Crowdsourcing Multiple Choice Science Questions

Johannes Welbl, Nelson F. Liu, Matt Gardner


Abstract
We present a novel method for obtaining high-quality, domain-targeted multiple choice questions from crowd workers. Generating these questions can be difficult without trading away originality, relevance or diversity in the answer options. Our method addresses these problems by leveraging a large corpus of domain-specific text and a small set of existing questions. It produces model suggestions for document selection and answer distractor choice which aid the human question generation process. With this method we have assembled SciQ, a dataset of 13.7K multiple choice science exam questions. We demonstrate that the method produces in-domain questions by providing an analysis of this new dataset and by showing that humans cannot distinguish the crowdsourced questions from original questions. When using SciQ as additional training data to existing questions, we observe accuracy improvements on real science exams.
Anthology ID:
W17-4413
Volume:
Proceedings of the 3rd Workshop on Noisy User-generated Text
Month:
September
Year:
2017
Address:
Copenhagen, Denmark
Editors:
Leon Derczynski, Wei Xu, Alan Ritter, Tim Baldwin
Venue:
WNUT
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
94–106
Language:
URL:
https://aclanthology.org/W17-4413
DOI:
10.18653/v1/W17-4413
Bibkey:
Cite (ACL):
Johannes Welbl, Nelson F. Liu, and Matt Gardner. 2017. Crowdsourcing Multiple Choice Science Questions. In Proceedings of the 3rd Workshop on Noisy User-generated Text, pages 94–106, Copenhagen, Denmark. Association for Computational Linguistics.
Cite (Informal):
Crowdsourcing Multiple Choice Science Questions (Welbl et al., WNUT 2017)
Copy Citation:
PDF:
https://aclanthology.org/W17-4413.pdf
Data
SciQSQuAD