Multilingual AMR Parsing with Noisy Knowledge Distillation

TL;DR

This paper introduces a multilingual AMR parser trained via noisy knowledge distillation from an English parser, achieving state-of-the-art results across multiple languages with a single model.

Contribution

It presents a novel approach to multilingual AMR parsing using noisy knowledge distillation and extensive pre-training, outperforming previous methods on several languages.

Findings

Surpassed previous results on German, Spanish, Italian, and Chinese by large margins.

Achieved comparable performance to state-of-the-art English-only parsers.

Demonstrated effectiveness of noisy input and precise output in distillation.

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 \textsc{Smatch} points on Chinese and on average 11.3 \textsc{Smatch} points). Our parser also achieves comparable performance on English to the latest state-of-the-art English-only parser.