Visual Cues and Error Correction for Translation Robustness
This addresses robustness issues in machine translation for noisy human-generated inputs, but it is incremental as it builds on existing techniques.
The paper tackled the problem of neural machine translation models being sensitive to input noise like misspellings, by introducing visual context and an error correction training regime, resulting in improved robustness to noisy texts while maintaining translation quality on clean texts in English-French and English-German experiments.
Neural Machine Translation models are sensitive to noise in the input texts, such as misspelled words and ungrammatical constructions. Existing robustness techniques generally fail when faced with unseen types of noise and their performance degrades on clean texts. In this paper, we focus on three types of realistic noise that are commonly generated by humans and introduce the idea of visual context to improve translation robustness for noisy texts. In addition, we describe a novel error correction training regime that can be used as an auxiliary task to further improve translation robustness. Experiments on English-French and English-German translation show that both multimodal and error correction components improve model robustness to noisy texts, while still retaining translation quality on clean texts.