Research map / Large language models / Alignment and preferences
Balancing Quality and Human Involvement: An Effective Approach to Interactive 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 |
|---|---|---|---|
| Train joint query/translate model with RL instead of confidence-threshold protocol | Interactive NMT (Transformer) · IWSLT14 De-En real IMT | BLEU: 46.06 -> 49.33 | Helped |
| Replace confidence baseline with RL approach at equal query frequency | Interactive NMT · IWSLT14 De-En simulated | BLEU: +2 BLEU on average | Helped |
| Replace confidence baseline with RL approach | Interactive NMT · IWSLT Zh-En simulated | BLEU: +1.1 BLEU on average | Helped |
| Replace confidence baseline with RL approach | Interactive NMT · IWSLT17 Fr-En simulated | BLEU: +1.7 BLEU on average | Helped |
| Add RL training on top of pretrained joint model (ablation) | Joint INMT model · IWSLT14 De-En simulated | BLEU | Helped |
| Add MLE auxiliary losses and 3:1 critic-actor asynchronous updates to actor-critic (ablation) | RL training · IWSLT De-En | training stability | Helped |
| Use RL-trained policy vs transformer alone on queried tokens like 'up' | RL approach · IWSLT14 De-En frequent tokens | recall: 0.237 -> 0.502 | Helped |
Closest papers by meaning
- Analyzing Uncertainty in Neural Machine Translation (ICML 2018 · similarity 0.88)
- Improving Simultaneous Machine Translation with Monolingual Data (AAAI 2023 · similarity 0.88)
- SimulPL: Aligning Human Preferences in Simultaneous Machine Translation (ICLR 2025 · similarity 0.87)
- A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models (ICLR 2024 · similarity 0.87)
- MT-Ranker: Reference-free machine translation evaluation by inter-system ranking (ICLR 2024 · similarity 0.87)
- Calibrating Translation Decoding with Quality Estimation on LLMs (NeurIPS 2025 · 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.