Research map / Architectures and efficiency / Attention and transformers
Search Engine Guided Neural 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 |
|---|---|---|---|
| add search-engine retrieved translation pairs with key-value memory and shallow fusion | attention NMT · JRC-Acquis Fr→En test | BLEU: 60.11 -> 65.11 | Helped |
| add retrieval-guided SEG-NMT | attention NMT · JRC-Acquis En→De test | BLEU: 43.98 -> 48.80 | Helped |
| add retrieval-guided SEG-NMT | attention NMT · JRC-Acquis En→Es test | BLEU: 50.48 -> 57.27 | Helped |
| key-value memory instead of CopyNet copying over retrieved target | retrieval-augmented NMT · JRC-Acquis Fr→En test | BLEU: 62.02 -> 65.11 | Helped |
| shallow fusion instead of deep fusion (ablation) | SEG-NMT · JRC-Acquis En-Fr dev | BLEU gain over NMT: +1.30 -> +5.21 | Helped |
| increase retrieved pairs beyond about four (ablation) | SEG-NMT · JRC-Acquis Fr→En | BLEU | Hurt |
| adaptive greedy coverage selection of retrieved pairs (ablation) | SEG-NMT · JRC-Acquis Fr→En | BLEU | Helped |
| higher fuzzy-match relevance of retrieved pair | SEG-NMT · JRC-Acquis Fr→En dev | BLEU gain | Helped |
Closest papers by meaning
- Neural Machine Translation Advised by Statistical Machine Translation (AAAI 2017 · similarity 0.88)
- Improved Neural Machine Translation with SMT Features (AAAI 2016 · similarity 0.87)
- Syntax-Directed Attention for Neural Machine Translation (AAAI 2018 · similarity 0.87)
- Joint Training for Neural Machine Translation Models with Monolingual Data (AAAI 2018 · similarity 0.86)
- Translating with Bilingual Topic Knowledge for Neural Machine Translation (AAAI 2019 · similarity 0.86)
- Acquiring Knowledge from Pre-Trained Model to Neural Machine Translation (AAAI 2020 · similarity 0.86)
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.