Convergent Privacy Framework for Multi-layer GNNs through Contractive Message Passing
CARIBOU is a privacy-preserving GNN framework that couples a Contractive Graph Layer (CGL) with convergent privacy accounting, enabling deeper aggregation under edge-level DP (EDP) and node-level DP (NDP) with superior utility.
Institutions
- University of California, Irvine (US)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2025-11-06
- DOI
- https://doi.org/10.5281/zenodo.17539659
- Primary Topic
- Privacy-Preserving Technologies in Data
- Type
- article
- Field-Weighted Citation Impact
- 0.00