Lecture 20: Bayesian Interpretations
Tomaso Poggio, Sayan Mukherjee

Description

This class focuses on a Bayesian interpretation of regularization, both in the case of the quadratic loss function and in the case of the SVM loss function. We discuss a Bayesian interpretation of the regularizer and of the various loss functions. We conclude with a critical discussion of techniques for synthesizing or selecting the kernel from a certain ensemble of data.

Slides

Slides for this lecture: PS, PDF.

Suggested Reading

Evgeniou, Pontil, and Poggio. Regularization Networks and Support Vector Machines. Advances in Computational Mathematics, 2000.