = {Geoffrey Gordon},
title = {Regret bounds for prediction problems},
booktitle = {Proceedings of 12th Annual Conference on Computational Learning Theory (COLT '99)},
year = {1999},
month = {July},
pages = {29 - 40},
}
Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.