update

README.md

@@ -1 +1,37 @@
-# HiFGL
+# HiFGL: A Hierarchical Framework for Cross-silo Cross-device Federated Graph Learning
+
+This is the official repository of the KDD'2024 Research Track Paper ["HiFGL: A Hierarchical Framework for Cross-silo Cross-device Federated Graph Learning"](https://dl.acm.org/doi/10.1145/3637528.3671660) ([Preprint](https://arxiv.org/abs/2406.10616)).
+
+### Introduction
+
+Federated Graph Learning (FGL) has emerged as a promising way to learn high-quality representations from distributed graph data with privacy preservation. Despite considerable efforts have been made for FGL under either cross-device or cross-silo paradigm, how to effectively capture graph knowledge in a more complicated cross-silo cross-device environment remains an under-explored problem. However, this task is challenging because of the inherent hierarchy and heterogeneity of decentralized clients, diversified privacy constraints in different clients, and the cross-client graph integrity requirement. To this end, in this paper, we propose a Hierarchical Federated Graph Learning (HiFGL) framework for cross-silo cross-device FGL. Specifically, we devise a unified hierarchical architecture to safeguard federated GNN training on heterogeneous clients while ensuring graph integrity. Moreover, we propose a Secret Message Passing (SecMP) scheme to shield unauthorized access to subgraph-level and node-level sensitive information simultaneously. Theoretical analysis proves that HiFGL achieves multi-level privacy preservation with complexity guarantees. Extensive experiments on real-world datasets validate the superiority of the proposed framework against several baselines. Furthermore, HiFGL's versatile nature allows for its application in either solely cross-silo or cross-device settings, further broadening its utility in real-world FGL applications.
+
+<p align="center">
+  <img src="./HiFGL.png" alt="Hierarchical Federated Graph Learning" width="100%">
+</p>
+
+### Requirements
+
+- Python 3.8
+- PyTorch 1.10
+- PyTorch Lightning 1.5.3
+- TorchMetrics 0.6
+
+### Run
+
+An example that appoints dataset with `cora` and epochs with `64`.
+
+```
+python ./code/main.py -D cora -E 64
+```
+
+### Citation
+
+```
+@article{guo2024hifgl,
+  title={HiFGL: A Hierarchical Framework for Cross-silo Cross-device Federated Graph Learning},
+  author={Guo, Zhuoning and Yao, Duanyi and Yang, Qiang and Liu, Hao},
+  journal={arXiv preprint arXiv:2406.10616},
+  year={2024}
+}
+```

code/args.py

@@ -12,7 +12,7 @@ def parse_global_args():
     parser.add_argument('--root', type=str, default='./')
     parser.add_argument('--data_folder', type=str, default='data')
     parser.add_argument('--args_folder', type=str, default='args')
-    parser.add_argument('-D', '--dataset_name', type=str)
+    parser.add_argument('-D', '--dataset_name', type=str, default='cora')
     parser.add_argument('-E', '--epochs', type=int, default=1024, help='epoch size')
     parser.add_argument('--train_batch_size', type=int, default=1, help='train batch size.')
     parser.add_argument('--val_batch_size', type=int, default=1, help='validation batch size.')

code/train.py

@@ -102,7 +102,7 @@ def test_step(self, *args, **kwargs):
         self.validation_step(*args, **kwargs)
 
     def _get_clients(self):
-        return {i: torch.load(f"../data/{self.hparams.dataset_name}/{i}_clients.pt") for i in ([-1] + list(range(1, self.hparams.num_clients + 1)))}
+        return {i: torch.load(f"./data/{self.hparams.dataset_name}/{i}_clients.pt") for i in ([-1] + list(range(1, self.hparams.num_clients + 1)))}
 
     def _count_classes(self, clients: dict):
         classes = set()