Hint-Based Training for Non-Autoregressive Machine Translation
arXiv:1909.06708
Abstract
Due to the unparallelizable nature of the autoregressive factorization, AutoRegressive Translation (ART) models have to generate tokens sequentially during decoding and thus suffer from high inference latency. Non-AutoRegressive Translation (NART) models were proposed to reduce the inference time, but could only achieve inferior translation accuracy. In this paper, we proposed a novel approach to leveraging the hints from hidden states and word alignments to help the training of NART models. The results achieve significant improvement over previous NART models for the WMT14 En-De and De-En datasets and are even comparable to a strong LSTM-based ART baseline but one order of magnitude faster in inference.
arXiv commentEMNLP-IJCNLP 2019
References in corpus (10)
- Distilling the Knowledge in a Neural Network
- Convolutional Sequence to Sequence Learning
- FitNets: Hints for Thin Deep Nets
- Non-Autoregressive Neural Machine Translation
- Neural Machine Translation in Linear Time
- An Actor-Critic Algorithm for Sequence Prediction
- Fast Decoding in Sequence Models using Discrete Latent Variables
- Sequence-Level Knowledge Distillation
- A Convolutional Encoder Model for Neural Machine Translation
- Attention-Guided Answer Distillation for Machine Reading Comprehension
Cited by in corpus (6)
- Aligned Cross Entropy for Non-Autoregressive Machine Translation
- LAVA NAT: A Non-Autoregressive Translation Model with Look-Around Decoding and Vocabulary Attention
- Duplex Sequence-to-Sequence Learning for Reversible Machine Translation
- Fine-Tuning by Curriculum Learning for Non-Autoregressive Neural Machine Translation
- Improving Non-autoregressive Neural Machine Translation with Monolingual Data
- Task-Level Curriculum Learning for Non-Autoregressive Neural Machine Translation