Non-Autoregressive Text Generation with Pre-trained Language Models
Autor: | David Vandyke, Deng Cai, Yixuan Su, Simon Baker, Yan Wang, Piji Li, Nigel Collier |
---|---|
Rok vydání: | 2021 |
Předmět: |
FOS: Computer and information sciences
Computer Science - Computation and Language Machine translation business.industry Computer science Inference Machine learning computer.software_genre Automatic summarization Autoregressive model Conditional independence Component (UML) Language model Artificial intelligence business computer Computation and Language (cs.CL) Decoding methods |
Zdroj: | EACL |
DOI: | 10.48550/arxiv.2102.08220 |
Popis: | 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. Comment: Accepted to EACL 2021 |
Databáze: | OpenAIRE |
Externí odkaz: |