Abstract | ||
---|---|---|
Large amounts of data has made neural machine translation (NMT) a big success in recent years. But it is still a challenge if we train these models on small-scale corpora. In this case, the way of using data appears to be more important. Here, we investigate the effective use of training data for low-resource NMT. In particular, we propose a dynamic curriculum learning (DCL) method to reorder training samples in training. Unlike previous work, we do not use a static scoring function for reordering. Instead, the order of training samples is dynamically determined in two ways - loss decline and model competence. This eases training by highlighting easy samples that the current model has enough competence to learn. We test our DCL method in a Transformer-based system. Experimental results show that DCL outperforms several strong baselines on three low-resource machine translation benchmarks and different sized data of WMT' 16 En-De. |
Year | Venue | DocType |
---|---|---|
2020 | COLING | Conference |
Volume | Citations | PageRank |
2020.coling-main | 0 | 0.34 |
References | Authors | |
0 | 9 |
Name | Order | Citations | PageRank |
---|---|---|---|
Chen Xu | 1 | 0 | 0.34 |
Bojie Hu | 2 | 0 | 2.37 |
Yufan Jiang | 3 | 0 | 0.34 |
Kai Feng | 4 | 0 | 0.34 |
Zeyang Wang | 5 | 0 | 0.34 |
Shen Huang | 6 | 64 | 14.51 |
Qi Ju | 7 | 0 | 0.34 |
Tong Xiao | 8 | 131 | 23.91 |
Jingbo Zhu | 9 | 703 | 64.21 |