Statistical Learning: Stability is Sufficient for Generalization and Necessary and Sufficient for Consistency of Empirical Risk Minimization
Author(s)Mukherjee, Sayan; Niyogi, Partha; Poggio, Tomaso; Rifkin, Ryan
MetadataShow full item record
Solutions of learning problems by Empirical Risk Minimization (ERM) need to be consistent, so that they may be predictive. They also need to be well-posed, so that they can be used robustly. We show that a statistical form of well-posedness, defined in terms of the key property of L-stability, is necessary and sufficient for consistency of ERM.
revised July 2003
AI, Theory of Learning, Great Discoveries, Consistency, ERM, Stability