Low-Resource Parsing with Crosslingual Contextualized Representations

TL;DR

This paper investigates how multilingual contextualized word representations can improve dependency parsing in low-resource languages by sharing parameters across languages, demonstrating significant benefits even without explicit crosslingual supervision.

Contribution

It evaluates the effectiveness of multilingual CWRs for crosslingual transfer in low-resource dependency parsing and analyzes the role of polyglot models in encoding lexical correspondence.

Findings

Multilingual CWRs significantly improve low-resource dependency parsing.

Polyglot language models better encode crosslingual lexical relations.

Shared parameters enable effective transfer without crosslingual supervision.

Abstract

Despite advances in dependency parsing, languages with small treebanks still present challenges. We assess recent approaches to multilingual contextual word representations (CWRs), and compare them for crosslingual transfer from a language with a large treebank to a language with a small or nonexistent treebank, by sharing parameters between languages in the parser itself. We experiment with a diverse selection of languages in both simulated and truly low-resource scenarios, and show that multilingual CWRs greatly facilitate low-resource dependency parsing even without crosslingual supervision such as dictionaries or parallel text. Furthermore, we examine the non-contextual part of the learned language models (which we call a "decontextual probe") to demonstrate that polyglot language models better encode crosslingual lexical correspondence compared to aligned monolingual language models. This analysis provides further evidence that polyglot training is an effective approach to crosslingual transfer.