Privacy vs Accuracy: Trade-offs in Federated Learning
الملخص
In federated learning, model updates from distributed sources are vulnerable to inference attacks that can reconstruct private training data from gradient information.Traditional approaches to this problem apply differential privacy uniformly across all model parameters, creating a significant trade-off where stronger privacy guarantees substantially degrade model utility.This thesis investigates how differential privacy can be selectively applied to gradient components to enhance privacy protection while maintaining model performance.We introduce selective gradient perturbation, a percentile-based approach that applies noise only to the most significant gradient components by magnitude.Through comprehensive experiments on the CIFAR-10 dataset using a simulated federated learning environment with 10 clients, we demonstrate that selective perturbation fundamentally transforms the conventional privacy-utility trade-off.Our findings reveal that by applying noise to only the top 10% of gradients, we achieve 74.38% accuracy while maintaining privacy protection, and counter-intuitively, stronger privacy settings (ε = 0.1) yielded better model performance (78.60% accuracy) than weaker protection (ε = 0.5, 66.56% accuracy).These results challenge the conventional wisdom that privacy necessarily comes at a significant cost to utility, offering a practical approach for implementing differential privacy in federated learning that achieves both strong privacy guarantees and high model utility.
الكلمات المفتاحية
غير متوفر
التخصصات
رسائل وأطاريح مشابهة
- Federated Learning of Fuzzy Rule-based Systems: Advances in Privacy, Security and Interpretability Mattia Daole · University of Pisa · 2026
- Federated Learning-Based Intrusion Detection System for Privacy-Preserving IoT Networks Prasanna Dhanireddy ; Khushboo Kumari · KTH Royal Institute of Technology · 2026
- Topology-dependent privacy risks in decentralized federated learning JOSÉ INÁCIO ANTUNES DE GOUVEIA · 2026
- Privacy-Preserving Federated Learning With Backdoor Resilience Yordanov Yordan · Karlsruhe Institute of Technology · 2026
- Designing for Privacy in Tabular Federated Learning Ivo Osterberg Nilsson ; Maximilian Birr Engvall · Jönköping University · 2026
- On the limits of privacy in federated learning : personalization and adversarial analysis Sargam Gupta · KTH Royal Institute of Technology · 2026
