Controllable Summarization with Constrained Markov Decision Process

Hou Pong Chan, Lu Wang, Irwin King


Abstract
We study controllable text summarization, which allows users to gain control on a particular attribute (e.g., length limit) of the generated summaries. In this work, we propose a novel training framework based on Constrained Markov Decision Process (CMDP), which conveniently includes a reward function along with a set of constraints, to facilitate better summarization control. The reward function encourages the generation to resemble the human-written reference, while the constraints are used to explicitly prevent the generated summaries from violating user-imposed requirements. Our framework can be applied to control important attributes of summarization, including length, covered entities, and abstractiveness, as we devise specific constraints for each of these aspects. Extensive experiments on popular benchmarks show that our CMDP framework helps generate informative summaries while complying with a given attribute’s requirement.1
Anthology ID:
2021.tacl-1.72
Volume:
Transactions of the Association for Computational Linguistics, Volume 9
Month:
Year:
2021
Address:
Cambridge, MA
Editors:
Brian Roark, Ani Nenkova
Venue:
TACL
SIG:
Publisher:
MIT Press
Note:
Pages:
1213–1232
Language:
URL:
https://aclanthology.org/2021.tacl-1.72
DOI:
10.1162/tacl_a_00423
Bibkey:
Cite (ACL):
Hou Pong Chan, Lu Wang, and Irwin King. 2021. Controllable Summarization with Constrained Markov Decision Process. Transactions of the Association for Computational Linguistics, 9:1213–1232.
Cite (Informal):
Controllable Summarization with Constrained Markov Decision Process (Chan et al., TACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.tacl-1.72.pdf
Video:
 https://aclanthology.org/2021.tacl-1.72.mp4