Paper: Linguistic Considerations in Automatic Question Generation

ACL ID P14-2053
Title Linguistic Considerations in Automatic Question Generation
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2014
Authors

As students read expository text, compre- hension is improved by pausing to answer questions that reinforce the material. We describe an automatic question generator that uses semantic pattern recognition to create questions of varying depth and type for self-study or tutoring. Throughout, we explore how linguistic considerations in- form system design. In the described sys- tem, semantic role labels of source sen- tences are used in a domain-independent manner to generate both questions and an- swers related to the source sentence. Eval- uation results show a 44% reduction in the error rate relative to the best prior systems, averaging over all metrics, and up to 61% reduction in the error rate on grammatical- ity judgments.