28th November 2003 - Boosting, Maximum Likelihood and Bregman divergence
Note that this is probably the last journal club meeting of 2003.
Shiro and Fernando will talk about the EM-like geometrical structure behind the Adaboost, which gives a proof of convergence with a related cost function. The main paper we will discuss is:
They will also present on generalization of Adaboost based on the following two papers:
- Information Geometry of U-Boost and Bregman Divergence by N. Murata, T. Takenouchi, T. Kanamori, and S. Eguchi
- Logistic Regression, AdaBoost and Bregman Distances by M. Collins, R. E. Schapire, and Y. Singer