Sequentially Controlled Text Generation

Alexander Spangher, Yao Ming, Xinyu Hua, Nanyun Peng


Abstract
While GPT-2 generates sentences that are remarkably human-like, longer documents can ramble and do not follow human-like writing structure. We study the problem of imposing structure on long-range text. We propose a novel controlled text generation task, sequentially controlled text generation, and identify a dataset, NewsDiscourse as a starting point for this task. We develop a sequential controlled text generation pipeline with generation and editing. We test different degrees of structural awareness and show that, in general, more structural awareness results in higher control- accuracy, grammaticality, coherency and topicality, approaching human-level writing performance.
Anthology ID:
2022.findings-emnlp.509
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2022
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6848–6866
Language:
URL:
https://aclanthology.org/2022.findings-emnlp.509
DOI:
Bibkey:
Cite (ACL):
Alexander Spangher, Yao Ming, Xinyu Hua, and Nanyun Peng. 2022. Sequentially Controlled Text Generation. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 6848–6866, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Sequentially Controlled Text Generation (Spangher et al., Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-emnlp.509.pdf