Coding-Enforced Resilient and Secure Aggregation for Hierarchical Federated Learning

TL;DR

This paper introduces H-SecCoGC, a coding-based hierarchical secure aggregation scheme for federated learning that enhances robustness, privacy, and efficiency under unreliable communication conditions.

Contribution

It presents a novel coding strategy that enforces structured aggregation, improving resilience and privacy in hierarchical federated learning.

Findings

Ensures accurate global models despite communication unreliability.

Avoids partial participation issues, enhancing robustness.

Demonstrates superior performance through theoretical and experimental analysis.

Abstract

Hierarchical federated learning (HFL) has emerged as an effective paradigm to enhance link quality between clients and the server. However, ensuring model accuracy while preserving privacy under unreliable communication remains a key challenge in HFL, as the coordination among privacy noise can be randomly disrupted. To address this limitation, we propose a robust hierarchical secure aggregation scheme, termed H-SecCoGC, which integrates coding strategies to enforce structured aggregation. The proposed scheme not only ensures accurate global model construction under varying levels of privacy, but also avoids the partial participation issue, thereby significantly improving robustness, privacy preservation, and learning efficiency. Both theoretical analyses and experimental results demonstrate the superiority of our scheme under unreliable communication across arbitrarily strong privacy guarantees