Security and Credit evaluation mechanism federated learning privacy protection scheme
Abstract
Federated learning has been widely applied in fields such as government affairs, transportation, and healthcare due to its advantages in protecting the privacy of multiple parties while achieving secure modeling. It effectively solves the problem of information silos in the industry. This paper proposes a privacy protection scheme for federated learning based on homomorphic encryption and BLS signature technology, in response to the lack of gradient protection and the easy tampering of aggregation models in existing federated learning. In this paper, homomorphic encryption is used to protect the confidentiality of gradients uploaded by the client.; using BLS signature to protect the integrity of the aggregation model and prevent attackers from tampering with the aggregation model; the reputation evaluation algorithm can identify and reward participants who provide high-quality data and actively participate in the training process. Finally, Experimental evaluation shows that our approach significantly improves the protection of gradient and aggregation models in terms of security, while controlling performance loss within an acceptable range. Compared with existing solutions, our approach achieves better performance while ensuring high security.
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.