Enhancing Mutual Trustworthiness in Federated Learning for Data-Rich Smart Cities

TL;DR

This work tackles trustworthiness in federated learning for data-rich smart cities by proposing a bilateral trust framework that accounts for both clients and federated servers. It combines $IQR$-based device trust, a reputation-driven server recommendation system, and a DST-based aggregation for bootstrapped server trust, all integrated through a mutual matching game to produce stable, trustworthy pairings. The approach achieves notable improvements in global model accuracy and significantly reduces selection of untrustworthy clients across multiple datasets, while providing a distributed, scalable mechanism suitable for IoT/IoV environments. The findings demonstrate the practical value of incorporating mutual trust and matching theory to enhance privacy-preserving learning in complex urban ecosystems, with future work focusing on collusion resilience and self-healing network properties.

Abstract

Federated learning is a promising collaborative and privacy-preserving machine learning approach in data-rich smart cities. Nevertheless, the inherent heterogeneity of these urban environments presents a significant challenge in selecting trustworthy clients for collaborative model training. The usage of traditional approaches, such as the random client selection technique, poses several threats to the system's integrity due to the possibility of malicious client selection. Primarily, the existing literature focuses on assessing the trustworthiness of clients, neglecting the crucial aspect of trust in federated servers. To bridge this gap, in this work, we propose a novel framework that addresses the mutual trustworthiness in federated learning by considering the trust needs of both the client and the server. Our approach entails: (1) Creating preference functions for servers and clients, allowing them to rank each other based on trust scores, (2) Establishing a reputation-based recommendation system leveraging multiple clients to assess newly connected servers, (3) Assigning credibility scores to recommending devices for better server trustworthiness measurement, (4) Developing a trust assessment mechanism for smart devices using a statistical Interquartile Range (IQR) method, (5) Designing intelligent matching algorithms considering the preferences of both parties. Based on simulation and experimental results, our approach outperforms baseline methods by increasing trust levels, global model accuracy, and reducing non-trustworthy clients in the system.

Figures (7)

• Figure 1: Bootstrapping Architecture
• Figure 2: Overview of the Proposed Method
• Figure 3: Accuracy of Global Model vs. Number of Rounds: Fashion MNIST Dataset.
• Figure 4: Accuracy of Global Model vs. Number of Rounds: MNIST Dataset.
• Figure 5: Accuracy Vs. Untrusted Client Rate: MNIST Dataset.