Bounding Excess Risk in Machine Learning

We will discuss a general approach to the problem of bounding the excess risk of learning algorithms based on empirical risk minimization (possibly penalized). This approach has been developed in the