Headed-Span-Based Projective Dependency Parsing

TL;DR

This paper introduces a headed-span-based approach for projective dependency parsing, utilizing a novel dynamic programming algorithm to achieve state-of-the-art results across multiple datasets.

Contribution

It presents a new headed span representation and an $O(n^3)$ dynamic programming algorithm for efficient global training and exact inference in dependency parsing.

Findings

Achieves state-of-the-art results on PTB, CTB, and UD datasets.

Introduces a novel headed span representation for dependency trees.

Develops an efficient $O(n^3)$ dynamic programming algorithm.

Abstract

We propose a new method for projective dependency parsing based on headed spans. In a projective dependency tree, the largest subtree rooted at each word covers a contiguous sequence (i.e., a span) in the surface order. We call such a span marked by a root word \textit{headed span}. A projective dependency tree can be represented as a collection of headed spans. We decompose the score of a dependency tree into the scores of the headed spans and design a novel $O(n^3)$ dynamic programming algorithm to enable global training and exact inference. Our model achieves state-of-the-art or competitive results on PTB, CTB, and UD. Our code is publicly available at \url{https://github.com/sustcsonglin/span-based-dependency-parsing}.