Sequence-Level Training for Non-Autoregressive Neural Machine Translation

Abstract

In recent years, Neural Machine Translation (NMT) has achieved notable results in various translation tasks. However, the word-by-word generation manner determined by the autoregressive mechanism leads to high translation latency of the NMT and restricts its low-latency applications. Non-Autoregressive Neural Machine Translation (NAT) removes the autoregressive mechanism and achieves significant decoding speedup through generating target words independently and simultaneously. Nevertheless, NAT still takes the word-level cross-entropy loss as the training objective, which is not optimal because the output of NAT cannot be properly evaluated due to the multimodality problem. In this article, we propose using sequence-level training objectives to train NAT models, which evaluate the NAT outputs as a whole and correlates well with the real translation quality. Firstly, we propose training NAT models to optimize sequence-level evaluation metrics (e.g., BLEU) based on several novel reinforcement algorithms customized for NAT, which outperforms the conventional method by reducing the variance of gradient estimation. Secondly, we introduce a novel training objective for NAT models, which aims to minimize the Bag-of-Ngrams (BoN) difference between the model output and the reference sentence. The BoN training objective is differentiable and can be calculated efficiently without doing any approximations. Finally, we apply a three-stage training strategy to combine these two methods to train the NAT model. We validate our approach on four translation tasks (WMT14 En$\leftrightarrow$De, WMT16 En$\leftrightarrow$Ro), which shows that our approach largely outperforms NAT baselines and achieves remarkable performance on all translation tasks. The source code is available at https://github.com/ictnlp/Seq-NAT.

Computational Linguistics Journal

References in corpus (13)

• Distilling the Knowledge in a Neural Network
• Sequence to Sequence Learning with Neural Networks
• Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
• Dual Learning for Machine Translation
• Professor Forcing: A New Algorithm for Training Recurrent Networks
• The Optimal Reward Baseline for Gradient-Based Reinforcement Learning
• Reward Augmented Maximum Likelihood for Neural Structured Prediction
• Aligned Cross Entropy for Non-Autoregressive Machine Translation
• Imitation Learning for Non-Autoregressive Neural Machine Translation
• Non-autoregressive Transformer by Position Learning
• An EM Approach to Non-autoregressive Conditional Sequence Generation
• Retrieving Sequential Information for Non-Autoregressive Neural Machine Translation
• Learning to Recover from Multi-Modality Errors for Non-Autoregressive Neural Machine Translation