Multi-Task Neural Models for Translating Between Styles Within and Across Languages

Abstract

Generating natural language requires conveying content in an appropriate style. We explore two related tasks on generating text of varying formality: monolingual formality transfer and formality-sensitive machine translation. We propose to solve these tasks jointly using multi-task learning, and show that our models achieve state-of-the-art performance for formality transfer and are able to perform formality-sensitive translation without being explicitly trained on style-annotated translation examples.

arXiv commentAccepted at the 27th International Conference on Computational Linguistics (COLING 2018)

References in corpus (4)

• Style Transfer from Non-Parallel Text by Cross-Alignment
• Sockeye: A Toolkit for Neural Machine Translation
• Toward Controlled Generation of Text
• Bi-Directional Neural Machine Translation with Synthetic Parallel Data

Cited by in corpus (9)

• Controlling the Output Length of Neural Machine Translation
• Multi-Task Learning in Natural Language Processing: An Overview
• Sequence-to-sequence Pre-training with Data Augmentation for Sentence Rewriting
• From Theories on Styles to their Transfer in Text: Bridging the Gap with a Hierarchical Survey
• Grammatical Error Correction and Style Transfer via Zero-shot Monolingual Translation
• Thank you BART! Rewarding Pre-Trained Models Improves Formality Style Transfer
• TextSETTR: Few-Shot Text Style Extraction and Tunable Targeted Restyling
• Improving Formality Style Transfer with Context-Aware Rule Injection
• Controlling Neural Machine Translation Formality with Synthetic Supervision