Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about GraphAce, a groundbreaking secure two-party graph analysis framework that dramatically reduces communication overhead in this 18-minute conference presentation from USENIX Security '25. Discover how researchers from Tsinghua University and Ant Group developed a novel approach to secure multi-party computation for graph analysis that breaks the traditional Omega(|V|+|E|) communication lower bound by achieving Theta(|V|) communication complexity per iteration. Explore the technical innovations including a mixed primitives system combining homomorphic encryption and secret sharing, along with the novel ChaosTable data structure that protects privacy during cross-party computation while eliminating network traffic related to edges. Understand how GraphAce addresses security challenges when utilizing local graph data from multiple parties, enabling applications in importance measures, fraud detection, and anti-money laundering without revealing individual data sources. Examine the impressive performance improvements showing speed increases of up to tens of thousands of times over previous methods while saving up to 99.99% in communication costs, making it the first secure framework capable of processing over 1 million vertices and 132 million edges in reasonable time - representing a 128x improvement in scale over previous reports.
