Paper: Temporal Restricted Boltzmann Machines for Dependency Parsing

ACL ID P11-2003
Title Temporal Restricted Boltzmann Machines for Dependency Parsing
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2011
Authors

We propose a generative model based on Temporal Restricted Boltzmann Machines for transition based dependency parsing. The parse tree is built incrementally using a shift- reduce parse and an RBM is used to model each decision step. The RBM at the current time step induces latent features with the help of temporal connections to the relevant previ- ous steps which provide context information. Our parser achieves labeled and unlabeled at- tachment scores of 88.72% and 91.65% re- spectively, which compare well with similar previous models and the state-of-the-art.