Secure Two-Party Feature Selection

TL;DR

This paper presents a secure four-round protocol for evaluating data value in trading scenarios, enabling privacy-preserving classification without trusted third parties, and proves its security in a specific adversary model.

Contribution

It introduces a novel provably secure protocol for privacy-preserving data valuation in machine learning classification tasks.

Findings

Protocol is secure under honest-but-curious model

Achieves data privacy during value evaluation

Efficient four-round communication protocol

Abstract

In this work, we study how to securely evaluate the value of trading data without requiring a trusted third party. We focus on the important machine learning task of classification. This leads us to propose a provably secure four-round protocol that computes the value of the data to be traded without revealing the data to the potential acquirer. The theoretical results demonstrate a number of important properties of the proposed protocol. In particular, we prove the security of the proposed protocol in the honest-but-curious adversary model.