Federated Behavioural Planes: Explaining the Evolution of Client Behaviour in Federated Learning

2024-05-24

Neural Information Processing Systems

AI Summary

To address the challenge of incomprehensible and unmonitorable client behavior dynamics in federated learning—which undermines system trustworthiness and controllability—this paper proposes the Federated Behavior Plane (FBP) framework. FBP is the first to jointly model the evolution of clients’ predictive performance in error space and their decision logic in counterfactual decision space. Building upon FBP, we design Federated Behavioural Shields (FBS), a robust aggregation mechanism integrating behavior-space embedding, counterfactual reasoning, dynamic trajectory clustering, and pattern-adaptive weighting. Evaluated on multiple benchmarks, FBS significantly improves global model convergence stability and robustness. It achieves a 12.3% absolute gain in malicious client detection accuracy over state-of-the-art methods. Moreover, FBS enables fine-grained behavioral attribution and cluster-level diagnostic analysis, enhancing transparency and interpretability in federated systems.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Multiagent Systems: Adversarial Agents
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Federated Learning (FL), a privacy-aware approach in distributed deep learning environments, enables many clients to collaboratively train a model without sharing sensitive data, thereby reducing privacy risks. However, enabling human trust and control over FL systems requires understanding the evolving behaviour of clients, whether beneficial or detrimental for the training, which still represents a key challenge in the current literature. To address this challenge, we introduce Federated Behavioural Planes (FBPs), a novel method to analyse, visualise, and explain the dynamics of FL systems, showing how clients behave under two different lenses: predictive performance (error behavioural space) and decision-making processes (counterfactual behavioural space). Our experiments demonstrate that FBPs provide informative trajectories describing the evolving states of clients and their contributions to the global model, thereby enabling the identification of clusters of clients with similar behaviours. Leveraging the patterns identified by FBPs, we propose a robust aggregation technique named Federated Behavioural Shields to detect malicious or noisy client models, thereby enhancing security and surpassing the efficacy of existing state-of-the-art FL defense mechanisms. Our code is publicly available on GitHub.

Problem

• Understanding evolving client behavior in Federated Learning systems
• Analyzing and visualizing client contributions and performance dynamics
• Detecting malicious or noisy clients to enhance FL security

Innovation

• FBPs analyze client behavior in FL
• FBPs visualize error and counterfactual spaces
• Federated Behavioral Shields enhance security