Structured sparsity in structured prediction
; Smith, N. A.
; Aguiar, P.
Figueiredo, M. A. T.
Structured sparsity in structured prediction, Proc Empirical Methods in Language Processing - EMNLP, Edinburgh, United Kingdom, Vol. , pp. - , July, 2011.
Digital Object Identifier:
Linear models have enjoyed great success in structured prediction in NLP. While a lot of progress has been made on efﬁcient training with several loss functions, the problem of endowing learners with a mechanism for feature selection is still unsolved. Common approaches employ ad hoc ﬁltering or L1 regularization; both ignore the structure of the feature space, preventing practicioners from encoding structural prior knowledge. We ﬁll this gap by adopting regularizers that promote structured sparsity, along with efﬁcient algorithms to handle them. Experiments on three tasks (chunking, entity recognition, and dependency parsing) show gains in performance, compactness, and model interpretability.