Low-Resource Parsing with Crosslingual Contextualized Representations
This addresses the challenge of dependency parsing for languages with small or nonexistent treebanks, which is an incremental improvement in crosslingual NLP methods.
The paper tackles the problem of dependency parsing for low-resource languages by using multilingual contextual word representations to enable crosslingual transfer from high-resource to low-resource languages, showing that this approach greatly facilitates parsing without needing crosslingual supervision like dictionaries or parallel text.
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.