Paper: Third-order Variational Reranking on Packed-Shared Dependency Forests

ACL ID D11-1137
Title Third-order Variational Reranking on Packed-Shared Dependency Forests
Venue Conference on Empirical Methods in Natural Language Processing
Session Main Conference
Year 2011
Authors

Weproposeanovel forest reranking algorithm for discriminative dependency parsing based on a variant of Eisner’s generative model. In our framework, we define two kinds of gener- ativemodelforreranking. Oneislearnedfrom training data offline and the other from a for- est generated by a baseline parser on the fly. The final prediction in the reranking stage is performed using linear interpolation of these models and discriminative model. In order to efficiently train the model from and decode on a hypergraph data structure representing a forest, we apply extended inside/outside and Viterbi algorithms. Experimental results show that our proposed forest reranking algorithm achieves significant improvement when com- pared with conventional approaches.