RobustFSM: Submodular Maximization in Federated Setting with Malicious Clients

2025-11-03

AI Summary

Malicious clients in federated learning can inject false information, severely degrading the robustness of submodular maximization—a critical task for representative subset selection under privacy constraints. Method: This paper proposes the first robust optimization framework specifically designed for federated submodular maximization. It integrates a robust aggregation mechanism with a dynamic client behavior assessment strategy within an iterative federated architecture, enabling secure information fusion and high-fidelity subset selection. The framework employs theoretically grounded outlier detection and weighted aggregation to withstand practical attacks, including Byzantine attacks and gradient poisoning. Results: Extensive experiments on multiple real-world datasets demonstrate that, under strong adversarial conditions, the proposed method improves subset selection quality by up to 200% over baseline federated submodular algorithms and significantly outperforms existing robust aggregation schemes. This work establishes a new paradigm for privacy-preserving, distributed representative sampling.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Submodular maximization is an optimization problem benefiting many machine learning applications, where we seek a small subset best representing an extremely large dataset. We focus on the federated setting where the data are locally owned by decentralized clients who have their own definitions for the quality of representability. This setting requires repetitive aggregation of local information computed by the clients. While the main motivation is to respect the privacy and autonomy of the clients, the federated setting is vulnerable to client misbehaviors: malicious clients might share fake information. An analogy is backdoor attack in conventional federated learning, but our challenge differs freshly due to the unique characteristics of submodular maximization. We propose RobustFSM, a federated submodular maximization solution that is robust to various practical client attacks. Its performance is substantiated with an empirical evaluation study using real-world datasets. Numerical results show that the solution quality of RobustFSM substantially exceeds that of the conventional federated algorithm when attacks are severe. The degree of this improvement depends on the dataset and attack scenarios, which can be as high as 200%

Problem

• Federated submodular maximization vulnerable to malicious clients
• RobustFSM defends against client attacks in distributed optimization
• Solution improves performance by up to 200% under severe attacks

Innovation

• Federated submodular maximization with malicious clients
• RobustFSM solution against various client attacks
• Empirical evaluation using real-world datasets validation