Dual Decomposition with Many Overlapping Components
; Smith, N. A.
Figueiredo, M. A. T.
; Aguiar, P.
Dual Decomposition with Many Overlapping Components, Proc Empirical Methods in Language Processing - EMNLP, Edinburgh, United Kingdom, Vol. , pp. - , July, 2011.
Digital Object Identifier:
Dual decomposition has been recently proposed as a way of combining complementary models, with a boost in predictive power. However, in cases where lightweight decompositions are not readily available (e.g., due to the presence of rich features or logical constraints), the original subgradient algorithm is inefﬁcient. We sidestep that difﬁculty by adopting an augmented Lagrangian method that accelerates model consensus by regularizing towards the averaged votes. We show how ﬁrst-order logical constraints can be handled efﬁciently, even though the corresponding subproblems are no longer combinatorial, and report experiments in dependency parsing, with state-of-the-art results.