550 - Learning Theory and Simulation

This is a theoretical course on machine learning. It will cover Lebesgue integration, weak and strong laws of large numbers, consistency, asymptotic normality, finite-time bounds, probabilistic almost correct learning and theoretical properties of learning methods based on simulation. The course is targeted at mathematically inclined students. The requirements for registering are either CPSC540 or an A+ in CPSC340.

Time: Mon 2:00-3:30pm, Wed 12:30-2:00pm

Location: ICICS 304.

Instructor: Nando de Freitas (nando at cs)

Office hours: Anytime for registered students (ICICS 183).

Recommended Books
  • All of Statistics.
  • A Probabilistic Theory of Pattern Recognition.
  • Prediction, Learning, and Games.
  • Simulation-Based Econometric Methods.

  •