Denoising autoencoders (non-autoregressive)

Reported on 6 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing6 results

Machine TranslationonWMT2014 German-English
BLEU score· 2018-02-19
25.43
best: 35.15 (Bi-SimCut)
SOTA
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901
Machine TranslationonIWSLT2015 German-English
BLEU score· 2018-02-19
32.43
best: 36.2 (PS-KD)
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901
Machine TranslationonIWSLT2015 English-German
BLEU score· 2018-02-19
27.01
best: 30 (PS-KD)
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901
Machine TranslationonWMT2016 Romanian-English
BLEU score· 2018-02-19
30.3
best: 40.3 (fast-noisy-channel-modeling)
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901
Machine TranslationonWMT2014 English-German
BLEU score· 2018-02-19
21.54
best: 35.14 (Transformer Cycle (Rev))
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901
Machine TranslationonWMT2016 English-Romanian
BLEU score· 2018-02-19
29.66
best: 34.7 (DeLighT)
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement arXiv:1802.06901