Bootstrapping Multilingual AMR with Contextual Word Alignments

TL;DR

This paper presents a method for creating multilingual AMR systems by projecting English AMR annotations onto other languages using weakly supervised, context-aware word alignments based on cross-lingual RoBERTa embeddings, achieving state-of-the-art results.

Contribution

Introduces a novel weakly supervised alignment technique leveraging contextual embeddings from XLM-R for multilingual AMR projection, surpassing previous benchmarks.

Findings

Outperforms existing methods for German, Italian, Spanish, and Chinese AMR tasks.

Uses cross-lingual RoBERTa embeddings for effective alignment.

Achieves highly competitive results in multilingual AMR parsing.

Abstract

We develop high performance multilingualAbstract Meaning Representation (AMR) sys-tems by projecting English AMR annotationsto other languages with weak supervision. Weachieve this goal by bootstrapping transformer-based multilingual word embeddings, in partic-ular those from cross-lingual RoBERTa (XLM-R large). We develop a novel technique forforeign-text-to-English AMR alignment, usingthe contextual word alignment between En-glish and foreign language tokens. This wordalignment is weakly supervised and relies onthe contextualized XLM-R word embeddings.We achieve a highly competitive performancethat surpasses the best published results forGerman, Italian, Spanish and Chinese.