Multilingual AMR Parsing with Noisy Knowledge Distillation

Deng Cai, Xin Li, Jackie Chun-Sing Ho, Lidong Bing, Wai Lam


Abstract
We study multilingual AMR parsing from the perspective of knowledge distillation, where the aim is to learn and improve a multilingual AMR parser by using an existing English parser as its teacher. We constrain our exploration in a strict multilingual setting: there is but one model to parse all different languages including English. We identify that noisy input and precise output are the key to successful distillation. Together with extensive pre-training, we obtain an AMR parser whose performances surpass all previously published results on four different foreign languages, including German, Spanish, Italian, and Chinese, by large margins (up to 18.8 Smatch points on Chinese and on average 11.3 Smatch points). Our parser also achieves comparable performance on English to the latest state-of-the-art English-only parser.
Anthology ID:
2021.findings-emnlp.237
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2021
Month:
November
Year:
2021
Address:
Punta Cana, Dominican Republic
Editors:
Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
Venue:
Findings
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
2778–2789
Language:
URL:
https://aclanthology.org/2021.findings-emnlp.237
DOI:
10.18653/v1/2021.findings-emnlp.237
Bibkey:
Cite (ACL):
Deng Cai, Xin Li, Jackie Chun-Sing Ho, Lidong Bing, and Wai Lam. 2021. Multilingual AMR Parsing with Noisy Knowledge Distillation. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 2778–2789, Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Multilingual AMR Parsing with Noisy Knowledge Distillation (Cai et al., Findings 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.findings-emnlp.237.pdf
Code
 jcyk/xamr
Data
OPUS-MT