arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1909.13302cs.CL
收藏

一种用于语法纠错的可控数据综合方法

Controllable Data Synthesis Method for Grammatical Error Correction

Beijing Language and Culture University(北京语言大学);Beijing University of Technology(北京工业大学);Shanghai University of Finance and Economics(上海财经大学)

Liner Yang, Chencheng Wang, Yun Chen, Yongping Du, Erhong Yang

摘要:

由于目前的语法错误纠正(GEC)任务中缺乏并行数据,基于序列到序列框架的模型不能得到充分的训练以获得更高的性能。本文提出了两种数据合成方法,可以控制合成数据的出错率和错误类型的比例。第一种方法是对单语语料库中的每个单词进行固定概率的破坏,包括替换、插入和删除。另一种方法是训练错误生成模型并进一步过滤模型的解码结果。对不同合成数据的实验表明,出错率为40%,且错误类型比例相同,能更好地提高模型性能。最后,我们合成了大约1亿个数据,并实现了与最先进技术相当的性能,它使用的数据是我们使用的数据的两倍。

英文摘要:

Due to the lack of parallel data in current Grammatical Error Correction (GEC) task, models based on Sequence to Sequence framework cannot be adequately trained to obtain higher performance. We propose two data synthesis methods which can control the error rate and the ratio of error types on synthetic data. The first approach is to corrupt each word in the monolingual corpus with a fixed probability, including replacement, insertion and deletion. Another approach is to train error generation models and further filtering the decoding results of the models. The experiments on different synthetic data show that the error rate is 40% and the ratio of error types is the same can improve the model performance better. Finally, we synthesize about 100 million data and achieve comparable performance as the state of the art, which uses twice as much data as we use.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑