Research map / Architectures and efficiency / RNNs and normalization
Guiding Non-Autoregressive Neural Machine Translation Decoding with Reordering Information
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 explicit pseudo-translation reordering guidance (ablation) | NAT translation · IWSLT16 En→De validation | BLEU: 24.57 -> 25.29 | Helped |
| use an AT GRU rather than an NAT reordering module (ablation) | ReorderNAT · IWSLT16 En→De validation | BLEU: 25.29 -> 30.26 | Helped |
| add length parallel decoding (LPD, s=7) (ablation) | ReorderNAT (NAT) · IWSLT16 En→De validation | BLEU: 25.29 -> 27.40 | Helped |
| replace deterministic guiding with non-deterministic guiding (ablation) | ReorderNAT (NAT) · IWSLT16 En→De validation | BLEU: 24.94 -> 25.29 | Helped |
| replace deterministic guiding with non-deterministic guiding (ablation) | ReorderNAT (AT) · IWSLT16 En→De validation | BLEU: 30.15 -> 30.26 | Helped |
| replace deterministic guiding with non-deterministic guiding (ablation) | ReorderNAT (AT) · WMT14 En→De | BLEU: 26.49 -> 26.51 | No clear effect |
| replace deterministic guiding with non-deterministic guiding (ablation) | ReorderNAT (NAT) · WMT14 En→De and De→En | BLEU: 22.79 -> 21.05; 27.28 -> 25.92 | Hurt |
| increase temperature from 0.2 to 0.5 or 1 (ablation) | ReorderNAT NDGD · IWSLT16 En→De validation | BLEU | Hurt |
| use temperature T=0.1 instead of T=0.2 (ablation) | ReorderNAT NDGD · IWSLT16 En→De validation | BLEU | Hurt |
| vary the proportion of fast_align training data (ablation) | ReorderNAT · IWSLT16 En→De validation | BLEU: 24.48, 24.58, 24.92, 25.07, 24.94; 25.19, 24.85, 25.22, 25.47, 25.29 | No clear effect |
| add explicit pseudo-translation reordering guidance (ablation) | NAT translation · IWSLT16 En→De validation | RIBES: 82.21 -> 82.35 | Helped |
| add explicit pseudo-translation reordering guidance (ablation) | NAT translation · IWSLT16 En→De validation | Dup: 50.09 -> 37.52 | Helped |
Closest papers by meaning
- Non-Autoregressive Neural Machine Translation with Enhanced Decoder Input (AAAI 2019 · similarity 0.91)
- Non-Autoregressive Machine Translation with Auxiliary Regularization (AAAI 2019 · similarity 0.90)
- Minimizing the Bag-of-Ngrams Difference for Non-Autoregressive Neural Machine Translation (AAAI 2020 · similarity 0.90)
- Rephrasing the Reference for Non-autoregressive Machine Translation (AAAI 2023 · similarity 0.90)
- Non-Autoregressive Neural Machine Translation (ICLR 2018 · similarity 0.89)
- AEQA-NAT : Adaptive End-to-end Quantization Alignment Training Framework for Non-autoregressive Machine Translation (ICML 2025 · similarity 0.88)
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.