Local differential privacy for federated learning

DOI10.1007/978-3-031-17140-6_10MaRDI QIDQ6173289zbMATH OpenFDO

Authors Pathum Chamikara Mahawaga Arachchige, Dongxi Liu, Seyit Camtepe, Surya Nepal, Marthie Grobler, Peter Bertok, Ibrahim Khalil

Publication date 21 July 2023

Published in Computer Security – ESORICS 2022 (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2202.06053

zbMATH Keywords

differential privacydistributed machine learningfederated learninglocal differential privacyprivacy-preserving distributed machine learningprivacy-preserving federated learning

Mathematics Subject Classification ID

Privacy of data (68P27) Learning and adaptive systems in artificial intelligence (68T05)

Abstract: Advanced adversarial attacks such as membership inference and model memorization can make federated learning (FL) vulnerable and potentially leak sensitive private data. Local differentially private (LDP) approaches are gaining more popularity due to stronger privacy notions and native support for data distribution compared to other differentially private (DP) solutions. However, DP approaches assume that the FL server (that aggregates the models) is honest (run the FL protocol honestly) or semi-honest (run the FL protocol honestly while also trying to learn as much information as possible). These assumptions make such approaches unrealistic and unreliable for real-world settings. Besides, in real-world industrial environments (e.g., healthcare), the distributed entities (e.g., hospitals) are already composed of locally running machine learning models (this setting is also referred to as the cross-silo setting). Existing approaches do not provide a scalable mechanism for privacy-preserving FL to be utilized under such settings, potentially with untrusted parties. This paper proposes a new local differentially private FL (named LDPFL) protocol for industrial settings. LDPFL can run in industrial settings with untrusted entities while enforcing stronger privacy guarantees than existing approaches. LDPFL shows high FL model performance (up to 98%) under small privacy budgets (e.g., epsilon = 0.5) in comparison to existing methods.

Recommendations

• Privacy-preserving federated learning on lattice quantization
• FL-MAC-RDP: federated learning over multiple access channels with Rényi differential privacy
• Utility optimization of federated learning with differential privacy
• CECMLP: new cipher-based evaluating collaborative multi-layer perceptron scheme in federated learning
• An optimal (, )-differentially private learning of distributed deep fuzzy models

Cites work

• Concentrated differential privacy: simplifications, extensions, and lower bounds
• Extremal mechanisms for local differential privacy
• Fully homomorphic encryption using ideal lattices

Cited in

(9)

• Utility optimization of federated learning with differential privacy
• FL-MAC-RDP: federated learning over multiple access channels with Rényi differential privacy
• scientific article; zbMATH DE number 7233838 (Why is no real title available?)
• Differentially Private Distributed Learning
• Advances and Open Problems in Federated Learning
• scientific article; zbMATH DE number 7415075 (Why is no real title available?)
• Privacy-preserving federated learning on lattice quantization
• PFLM: privacy-preserving federated learning with membership proof
• CORK: a privacy-preserving and lossless federated learning scheme for deep neural network

This page was built for publication: Local differential privacy for federated learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6173289)