Efficient Homomorphic Attribute Federated Learning Privacy Protection Algorithm for Heterogeneous Device
Abstract
This paper explores the challenges faced in Federated Learning (FL) with a focus on privacy protection and the effective management of heterogeneous devices. FL, while promoting a privacy-centric model by using decentralized data for training, encounters significant risks due to the potential leakage of sensitive information from gradients shared among users. These challenges are exacerbated in environments with devices of varied computational capabilities and diverse network conditions. Existing solutions in FL are insufficient in effectively addressing these dual concerns of privacy and device heterogeneity. To address these limitations, we introduce the Efficient Quantization Homomorphic Federated Learning Framework (EQ-HFL). This innovative solution is specifically designed to enhance the efficiency and utility of FL in environments with heterogeneous devices. EQ- HFL stands out by proactively involving these devices in the encryption process and customizing homomorphic encryption quantization to suit their capacities. Additionally, our framework incorporates a novel data detection algorithm that intelligently assigns weights to client gradients, tailored to their unique data distribution. This approach is particularly effective in managing the challenge of non-identically distributed (non-iid) data, which is a common hurdle in FL. A key innovation of EQ-HFL is its hierarchical quantization scheme, which significantly speeds up the homomorphic encryption and decryption…
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.