Levenshtein Transformer

Abstract

Modern neural sequence generation models are built to either generate tokens step-by-step from scratch or (iteratively) modify a sequence of tokens bounded by a fixed length. In this work, we develop Levenshtein Transformer, a new partially autoregressive model devised for more flexible and amenable sequence generation. Unlike previous approaches, the atomic operations of our model are insertion and deletion. The combination of them facilitates not only generation but also sequence refinement allowing dynamic length changes. We also propose a set of new training techniques dedicated at them, effectively exploiting one as the other's learning signal thanks to their complementary nature. Experiments applying the proposed model achieve comparable performance but much-improved efficiency on both generation (e.g. machine translation, text summarization) and refinement tasks (e.g. automatic post-editing). We further confirm the flexibility of our model by showing a Levenshtein Transformer trained by machine translation can straightforwardly be used for automatic post-editing.

17 pages (6 pages appendix). Camera ready, accepted by NeurIPS 2019

Cited by in corpus (21)

• HAT: Hardware-Aware Transformers for Efficient Natural Language Processing
• Enable Deep Learning on Mobile Devices: Methods, Systems, and Applications
• Aligned Cross Entropy for Non-Autoregressive Machine Translation
• Non-Autoregressive Machine Translation with Disentangled Context Transformer
• Understanding and Improving Lexical Choice in Non-Autoregressive Translation
• Incorporating BERT into Parallel Sequence Decoding with Adapters
• Understanding Knowledge Distillation in Non-autoregressive Machine Translation
• FastCorrect: Fast Error Correction with Edit Alignment for Automatic Speech Recognition
• BANG: Bridging Autoregressive and Non-autoregressive Generation with Large Scale Pretraining
• Cascaded Text Generation with Markov Transformers
• Fully Non-autoregressive Neural Machine Translation: Tricks of the Trade
• Multi-Task Learning with Shared Encoder for Non-Autoregressive Machine Translation
• Insertion-Deletion Transformer
• Neighbors Are Not Strangers: Improving Non-Autoregressive Translation under Low-Frequency Lexical Constraints
• Task-Level Curriculum Learning for Non-Autoregressive Neural Machine Translation
• Integrated Training for Sequence-to-Sequence Models Using Non-Autoregressive Transformer
• Learning to Recover from Multi-Modality Errors for Non-Autoregressive Neural Machine Translation
• UCorrect: An Unsupervised Framework for Automatic Speech Recognition Error Correction
• Search Spaces for Neural Model Training
• Learning Discrete Energy-based Models via Auxiliary-variable Local Exploration
• Sketch and Refine: Towards Faithful and Informative Table-to-Text Generation