Machine Learning Day

BMM Summer Course 2017

Marine Biological Laboratory, Woods Hole, MA

Sat. 19 Aug. 2017
Theory: Morning (9:15 - 12:00), Speck Auditorium
Practice/Labs: Afternoon (1:30 - 4:30), Loeb 306

Instructor: Lorenzo Rosasco (lrosasco at
Universita' di Genova, Istituto Italiano di Tecnologia, MIT
TAs: Georgios Evangelopoulos (gevang at

Course Description

Learning, its principles and computational implementations, is at the very core of intelligence. Machine Learning (ML) is the key to developing intelligent systems and analyzing data in science and engineering. ML engines enable intelligent technologies such as Alexa, Siri, Cortana, Google Now, Watson, AplhaGo, or self driving cars, to name a few. At the same time ML methods help make sense of the flood of online or biological data, forming the basis of a new Science of Data.

This one day course provides an introduction to essential concepts and algorithms at the core of ML. Theory classes in the morning are complemented by hands-on lab sessions in the afternoon.


Time Session Material
9:00 Theory Local Methods, Bias-Variance and Model Selection slides
Theory Regularization: Linear and Kernel Least Squares slides
Theory Variable Selection and Dimensionality Reduction slides
1:30 (Optional) MATLAB Warm-Up: Data Generation code
Practice k-NN and Cross-validation code
Practice Regularized Least Squares (RLS) code
Practice Kernel RLS code/data
Practice PCA and Orthogonal Matching Pursuit (OMP) code
(Optional) Learning with Real Data code/data


Basic Probability, Calculus, Linear Algebra


  • L. Rosasco, Introductory Machine Learning Notes, Draft, 2016 (pdf).
  • T. Hastie, R. Tibshirani and J. Friedman, The Elements of Statistical Learning, 2nd Ed., 2009 (pdf available on authors' website).

Further reading

  • T. Poggio and S. Smale, Mathematics of Learning: Dealing with Data, Notices of the AMS, 2003 (pdf).
  • P. Domingos, A few useful things to know about Machine Learning, Comm. ACM, 55 (10), 2012 (pdf).