FederatedTrustee: Pos-Hoc Explainability for Trustworthy Federated Learning

• Grant number: 25/11909-7
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2025
• End date: July 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Ronaldo Alves Ferreira
• Grantee: Maria Luiza Brito Pagliosa
• Host Institution: Faculdade de Computação. Universidade Federal de Mato Grosso do Sul (UFMS). Campo Grande , SP, Brazil
• Associated research grant: 23/00811-0 - EcoSustain - Computer and Data Science for the Environment, AP.TEM

Abstract

Federated Learning (FL) enables decentralized model training while preserving user privacy. However, it poses unresolved challenges in trust management and model interpretability, as servers have limited visibility into the quality and intent of client updates. This project aims to develop FederatedTrustee, a tool that incorporates explainable AI (XAI) techniques into the FL pipeline to enhance transparency, robustness, and reliability. Inspired by global post-hoc methods such as Trustee, the tool will assess client contributions and generate interpretable surrogate models to explain the behavior of the aggregated model. The project will also investigate privacy-preserving strategies to mitigate potential leakage from explanation mechanisms like decision trees.