Spelling Error Correction Using a Nested RNN Model and Pseudo Training Data

Abstract

We propose a nested recurrent neural network (nested RNN) model for English spelling error correction and generate pseudo data based on phonetic similarity to train it. The model fuses orthographic information and context as a whole and is trained in an end-to-end fashion. This avoids feature engineering and does not rely on a noisy channel model as in traditional methods. Experiments show that the proposed method is superior to existing systems in correcting spelling errors.

6 pages, 1 figure

References in corpus (2)

• Character-Aware Neural Language Models
• A Multilayer Convolutional Encoder-Decoder Neural Network for Grammatical Error Correction

Cited by in corpus (5)

• Combating Adversarial Misspellings with Robust Word Recognition
• GitHub Typo Corpus: A Large-Scale Multilingual Dataset of Misspellings and Grammatical Errors
• Misspelling Correction with Pre-trained Contextual Language Model
• Comparison of Interactive Knowledge Base Spelling Correction Models for Low-Resource Languages
• Context-aware Stand-alone Neural Spelling Correction