VERTICES: Efficient Two-Party Vertical Federated Linear Model with TTP-aided Secret Sharing

TL;DR

This paper introduces VERTICES, an efficient two-party vertical federated linear model leveraging TTP-aided secret sharing, significantly reducing training time compared to existing MPC-based methods.

Contribution

It proposes a novel secret sharing scheme with a trusted coordinator for VFL, achieving substantial speedups over traditional MPC frameworks.

Findings

Training acceleration of 2.5x to 6.6x compared to existing MPC methods.

Maintains strong privacy guarantees with improved efficiency.

Applicable to two-party vertical federated linear models.

Abstract

Vertical Federated Learning (VFL) has emerged as one of the most predominant approaches for secure collaborative machine learning where the training data is partitioned by features among multiple parties. Most VFL algorithms primarily rely on two fundamental privacy-preserving techniques: Homomorphic Encryption (HE) and secure Multi-Party Computation (MPC). Though generally considered with stronger privacy guarantees, existing general-purpose MPC frameworks suffer from expensive computation and communication overhead and are inefficient especially under VFL settings. This study centers around MPC-based VFL algorithms and presents a novel approach for two-party vertical federated linear models via an efficient secret sharing (SS) scheme with a trusted coordinator. Our approach can achieve significant acceleration of the training procedure in vertical federated linear models of between 2.5x and 6.6x than other existing MPC frameworks under the same security setting.