Research map / Label-efficient and robust learning / Recognition and long-tail
Dual Learning for Machine Translation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Train with dual-NMT instead of NMT | NMT · En→Fr (Large), newstest2014 | BLEU: 29.92 -> 32.06 | Helped |
| Train with dual-NMT instead of NMT | NMT · Fr→En (Large), newstest2014 | BLEU: 27.49 -> 29.78 | Helped |
| Train with dual-NMT instead of NMT | NMT · En→Fr (Small), newstest2014 | BLEU: 25.32 -> 28.73 | Helped |
| Train with dual-NMT instead of NMT | NMT · Fr→En (Small), newstest2014 | BLEU: 22.27 -> 27.50 | Helped |
| Train with dual-NMT instead of pseudo-NMT | pseudo-NMT · En→Fr (Large), newstest2014 | BLEU: 30.40 -> 32.06 | Helped |
| Train with dual-NMT instead of pseudo-NMT | pseudo-NMT · Fr→En (Large), newstest2014 | BLEU: 27.66 -> 29.78 | Helped |
| Train with dual-NMT instead of pseudo-NMT | pseudo-NMT · En→Fr (Small), newstest2014 | BLEU: 25.63 -> 28.73 | Helped |
| Train with dual-NMT instead of pseudo-NMT | pseudo-NMT · Fr→En (Small), newstest2014 | BLEU: 23.24 -> 27.50 | Helped |
| Use dual-NMT rather than NMT for round-trip translation | NMT · En→Fr→En (Small), test set | BLEU: 28.28 -> 48.94 | Helped |
| Use dual-NMT rather than NMT for round-trip translation | NMT · Fr→En→Fr (Small), test set | BLEU: 32.63 -> 50.38 | Helped |
| Use 10% rather than all bilingual data for warm start | dual-NMT · Fr→En, newstest2014 | BLEU: 29.78 -> 27.50 | Mixed |
Closest papers by meaning
- Dual Transfer Learning for Neural Machine Translation with Marginal Distribution Regularization (AAAI 2018 · similarity 0.92)
- Multi-Agent Dual Learning (ICLR 2019 · similarity 0.90)
- Model-Level Dual Learning (ICML 2018 · similarity 0.90)
- Unsupervised Neural Machine Translation (ICLR 2018 · similarity 0.89)
- Improving Simultaneous Machine Translation with Monolingual Data (AAAI 2023 · similarity 0.89)
- A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models (ICLR 2024 · similarity 0.87)
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.