A Consistent Regularization Approach for Structured Prediction

Part of Advances in Neural Information Processing Systems 29 (NIPS 2016)

Bibtex Metadata Paper Reviews Supplemental

Authors

Carlo Ciliberto, Lorenzo Rosasco, Alessandro Rudi

Abstract

We propose and analyze a regularization approach for structured prediction problems. We characterize a large class of loss functions that allows to naturally embed structured outputs in a linear space. We exploit this fact to design learning algorithms using a surrogate loss approach and regularization techniques. We prove universal consistency and finite sample bounds characterizing the generalization properties of the proposed method. Experimental results are provided to demonstrate the practical usefulness of the proposed approach.