GraphAce: Secure Two-Party Graph Analysis Achieving Communication Efficiency
Jiping Yu
34th USENIX Security Symposium (USENIX Security '25) · Day 2 · Privacy 1: Differential Privacy and Audit
In an increasingly data-driven world, the ability to analyze vast datasets is paramount for extracting insights, detecting anomalies, and making informed decisions. However, a significant challenge arises when these datasets, particularly complex graph structures, are distributed across multiple entities, each holding sensitive information that cannot be directly shared due to privacy concerns, regulatory compliance, or proprietary interests. The talk "GraphAce: Secure Two-Party Graph Analysis Achieving Communication Efficiency" by Jiping Yu from China University, a collaborative effort with Kuni Xiaoi, Xiaoi Juan Hong, and Wuang Chen, addresses this critical problem head-on.
AI review
Legitimate academic systems security research with a real complexity result — O(V) communication per iteration for secure two-party graph analysis is a meaningful contribution over GraphSC's edge-dependent overhead. The work is technically sound and earned artifact badges, but it sits firmly in the privacy-preserving computation lane of academic cryptography, not offensive or defensive security operations.