Randomized Dimensionality Reduction in Machine Learning
- Christos Boutsidis | T. J. Watson Research Center
We show how certain random projections and random sampling methods can be used to design efficient dimensionality reduction techniques for two popular machine learning problems: (i) K-means Clustering, and (ii) Canonical Correlation Analysis. In both cases, we argue that randomized dimensionality reduction is provably efficient.
Speaker Details
Christos Boutsidis joined the Business Analytics and Mathematical Sciences Department of IBM Research in August 2011, shortly after receiving his Ph.D. in Computer Science from Rensselaer Polytechnic Institute. He belongs to the Predictive Modeling and Optimization group, managed by Aleksandra Mojsilovic. The mission of our group is to conduct leading-edge research in probability theory, applied math, statistical modeling, and machine learning, and to apply this expertise to solve challenging problems in business analytics for IBM and its clients.
-
-
Jeff Running
-
-
