Non-Autoregressive Text Generation with Pre-trained Language Models
This work improves non-autoregressive generation for faster text generation applications, though it is incremental as it builds on existing methods.
The authors tackled the problem of low generation quality in non-autoregressive text generation by using BERT as a backbone and introducing mechanisms to address output length inflexibility and token independence, achieving competitive performance with autoregressive models on tasks like text summarization and machine translation.
Non-autoregressive generation (NAG) has recently attracted great attention due to its fast inference speed. However, the generation quality of existing NAG models still lags behind their autoregressive counterparts. In this work, we show that BERT can be employed as the backbone of a NAG model to greatly improve performance. Additionally, we devise mechanisms to alleviate the two common problems of vanilla NAG models: the inflexibility of prefixed output length and the conditional independence of individual token predictions. Lastly, to further increase the speed advantage of the proposed model, we propose a new decoding strategy, ratio-first, for applications where the output lengths can be approximately estimated beforehand. For a comprehensive evaluation, we test the proposed model on three text generation tasks, including text summarization, sentence compression and machine translation. Experimental results show that our model significantly outperforms existing non-autoregressive baselines and achieves competitive performance with many strong autoregressive models. In addition, we also conduct extensive analysis experiments to reveal the effect of each proposed component.