Title | ||
---|---|---|
Bi-Directional Differentiable Input Reconstruction for Low-Resource Neural Machine Translation. |
Abstract | ||
---|---|---|
We aim to better exploit the limited amounts of parallel text available in low-resource settings by introducing a differentiable reconstruction loss for neural machine translation (NMT). We reconstruct the input from sampled translations and leverage differentiable sampling and bi-directional NMT to build a compact model that can be trained end-to-end. This approach achieves small but consistent BLEU improvements on four language pairs in both translation directions, and outperforms an alternative differentiable reconstruction strategy based on hidden states. |
Year | DOI | Venue |
---|---|---|
2018 | 10.18653/v1/n19-1043 | north american chapter of the association for computational linguistics |
Field | DocType | Volume |
BLEU,Computer science,Machine translation,Exploit,Differentiable function,Artificial intelligence,Sampling (statistics),Machine learning | Journal | abs/1811.01116 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xing Niu | 1 | 135 | 10.15 |
Weijia Xu | 2 | 0 | 5.75 |
Marine Carpuat | 3 | 587 | 51.99 |