PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

MKPM: a Multiclass extension of the Kernel Projection Machine
Sylvain Takerkart and Liva Ralaivola
In: CVPR 2011, 20-25 Jun 2011, Colorado Springs, USA.


We introduce Multiclass Kernel Projection Machines (MKPM), a new formalism that extends the Kernel Projection Machine framework to the multiclass case. Our formulation is based on the use of output codes and it implements a co-regularization scheme by simultaneously constraining the projection dimensions associated with the individual predictors that constitue the global classifier. In order to solve the optimization problem posed by our formulation, we propose an efficient dynamic programming approach. Numerical simulations conducted on a few pattern recognition problems illustrate the soundness of our approach.

EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Theory & Algorithms
ID Code:7708
Deposited By:Liva Ralaivola
Deposited On:17 March 2011