Dependency Parsing with Dynamic Bayesian Network

TL;DR

This paper introduces a Dynamic Bayesian Network approach to dependency parsing, enabling local classification methods to uncover syntactic structures in English sentences, demonstrated effectively on Wall Street Journal data.

Contribution

It presents a novel application of Dynamic Bayesian Networks for dependency parsing, addressing non-locality issues in traditional automata-based methods.

Findings

Successfully learned syntactic dependencies from labeled data

Effective on Wall Street Journal corpus

Improves parsing accuracy with probabilistic modeling

Abstract

Exact parsing with finite state automata is deemed inappropriate because of the unbounded non-locality languages overwhelmingly exhibit. We propose a way to structure the parsing task in order to make it amenable to local classification methods. This allows us to build a Dynamic Bayesian Network which uncovers the syntactic dependency structure of English sentences. Experiments with the Wall Street Journal demonstrate that the model successfully learns from labeled data.