Knowledge-Driven Distractor Generation for Cloze-Style Multiple Choice Questions

Abstract

In this paper, we propose a novel configurable framework to automatically generate distractive choices for open-domain cloze-style multiple-choice questions. The framework incorporates a general-purpose knowledge base to effectively create a small distractor candidate set, and a feature-rich learning-to-rank model to select distractors that are both plausible and reliable. Experimental results on a new dataset across four domains show that our framework yields distractors outperforming previous methods both by automatic and human evaluation. The dataset can also be used as a benchmark for distractor generation research in the future.

Cite

Text

Ren and Zhu. "Knowledge-Driven Distractor Generation for Cloze-Style Multiple Choice Questions." AAAI Conference on Artificial Intelligence, 2021. doi:10.1609/AAAI.V35I5.16559

Markdown

[Ren and Zhu. "Knowledge-Driven Distractor Generation for Cloze-Style Multiple Choice Questions." AAAI Conference on Artificial Intelligence, 2021.](https://mlanthology.org/aaai/2021/ren2021aaai-knowledge/) doi:10.1609/AAAI.V35I5.16559

BibTeX

@inproceedings{ren2021aaai-knowledge,
  title     = {{Knowledge-Driven Distractor Generation for Cloze-Style Multiple Choice Questions}},
  author    = {Ren, Siyu and Zhu, Kenny Q.},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2021},
  pages     = {4339-4347},
  doi       = {10.1609/AAAI.V35I5.16559},
  url       = {https://mlanthology.org/aaai/2021/ren2021aaai-knowledge/}
}