Privacy preserving poisoning attack detection in federated learning using CKKS homomorphic encryption & hybrid approach
Abstract
The field of federated learning (FL) has seen growing adoption because it provides privacy-protected methods for training machine learning models across distributed datasets. FL systems are exposed to poisoning attacks because malicious clients use manipulated model updates to harm system performance while compromising security. The study introduces a privacy- preserving approach to detect poisoning attacks inside federated learning through Intrusion Detection Systems (IDS). It utilize homomorphic encryption to protect privacy during model updates and using a combine technique that is apply autoencoders for anomaly detection, with clustering techniques to identify outliers indicative of malicious activity. The goal is to create an encryption based security system which protects federation learning process while efficiently identifying attack attempts in diverse network environment. This research fills a gap in existing methods by integrating privacy-preserving techniques with advanced anomaly detection to safeguard IDS models in federated settings.
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.