Research map / Architectures and efficiency / RNNs and normalization
Modeling Fluency and Faithfulness for Diverse 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 the fluency-and-faithfulness evaluation module and its L_c guidance | Transformer · CN→EN (MT03, MT04, MT05, MT06, MT08) | BLEU: 44.74 -> 46.20; 46.27 -> 47.39; 44.16 -> 46.22; 43.29 -> 45.63; 34.72 -> 36.78 | Helped |
| Add the fluency-and-faithfulness evaluation module and its L_c guidance | Transformer · WMT16 EN→RO | BLEU: 32.85 -> 34.00 | Helped |
| Add the fluency-and-faithfulness evaluation module and its L_c guidance | Transformer · WMT14 EN→DE | BLEU: 27.21 -> 27.35 | No clear effect |
| Replace the proposed L_c loss with KL divergence | Our Method · CN→EN test sets | BLEU: 45.17 vs 46.20; 46.86 vs 47.39; 44.51 vs 45.63; 36.03 vs 36.78 | Hurt |
| Replace the proposed L_c loss with KL divergence | Our Method · WMT14 EN→DE | BLEU: 27.35 -> 27.55 | Helped |
| Add the bag-of-words loss | Transformer · CN→EN MT05 | BLEU: 44.16 -> 45.03 | Helped |
| Add faithfulness cross-attention (ablation) | Our Method · CN→EN test average | BLEU: 43.51 -> 44.44 | Helped |
| Add the L_c guiding loss (ablation) | Our Method · CN→EN test average | BLEU: 43.18 -> 44.44 | Helped |
| Add the evaluation decoder (ablation) | Transformer · CN→EN test average | BLEU: 42.63 -> 44.44 | Helped |
| Add the evaluation module and guidance training | Transformer · CN→EN test average | 1-gram accuracy: 79.10 -> 79.82 | Helped |
| Add the evaluation module and guidance training | Transformer · CN→EN test average | 2-gram accuracy: 52.72 -> 54.18 | Helped |
| Add the evaluation module and guidance training | Transformer · CN→EN test average | 3-gram accuracy: 35.34 -> 36.90 | Helped |
Closest papers by meaning
- Analyzing Uncertainty in Neural Machine Translation (ICML 2018 · similarity 0.89)
- Neural Machine Translation with Adequacy-Oriented Learning (AAAI 2019 · similarity 0.87)
- Balancing Quality and Human Involvement: An Effective Approach to Interactive Neural Machine Translation (AAAI 2020 · similarity 0.87)
- A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models (ICLR 2024 · similarity 0.86)
- Calibrating Translation Decoding with Quality Estimation on LLMs (NeurIPS 2025 · similarity 0.86)
- Quantifying the Plausibility of Context Reliance in Neural Machine Translation (ICLR 2024 · similarity 0.85)
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.