Paper: Generation under Space Constraints

ACL ID C08-2020
Title Generation under Space Constraints
Venue International Conference on Computational Linguistics
Session Poster Session
Year 2008

Reasoning about how much to generate when space is limited is a challenge for generation systems. This paper presents two algorithms that exploit the discourse structure to decide which content to drop when there are space restrictions, in the context of producing documents from pre-authored text fragments. We analyse the effectiveness of both algorithms and show that the second is near optimal.