More data means less inference: A pseudo-max approach to structured learning

Part of Advances in Neural Information Processing Systems 23 (NIPS 2010)

Bibtex Metadata Paper Supplemental

Authors

David Sontag, Ofer Meshi, Amir Globerson, Tommi Jaakkola

Abstract

The problem of learning to predict structured labels is of key importance in many applications. However, for general graph structure both learning and inference in this setting are intractable. Here we show that it is possible to circumvent this difficulty when the input distribution is rich enough via a method similar in spirit to pseudo-likelihood. We show how our new method achieves consistency, and illustrate empirically that it indeed performs as well as exact methods when sufficiently large training sets are used.