SoK: Privacy-Preserving Collaborative Tree-based Model Learning

TL;DR

This survey reviews privacy-preserving methods for collaborative training of tree-based models, analyzing their techniques, threat models, and information leakage to guide future research in secure machine learning.

Contribution

It systematically categorizes existing approaches and introduces a novel framework for analyzing information leakage in distributed tree-based model learning.

Findings

Identifies strengths and limitations of current privacy-preserving methods

Provides a framework for analyzing information leakage

Highlights gaps in existing security guarantees

Abstract

Tree-based models are among the most efficient machine learning techniques for data mining nowadays due to their accuracy, interpretability, and simplicity. The recent orthogonal needs for more data and privacy protection call for collaborative privacy-preserving solutions. In this work, we survey the literature on distributed and privacy-preserving training of tree-based models and we systematize its knowledge based on four axes: the learning algorithm, the collaborative model, the protection mechanism, and the threat model. We use this to identify the strengths and limitations of these works and provide for the first time a framework analyzing the information leakage occurring in distributed tree-based model learning.