42 papers with code • 0 benchmarks • 4 datasets
Spelling correction is the task of detecting and correcting spelling mistakes.
These leaderboards are used to track progress in Spelling Correction
We show that MoNoise beats the state-of-the-art on different normalization benchmarks for English and Dutch, which all define the task of normalization slightly different.
We identify three key ingredients of high-quality tokenization repair, all missing from previous work: deep language models with a bidirectional component, training the models on text with spelling errors, and making use of the space information already present.
We extend a current sequence-tagging approach to Grammatical Error Correction (GEC) by introducing specialised tags for spelling correction and morphological inflection using the SymSpell and LemmInflect algorithms.
However, we note a critical flaw in the process of tagging one character to another, that the correction is excessively conditioned on the error.
Inspired by the findings from the Cmabrigde Uinervtisy effect, we propose a word recognition model based on a semi-character level recurrent neural network (scRNN).