Choice-75: A Dataset on Decision Branching in Script Learning

Zhaoyi Hou, Li Zhang, Chris Callison-Burch


Abstract
Script learning studies how daily events unfold. It enables machines to reason about narratives with implicit information. Previous works mainly consider a script as a linear sequence of events while ignoring the potential branches that arise due to people’s circumstantial choices. We hence propose Choice-75, the first benchmark that challenges intelligent systems to make decisions given descriptive scenarios, containing 75 scripts and more than 600 scenarios. We also present preliminary results with current large language models (LLM). Although they demonstrate overall decent performances, there is still notable headroom in hard scenarios.
Anthology ID:
2024.lrec-main.285
Volume:
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)
Month:
May
Year:
2024
Address:
Torino, Italia
Editors:
Nicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessandro Lenci, Sakriani Sakti, Nianwen Xue
Venues:
LREC | COLING
SIG:
Publisher:
ELRA and ICCL
Note:
Pages:
3215–3223
Language:
URL:
https://aclanthology.org/2024.lrec-main.285
DOI:
Bibkey:
Cite (ACL):
Zhaoyi Hou, Li Zhang, and Chris Callison-Burch. 2024. Choice-75: A Dataset on Decision Branching in Script Learning. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 3215–3223, Torino, Italia. ELRA and ICCL.
Cite (Informal):
Choice-75: A Dataset on Decision Branching in Script Learning (Hou et al., LREC-COLING 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-5/2024.lrec-main.285.pdf
Optional supplementary material:
 2024.lrec-main.285.OptionalSupplementaryMaterial.zip