On Robustness Properties of Convex Risk Minimization Methods for Pattern Recognition

The paper brings together methods from two disciplines: machine learning theory and robust statistics. We argue that robustness is an important aspect and we show that many existing machine learnin...