DHSampling: Diversity-based Hyperedge Sampling in GNN Learning

"DHSampling: Diversity-based Hyperedge Sampling in GNN Learning with Application to Medical Image Classification", Workshop on MLMI, MICCAI 2024, [Jiameng Liu, Furkan Pala, Islem Rekik, and Dinggang Shen]

Install

• Clone this repo and install corresponding requirement in requirements.txt

  git clone https://github.com/basiralab/DHSampling.git

  pip install -r requirements

Data Preparation

We train and validate our proposed DHSampling on two publicly available MedMNIST data (i.e.,OrganCMNIST, OrganSMNIST). For reproducing the code, you need to download the MedMNIST data and convert to Graph as following:

1. pip install medmnist

   1. Download OrganCMNIST and OrganSMNIST data according to this repo

2. Data preprocess for MedMNIST dataset

   python ./DataPrepare/Process.py --medmnist_origin /folder/to/original/medmnist --medmnist_target /folder/to/processed/data

3. Convert processed data to graph

   python ./DataPrepare/GraphConstruction.py --medmnist_npy /folder/to/processed/data --medmnist_garph /folder/to/save/graph

4. Copy corresponding converted graph into ./GCN/input, ./GAT/input

Training and Evaluation

• Script for DHSampling on OrganCMNIST based on GCN
  • Change file_folder=organc in file ./GCN/src/parse.py
  • sh ./run/OrganCMNIST_GCN.sh
• Script for DHSampling on OrganSMNIST based on GCN
  • Change file_folder=organs in file ./GCN/src/parse.py
  • sh ./run/OrganSMNIST_GCN.sh
• Script for DHSampling on OrganCMNIST based on GAT
  • Change file_folder=organc in file ./GAT/src/parse.py
  • sh ./run/OrganCMNIST_GAT.sh
• Script for DHSampling on OrganSMNIST based on GAT
  • Change file_folder=organs in file ./GAT/src/parse.py
  • sh ./run/OrganSMNIST_GAT.sh

---

This implementation is highly inspired by ClusterGCN in this Repo

License

Copyright IDEA Lab, School of Biomedical Engineering, ShanghaiTech University, Shanghai, China. & BASIRA Lab, Department of Computing, Imperical College London

Licensed under the the GPL (General Public License);
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Repo for Diversity-based Hyperedges Sampling in GNN Learning
Contact: JiamengLiu.PRC@gmail.com