Inducing history representations for broad coverage statistical parsing

James Henderson
2003 Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology - NAACL '03   unpublished
We present a neural network method for inducing representations of parse histories and using these history representations to estimate the probabilities needed by a statistical left-corner parser. The resulting statistical parser achieves performance (89.1% F-measure) on the Penn Treebank which is only 0.6% below the best current parser for this task, despite using a smaller vocabulary size and less prior linguistic knowledge. Crucial to this success is the use of structurally determined soft
more » ... y determined soft biases in inducing the representation of the parse history, and no use of hard independence assumptions. Estimating the Parameters of the Probability Model The parsing system we propose consists of two components, one which estimates the parameters of a probability model for phrase structure trees, and one which searches for the most probable phrase structure tree given these parameters. The probability model we use is generative and history-based. At each step, the model's stochastic process generates a characteristic of the tree Edmonton,
doi:10.3115/1073445.1073459 fatcat:actelhb53zarff2qnxr7rtz4ka