Pivot-based Transfer Learning for Neural Machine Translation between Non-English Languages
This addresses translation challenges for languages with limited parallel data, though it is incremental as it builds on existing pivot-based methods.
The paper tackled the problem of neural machine translation between non-English languages by proposing pre-training strategies using a pivot language, resulting in a significant improvement of up to +2.6% BLEU on WMT 2019 tasks.
We present effective pre-training strategies for neural machine translation (NMT) using parallel corpora involving a pivot language, i.e., source-pivot and pivot-target, leading to a significant improvement in source-target translation. We propose three methods to increase the relation among source, pivot, and target languages in the pre-training: 1) step-wise training of a single model for different language pairs, 2) additional adapter component to smoothly connect pre-trained encoder and decoder, and 3) cross-lingual encoder training via autoencoding of the pivot language. Our methods greatly outperform multilingual models up to +2.6% BLEU in WMT 2019 French-German and German-Czech tasks. We show that our improvements are valid also in zero-shot/zero-resource scenarios.