Sentence Selection for QA

This software accompanies the M.Sc. Thesis: "Sentence selection from biomedical documents for question answering".

It also accompanies the following paper:

G. Brokos, P. Liosis, R. McDonald, D. Pappas and I. Androutsopoulos, "AUEB at BioASQ 6: Document and Snippet Retrieval". Proceedings of the workshop BioASQ: Large-scale Biomedical Semantic Indexing and Question Answering, at the Conference on Empirical Methods in Natural Language Processing (EMNLP 2018), Brussels, Belgium, 2018.[PDF].

---

Specification

• BM25: Implementation of BM25 ranking function.
• DL_Models: Deep Learning model used.
• Extract_BioASQ_Data: Extract data from BioASQ's published json files and create datasets with various formats.