AUEB at BioASQ 6: Document and Snippet Retrieval

TL;DR

This paper describes AUEB's top-performing deep learning models for BioASQ 6's document and snippet retrieval tasks, demonstrating the effectiveness of text-only deep learning architectures in biomedical information retrieval.

Contribution

Introduction of novel deep learning extensions tailored for text-only biomedical document and snippet retrieval tasks, achieving top or near-top results in BioASQ 6.

Findings

Models scored at the top or near the top in all challenge batches

Deep learning architectures effectively handle biomedical retrieval tasks

Text-only models are highly effective for document and snippet retrieval

Abstract

We present AUEB's submissions to the BioASQ 6 document and snippet retrieval tasks (parts of Task 6b, Phase A). Our models use novel extensions to deep learning architectures that operate solely over the text of the query and candidate document/snippets. Our systems scored at the top or near the top for all batches of the challenge, highlighting the effectiveness of deep learning for these tasks.