Adversarial Grammatical Error Correction | Technology Org

Adversarial Grammatical Error Correction | Technology Org

Grammatical error correction is a popular natural language processing task that creates systems for automatically correcting errors in written text.

A recent paper on proposes a generative adversarial training based grammatical error correction system. The generator is trained to rewrite a grammatically incorrect sentence into a correct one. The discriminator learns to determine if the generated sentence is a meaning-preserving and grammatically correct rewrite of the input sentence.

Image credit: Pxhere, CC0 Public Domain

During the adversarial training between the two models, the discriminator learns to distinguish if a given input is human or artificially generated, while the generator learns to provide high-quality examples capable of tricking the discriminator. Therefore, the difference between natural and artificial sentences is minimized. It is shown that the proposed framework achieves better results than baselines.

Recent works in Grammatical Error Correction (GEC) have leveraged the progress in Neural Machine Translation (NMT), to learn rewrites from parallel corpora of grammatically incorrect and corrected sentences, achieving state-of-the-art results. At the same time, Generative Adversarial Networks (GANs) have been successful in generating realistic texts across many different tasks by learning to directly minimize the difference between human-generated and synthetic text. In this work, we present an adversarial learning approach to GEC, using the generator-discriminator framework. The generator is a Transformer model, trained to produce grammatically correct sentences given grammatically incorrect ones. The discriminator is a sentence-pair classification model, trained to judge a given pair of grammatically incorrect-correct sentences on the quality of grammatical correction. We pre-train both the discriminator and the generator on parallel texts and then fine-tune them further using a policy gradient method that assigns high rewards to sentences which could be true corrections of the grammatically incorrect text. Experimental results on FCE, CoNLL-14, and BEA-19 datasets show that Adversarial-GEC can achieve competitive GEC quality compared to NMT-based baselines.