Contextual Text Denoising with Masked Language Models

arXiv:1910.14080v2 Announce Type: replace-cross
Abstract: Recently, with the help of deep learning models, significant advances have been made in different Natural Language Processing (NLP) tasks. Unfortunately, state-of-the-art models are vulnerable to noisy texts. We propose a new contextual text denoising algorithm based on the ready-to-use masked language model. The proposed algorithm does not require retraining of the model and can be integrated into any NLP system without additional training on paired cleaning training data. We evaluate our method under synthetic …

abstract advances algorithm art arxiv cs.cl cs.lg deep learning denoising language language model language models language processing natural natural language natural language processing nlp processing state state-of-the-art models tasks text type vulnerable