A Benchmark Corpus of English Misspellings and a Minimally-supervised Model for Spelling Correction

摘要

Spelling correction has attracted a lot of attention in the NLP community. However, models have been usually evaluated on artificially-created or proprietary corpora. A publicly-available corpus of authentic misspellings, annotated in context, is still lacking. To address this, we present and release an annotated data set of 6,121 spelling errors in context, based on a corpus of essays written by English language learners. We also develop a minimally-supervised context-aware approach to spelling correction. It achieves strong results on our data: 88.12% accuracy. This approach can also train with a minimal amount of annotated data (performance reduced by less than 9c). Furthermore, this approach allows easy portability to new domains. We evaluate our model on data from a medical domain and demonstrate that it rivals the performance of a model trained and tuned on in-domain data.