BibleQA

Summary

BibleQA is a dataset and code repository for biblical question answering, created as part of the paper 'Finding Answers from the Word of God: Domain Adaptation for Neural Networks in Biblical Question Answering' (arXiv:1810.12118, IJCNN 2018). The dataset contains Bible trivia questions with answers drawn from multiple Bible translations (KJV, ASV, YLT, WEB). The repository includes baseline models using CNN, BiDAF, and simple neural approaches, with experiments on transfer learning from SQuAD showing that pre-training improves accuracy and that modern Bible translations yield better results.

Keywords

bible qa question-answering biblical-nlp domain-adaptation transfer-learning squad

Topics

NLP / Religion

Research notes

• GitHub repository (8 stars, 8 forks) by Helen Jiahe Zhao and Jiamou Liu. Paper accepted at IJCNN 2018. Contains data, preprocessing scripts, and multiple baseline model implementations (CNN, BiDAF, simple).