Description
Text correction systems (e.g., spell checkers) have been used to improve the quality of computerized text by detecting and correcting errors. In this project, we investigated the usage of GECToR, predominantly used for correcting grammatical errors in well-composed and polished texts, for use in "noisy" textual environments rife with less structured language. We have found that GECToR is insufficiently robust for correcting spelling errors and word variances, particularly, suffering from overcorrections (text rewrites). Thus, we proposed a text corrector with a two-stage structure to alleviate overcorrection issues. Our method compares a rule-based error corrector with that of an attention neural error corrector with contextual attention. This novel architecture allows GECToR to produce corrections based on the erroneous text and its context without needing an end-to-end structure.