Abstract

We report on our system for textual inference and question entailment in the medical domain for the ACL BioNLP 2019 Shared Task, MEDIQA. Textual inference is the task of finding the semantic relationships between pairs of text. Question entailment involves identifying pairs of questions which have similar semantic content. To improve upon medical natural language inference and question entailment approaches to further medical question answering, we propose a system that incorporates open-domain and biomedical domain approaches to improve semantic understanding and ambiguity resolution. Our models achieve 80% accuracy on medical natural language inference (6.5% absolute improvement over the original baseline), 48.9% accuracy on recognising medical question entailment, 0.248 Spearman’s rho for question answering ranking and 68.6% accuracy for question answering classification.

Highlights

  • Medical health search is the second most searched thematic query, representing 5% of all queries on Google (Cocco et al, 2018)

  • As a means to retrieve these questions that are already answered by experts, question entailment has been proposed to discern relationships between pairs of questions

  • We detail our approach in MEDIQA which addresses some of the problems with biomedical text such as utilising deep contextual relationships between words within a sentence for semantic understanding and ambiguity associated with esoteric terminology, abbreviations, and patient colloquialism

Read more

Summary

Introduction

Medical health search is the second most searched thematic query, representing 5% of all queries on Google (Cocco et al, 2018). Many queries are semantically identical and are potentially already answered by experts (Abacha and Demner-Fushman, 2016). These questions may not be directly retrievable due to semantic ambiguity involving abbreviations (Wu et al, 2017), patient colloquialism (Graham and Brookey, 2008) or esoteric terminology (Lee et al, 2019). We detail our approach in MEDIQA which addresses some of the problems with biomedical text such as utilising deep contextual relationships between words within a sentence for semantic understanding and ambiguity associated with esoteric terminology, abbreviations, and patient colloquialism. Proceedings of the BioNLP 2019 workshop, pages 478–487 Florence, Italy, August 1, 2019. c 2019 Association for Computational Linguistics

Datasets
Our System
Results and Discussion
Conclusions
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call