Bootstrapping Multilingual AMR with Contextual Word Alignments
Computation and Language
2022-05-09 v1 Artificial Intelligence
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.
Keywords
Cite
@article{arxiv.2102.02189,
title = {Bootstrapping Multilingual AMR with Contextual Word Alignments},
author = {Janaki Sheth and Young-Suk Lee and Ramon Fernandez Astudillo and Tahira Naseem and Radu Florian and Salim Roukos and Todd Ward},
journal= {arXiv preprint arXiv:2102.02189},
year = {2022}
}