Private Two-Party Set Intersection Protocol in Rational Model

Many data mining algorithms use privacy preserving set intersection operations. Private set operations have considered semi-honest and malicious adversarial models in cryptographic settings. Protocols in semi-honest model, requiring light computations, provide weak security. Protocols in malicious model guarantee strong security at the price of expensive computations like homomorphic encryption and zero-knowledge proof. However, practical implementations require robust and efficient protocols. In this paper, we build efficient and private set intersection avoiding the use of expensive tools like homomorphic encryption and zero-knowledge proof. Our proposed set intersection protocol is constructed in game-theoretic model. In our model, the parties are viewed as rational whereby they are assumed (only) to act in their self-interest. Our protocol satisfies computational Nash equilibrium.