Towards Zero-Shot Multilingual Synthetic Question and Answer Generation for Cross-Lingual Reading Comprehension
Siamak Shakeri Google Research、Noah Constant Google Research \ANDMihir Sanjay Kale Google Research、Linting Xue Google Research Corresponding author.
Abstract
We propose a simple method to generate multilingual question and answer pairs on a large scale through the use of a single generative model. These synthetic samples can be used to improve the zero-shot performance of multilingual QA models on target languages. Our proposed multi-task training of the generative model only requires the labeled training samples in English, thus removing the need for such samples in the target languages, making it applicable to far more languages than those with labeled data. Human evaluations indicate the majority of such samples are grammatically correct and sensible. Experimental results show our proposed approach can achieve large gains on the XQuAD dataset, reducing the gap between zero-shot and supervised performance of smaller QA models on various languages.
中文速览
多语言问答数据严重匮乏，导致阅读理解模型在英语以外的语言上表现远落后于有监督训练水平。研究者提出用一个多语言生成模型（基于mT5）自动批量生成目标语言的问答对：训练时只需英语标注数据（SQuAD），同时加入多语言掩码语言模型（MLM）任务进行多任务微调，防止模型"遗忘"多语言能力，从而让它在推理阶段直接对非英语文章生成对应语言的问题和答案。生成的数据再经过"回环过滤"（round-trip filtering）去噪后，用来扩充目标语言的训练集，在XQuAD基准上最高提升9个F1绝对点，显著缩小零样本与有监督之间的差距。这项工作的价值在于整个流程对目标语言完全零样本，无需任何人工标注或翻译资源，理论上可推广到任何多语言模型支持的语言，为资源匮乏语言的问答系统开发提供了可行路径。
原文 arXiv:2010.12008;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2010.12008v3