Unsupervised Pretraining for Neural Machine Translation Using Elastic Weight Consolidation

By Dusan Varis, Ondrej Bojar. 2019

19 citations recorded by the source corpus.

This work presents our ongoing research of unsupervised pretraining in neural machine translation (NMT). In our method, we initialize the weights of the encoder and decoder with two language models that are trained with monolingual data and then fine-tune the model on parallel data using Elastic Weight Consolidation (EWC) to avoid forgetting of the original language modeling task. We compare the regularization by EWC with the previous work that focuses on regularization by language modeling objectives. The positive result is that using EWC with the decoder achieves BLEU scores similar to the previous work. However, the model converges 2-3 times faster and does not require the original unlabeled training data during the fine-tuning stage. In contrast, the regularization using EWC is less effective if the original and new tasks are not closely related. We show that initializing the bidirectional NMT encoder with a left-to-right language model and forcing the model to remember the original left-to-right language modeling task limits the learning capacity of the encoder for the whole bidirectional context.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Fei Mi
2. Minlie Huang
3. Boi Faltings
4. Liangwei Chen
5. Mengjie Zhao
6. Thomas Scialom
7. Yanzhe Zhang
8. Zhuang Li

Citations in the corpus

• Continual Learning for Natural Language Generation in Task-oriented Dialog Systems (2020, background): “The catastrophic forgetting issue is also noticed in domain adaptation setups for neural machine translation (Saunders et al., 2019; Thompson et al., 2019; Varis and Bojar, 2019) and the reading comprehension task (Xu et al., 2019).”