AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Cross-model Back-translated Distillation for Unsupervised Machine Translation

Xuan-Phi Nguyen, Shafiq Joty, Thanh-Tung Nguyen, Kui Wu, Ai Ti Aw

ICML 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 12 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add cross-model back-translated distillation to XLMXLM · Base WMT, En-Fr, Fr-En, En-De, De-En, En-Ro, Ro-EnBLEU: 33.0 -> 35.4; 31.5 -> 33.0; 23.9 -> 26.1; 29.3 -> 31.5; 30.6 -> 32.2; 27.9 -> 29.2Improved
Add cross-model back-translated distillation to NMTNMT · Base WMT, En-Fr, Fr-En, En-De, De-En, En-Ro, Ro-EnBLEU: 24.7 -> 26.6; 24.5 -> 25.7; 14.5 -> 16.6; 18.2 -> 20.5; 16.7 -> 18.1; 16.3 -> 17.8Improved
Add cross-model back-translated distillation to PBSMTPBSMT · Base WMT, En-Fr, Fr-En, En-De, De-En, En-Ro, Ro-EnBLEU: 17.1 -> 21.6; 16.4 -> 20.6; 10.9 -> 15.0; 13.6 -> 17.7; 10.5 -> 11.3; 11.7 -> 14.5Improved
Add CBD to NMTNMT · IWSLT En-Fr and En-De, four translation directionsBLEU: 29.6 -> 31.8; 30.7 -> 31.8; 15.8 -> 18.4; 19.1 -> 21.7Improved
Use one agent for one-way synthetic data rather than cross-model back-translation (ablation)CBD · Base WMT En-FrBLEU: 26.6 -> 24.5Hurt
Use two agents for separate forward translations without cross-model back-translation (ablation)CBD · Base WMT En-FrBLEU: 26.6 -> 24.6Hurt
Use the same agent for forward and backward translation (ablation)CBD · Base WMT En-FrBLEU: 26.6 -> 24.8Hurt
Increase the number of agents from 2 to 3 (ablation)GCBD · IWSLT En-Fr, Fr-En, En-De, De-EnBLEU: 31.8 -> 32.8; 31.8 -> 32.1; 18.4 -> 19.2; 21.7 -> 22.2Improved
Increase the number of agents from 3 to 4 (ablation)GCBD · IWSLT En-Fr, Fr-En, En-De, De-EnBLEU: 32.8 -> 32.3; 32.1 -> 32.0; 19.2 -> 19.1; 22.2 -> 21.9Hurt
Replace CBD with ensemble knowledge distillationUnsupervised MT with NMT agents · Base WMT En-Fr, Fr-En, En-De, De-En, En-Ro, Ro-EnBLEU: CBD: 26.6, 25.7, 16.6, 20.5, 18.1, 17.8; Ens-Distil: 17.3, 20.0, 3.5, 3.7, 1.2, 1.1Hurt
Use temperature sampling instead of deterministic CBDXLM · Base WMT En-Fr, Fr-En, En-De, De-EnBLEU: 33.0 -> 33.5; 31.5 -> 32.2; 23.9 -> 24.3; 29.3 -> 30.2Mixed
Use target noising during iterative back-translationXLM · Base WMT En-Fr, Fr-En, En-De, De-EnBLEU: 33.0 -> 32.8; 31.5 -> 30.7; 23.9 -> 24.0; 29.3 -> 29.6Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy