Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published March 2000 | Published
Journal Article Open

No Free Lunch for Noise Prediction

Abstract

No-free-lunch theorems have shown that learning algorithms cannot be universally good. We show that no free funch exists for noise prediction as well. We show that when the noise is additive and the prior over target functions is uniform, a prior on the noise distribution cannot be updated, in the Bayesian sense, from any finite data set. We emphasize the importance of a prior over the target function in order to justify superior performance for learning systems.

Additional Information

© 2000 Massachusetts Institute of Technology. Received May 29, 1998; accepted March 31, 1999. Posted Online March 13, 2006. We thank Yaser Abu-Mostafa and the members of the Caltech Learning Systems Group for helpful comments. In addition, two anonymous referees provided useful comments.

Attached Files

Published - MAGnc00a.pdf

Files

MAGnc00a.pdf
Files (2.6 MB)
Name Size Download all
md5:e38136b3d39d6a9195137ce58177d213
2.6 MB Preview Download

Additional details

Created:
August 19, 2023
Modified:
October 24, 2023