NIPS2009_0135_slide - Analysis of SVM with Indefinite...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Analysis of SVM with Indefinite Kernels Y. Ying, C. Campbell and M. Girolami SVM classification with non-positive semi-definite (indefinite) kernel (min-max problem) [Luss & d’Aspremont, NIPS’07] Approximate approaches (smoothing techniques) PGM and ACCPM [Luss & d’Aspremont, NIPS’07] Semi-infinite quadratically constrained linear programming (SIQCLP) [Chen & Ye, ICML’08] Our Contribution We prove that the objective function is, indeed, differentiable with Lipschitz continuous gradient We successfully apply Nesterov’s smoothing optimization approach to efficiently solve SVM with indefinite kernels ...
View Full Document

Ask a homework question - tutors are online