Contextual Text Denoising with Masked Language Models
This addresses the problem of noisy text handling for NLP systems, offering a practical, incremental improvement.
The paper tackles the vulnerability of state-of-the-art NLP models to noisy texts by proposing a contextual text denoising algorithm based on masked language models, which improves performance in downstream tasks without requiring retraining or paired cleaning data.
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 noise and natural noise and show that the proposed algorithm can use context information to correct noise text and improve the performance of noisy inputs in several downstream tasks.