Make The Most of Prior Data: A Solution for Interactive Text Summarization with Preference Feedback

Duy-Hung Nguyen, Nguyen Viet Dung Nghiem, Bao-Sinh Nguyen, Dung Tien Tien Le, Shahab Sabahi, Minh-Tien Nguyen, Hung Le


Abstract
For summarization, human preferences is critical to tame outputs of the summarizer in favor of human interests, as ground-truth summaries are scarce and ambiguous. Practical settings require dynamic exchanges between humans and AI agents wherein feedback is provided in an online manner, a few at a time. In this paper, we introduce a new framework to train summarization models with preference feedback interactively. By properly leveraging offline data and a novel reward model, we improve the performance regarding ROUGE scores and sample-efficiency. Our experiments on three various datasets confirm the benefit of the proposed framework in active, few-shot and online settings of preference learning.
Anthology ID:
2022.findings-naacl.147
Volume:
Findings of the Association for Computational Linguistics: NAACL 2022
Month:
July
Year:
2022
Address:
Seattle, United States
Editors:
Marine Carpuat, Marie-Catherine de Marneffe, Ivan Vladimir Meza Ruiz
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1919–1930
Language:
URL:
https://aclanthology.org/2022.findings-naacl.147
DOI:
10.18653/v1/2022.findings-naacl.147
Bibkey:
Cite (ACL):
Duy-Hung Nguyen, Nguyen Viet Dung Nghiem, Bao-Sinh Nguyen, Dung Tien Tien Le, Shahab Sabahi, Minh-Tien Nguyen, and Hung Le. 2022. Make The Most of Prior Data: A Solution for Interactive Text Summarization with Preference Feedback. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 1919–1930, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
Make The Most of Prior Data: A Solution for Interactive Text Summarization with Preference Feedback (Nguyen et al., Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-naacl.147.pdf
Video:
 https://aclanthology.org/2022.findings-naacl.147.mp4
Data
BillSumReddit TIFU