GAP: Differentially Private Graph Neural Networks with Aggregation Perturbation
Description
The method is a novel differentially private Graph Neural Network (GNN) based on Aggregation Perturbation (GAP), which can statistically obfuscate the presence of a single edge (edge-level privacy) or a single node and all its adjacent edges (node-level privacy), thus providing formal privacy guarantees with competitive classification performance.
Publications
Sajadmanes, S. and Gatica-Perez, D. (2023) GAP: Differentially Private Graph Neural Networks with Aggregation Perturbation. 32nd USENIX Security Symposium (USENIX Security 23)
Links
- Repository: https://github.com/sisaman/GAP
Advantages
The method can be applied to both edge-level privacy and node-level privacy while learning in GNNs.
Applications
Applications involving graph data where node-level or edge-level privacy need to be preserved.
Technology Readiness Level
TRL 3
Contact us for more information
- Interested in using our technologies?
- Interested to know more about the licensing possibilities and conditions?