Bayesian Similarity (Extended Abstract)

Thomas Breuel
Proceedings of the 2003 Snowbird Workshop on Learning, ?, 2003

Abstract:

Judging similarity is a key human skill and at the heart of many pattern recognition methods: good models of similarity allow us to extrapolate to novel data from fewer samples. Similarity has been studied extensively in areas such as pattern recognition and adaptive nearest neighbors [10, 6, 1], character recognition, text-based information retrieval [4, 7], and case-based reasoning [5]. Closely related to models of similarity is the problem of multitask learning [3], where we attempt to use information learned from one task to improve performance on a related task. In this talk, I explore the relationship between Bayes optimal classification, similarity measures, and multitask learning. The key for making the connection is the conditional distribution P (S |x, x′ ), where S is a boolean variable indicating whether samples x and x′ come from the same class.

Files:

  BayesianSimilarity.pdf

BibTex:

@inproceedings{ BREU2003,
	Title = {Bayesian Similarity (Extended Abstract)},
	Author = {Thomas Breuel},
	BookTitle = {Proceedings of the 2003 Snowbird Workshop on Learning},
	Year = {2003},
	Publisher = {?}
}

     
Last modified:: 30.08.2016