9 - 11 December 2015

Workshop on advances in the Machine Learning field of Structured Prediction.

Invited speaker: Xavier Carreras and Ariadna Quattoni: "Low-rank Matrix Learning for Compositional Objects, Strings and Trees"

Abstract: Compositional structures abound in NLP, ranging from bilexical relations, entities in context, sequences and trees.  The focus of this talk is in latent-variable compositional models for structured objects that: induce a latent n-dimensional representation for each element of the structure; and learn operators for composing such elements into structures.