CoAD: Automatic Diagnosis through Symptom and Disease Collaborative Generation

Authors: Huimin Wang Wai-Chung Kwan, Kam-Fai Wong, Yefeng Zheng

Code of our paper: CoAD: Automatic Diagnosis through Symptom and Disease Collaborative Generation

Environment Setup:

conda env create --name coad --file environment.yml

Dataset

We only download the dxy and muzhi dataset due to license issue.

cd data
chmod u+x download_dataset.sh
./download_dataset.sh

Training

# dataset: Either dxy, muzhi.
python train.py \
--dataset dxy \
--do_sym_augmentation \
--do_dis_augmentation \
--do_train

Testing

python train.py 
--dataset dxy \
--checkpoint $checkpoint_path \
--do_test