Send me an email to schedule an appointment online.
Topics in Learning Theory
Stochastic bandits [pdf]
First order methods for online convex optimization
The exponentially weighted average (EWA) forecaster
Introduction to online learning
Lectures 9 and 10:
Complement: Pinsker's inequality
Gradient descent for smooth and strongly convex functions; Accelerated gradient descent.
Gradient descent for convex functions.
Introduction to convex optimization.
Convex approach to binary classification.
Rademacher complexity and VC dimension.
Introduction to empirical risk minimization (ERM); Estimation-Approximation tradeoff; ERM with a finite class and a bounded loss; Noiseless case.
Conditional probabilities and expectation; Optimal predictors; Examples of the square and binary losses.
Introduction to supervised learning; Learning sample; Loss functions, Risk and Excess risk.