Research map / Label-efficient and robust learning / Recognition and long-tail
Towards Reliable Neural Machine Translation with Consistency-Aware Meta-Learning
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 |
|---|---|---|---|
| Apply CAML to Transformer-base | Transformer on WMT14 En→De · WMT14 En→De | BLEU: 1.11 | Helped |
| Apply CAML to Transformer-big | Transformer on WMT14 En→De · WMT14 En→De | BLEU: 0.85 | Helped |
| Combine CAML with synthetic monolingual data | Transformer on WMT14 En→De · WMT14 En→De | BLEU: 1.4/1.06 | Helped |
| Apply CAML with synthetic data | Transformer on WMT17 Zh→En · WMT17 Zh→En | BLEU: 1.01 | Helped |
| Apply CAML with synthetic data | Transformer on WMT16 Ro→En · WMT16 Ro→En | BLEU: 1.79/1.50 | Helped |
| Apply CAML to Mask-Predict | Non-autoregressive translation on En→De · WMT14 En→De | BLEU: +1.07 BLEU | Helped |
| Apply CAML to multilingual many-to-one translation | TED M2O · Related / Diverse language pairs | BLEU: 1.51/1.09 | Helped |
| Add CAML to Deep Transformer | Deep Transformer on En→De · WMT14 En→De | BLEU: 1.44 | Helped |
| Add CAML to mBART | mBART on En→De · WMT14 En→De | BLEU: 1.41 | Helped |
| Use augmented source sentences with original training data (ablation) | Transformer on WMT14 En→De · WMT14 En→De | BLEU: +0.34 | Helped |
| Use standard MAML instead of CAML (ablation) | Transformer on WMT14 En→De · WMT14 En→De | BLEU: -0.12 | Hurt |
| Use L_W and L_S with multi-task learning (ablation) | Transformer on WMT14 En→De · WMT14 En→De | BLEU: +0.48 | Helped |
Closest papers by meaning
- MetaMT, a Meta Learning Method Leveraging Multiple Domain Data for Low Resource Machine Translation (AAAI 2020 · similarity 0.91)
- Mirror-Generative Neural Machine Translation (ICLR 2020 · similarity 0.89)
- Meta-Curriculum Learning for Domain Adaptation in Neural Machine Translation (AAAI 2021 · similarity 0.89)
- A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models (ICLR 2024 · similarity 0.87)
- AEQA-NAT : Adaptive End-to-end Quantization Alignment Training Framework for Non-autoregressive Machine Translation (ICML 2025 · similarity 0.86)
- Calibrating Translation Decoding with Quality Estimation on LLMs (NeurIPS 2025 · 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.