Don’t Add, don’t Miss: Effective Content Preserving Generation from Pre-Selected Text Spans

Aviv Slobodkin, Avi Caciularu, Eran Hirsch, Ido Dagan


Abstract
The recently introduced Controlled Text Reduction (CTR) task isolates the text generation step within typical summarization-style tasks. It does so by challenging models to generate coherent text conforming to pre-selected content within the input text (“highlights”). This framing enables increased modularity in summarization-like tasks, allowing to couple a single CTR model with various content-selection setups and modules. However, there are currently no reliable CTR models, while the performance of the existing baseline for the task is mediocre, falling short of practical utility. Here, we address this gap by introducing a high-quality, open-source CTR model that tackles two prior key limitations: inadequate enforcement of the content-preservation constraint, and suboptimal silver training data. Addressing these, we amplify the content-preservation constraint in both training, via RL, and inference, via a controlled decoding strategy. Further, we substantially improve the silver training data quality via GPT-4 distillation. Overall, pairing the distilled dataset with the highlight-adherence strategies yields marked gains over the current baseline, of up to 30 ROUGE-L points, providing a reliable CTR model for downstream use.
Anthology ID:
2023.findings-emnlp.852
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2023
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
12784–12800
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.852
DOI:
10.18653/v1/2023.findings-emnlp.852
Bibkey:
Cite (ACL):
Aviv Slobodkin, Avi Caciularu, Eran Hirsch, and Ido Dagan. 2023. Don’t Add, don’t Miss: Effective Content Preserving Generation from Pre-Selected Text Spans. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 12784–12800, Singapore. Association for Computational Linguistics.
Cite (Informal):
Don’t Add, don’t Miss: Effective Content Preserving Generation from Pre-Selected Text Spans (Slobodkin et al., Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-emnlp.852.pdf