Mining Big Data Using Kernel

Dr. Kai Zhang
Siemens Corporate Research


Knowledge discovery from big data has dominated modern science discovery in mang areas. In this talk, I will focus on scaling up kernel methods in large scale unsupervised, semi-supervised, and supervised learning, as well as kernel learning problems. The main theme is to use matrix low-rank approximation as a basic building block both in alleviating the computational burdens and to design novel computational models for various learning paradigms. I will also talk about sparse learning in in bioinformatics applications as well as topics in complex networks if time allows.