Abstract:

The course will cover the following topics: 
(1) Learning features: A structured model. Introduction to clustering and features extraction. Learning a model using regression and logit regression. Sparsity in high dimension.

(2) Automatic representation learning with Non­Negative Matrix Implementation. Matrix factorization and SVD methods. NMF. Completion using NMF and recommendation.

(3) Using spark, hadoop. Principle of map reduce. Divide and conquer method. Stochastic gradient optimization.

(4) Application to movie lens recommendation. Admission is limited to 24 students (contact: tasio@eio.uva.es)