期刊名称：Conference on European Chapter of the Association for Computational Linguistics (EACL)
出版年度：2021
卷号：2021
页码：394-404
DOI：10.18653/v1/2021.eacl-main.30
语种：English
出版社：ACL Anthology
摘要：We develop high performance multilingual Abstract Meaning Representation (AMR) systems by projecting English AMR annotations to other languages with weak supervision. We achieve this goal by bootstrapping transformer-based multilingual word embeddings, in particular those from cross-lingual RoBERTa (XLM-R large). We develop a novel technique for foreign-text-to-English AMR alignment, using the contextual word alignment between English and foreign language tokens. This word alignment is weakly supervised and relies on the contextualized XLM-R word embeddings. We achieve a highly competitive performance that surpasses the best published results for German, Italian, Spanish and Chinese.
