# DoubleTransfer at MEDIQA 2019: Multi-Source Transfer Learning for Natural Language Understanding in the Medical Domain

## TL;DR

This paper presents a multi-source transfer learning approach using MT-DNN and SciBERT, fine-tuned on multiple tasks, to improve medical domain natural language understanding, achieving top performance in MEDIQA 2019.

## Contribution

It introduces a multi-source transfer learning method combining MT-DNN and SciBERT with multi-task fine-tuning for medical NLP tasks.

## Key findings

- Achieved first place in the MEDIQA 2019 QA task.
- Effective transfer learning from general and medical domain data.
- Improved performance across NLI, RQE, and QA tasks.

## Abstract

This paper describes our competing system to enter the MEDIQA-2019 competition. We use a multi-source transfer learning approach to transfer the knowledge from MT-DNN and SciBERT to natural language understanding tasks in the medical domain. For transfer learning fine-tuning, we use multi-task learning on NLI, RQE and QA tasks on general and medical domains to improve performance. The proposed methods are proved effective for natural language understanding in the medical domain, and we rank the first place on the QA task.