# Cross-model Back-translated Distillation for Unsupervised Machine Translation

arXiv: https://arxiv.org/abs/2006.02163 · HTML: https://mlsanity.com/paper/2006.02163/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Jun 2020.

## Built on by (most cited first, 6 of 6)

- [Unsupervised Neural Machine Translation with Generative Language Models Only](https://mlsanity.com/paper/2110.05448/) (2021)
- [Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation](https://mlsanity.com/paper/2203.08394/) (2022)
- [AUGVIC: Exploiting BiText Vicinity for Low-Resource NMT](https://mlsanity.com/paper/2106.05141/) (2021)
- [Refining Low-Resource Unsupervised Translation by Language Disentanglement of Multilingual Model](https://mlsanity.com/paper/2205.15544/) (2022)
- [CantonMT: Cantonese to English NMT Platform with Fine-Tuned Models Using Synthetic Back-Translation Data](https://mlsanity.com/paper/2403.11346/) (2024)
- [Quick Back-Translation for Unsupervised Machine Translation](https://mlsanity.com/paper/2312.00912/) (2023)

## Abstract (arXiv)

Recent unsupervised machine translation (UMT) systems usually employ three main principles: initialization, language modeling and iterative back-translation, though they may apply them differently. Crucially, iterative back-translation and denoising auto-encoding for language modeling provide data diversity to train the UMT systems. However, the gains from these diversification processes has seemed to plateau. We introduce a novel component to the standard UMT framework called Cross-model Back-translated Distillation (CBD), that is aimed to induce another level of data diversification that existing principles lack. CBD is applicable to all previous UMT approaches. In our experiments, CBD achieves the state of the art in the WMT'14 English-French, WMT'16 English-German and English-Romanian bilingual unsupervised translation tasks, with 38.2, 30.1, and 36.3 BLEU respectively. It also yields 1.5-3.3 BLEU improvements in IWSLT English-French and English-German tasks. Through extensive experimental analyses, we show that CBD is effective because it embraces data diversity while other similar variants do not.