SecuFL-IoT: an adaptive privacy-preserving federated learning framework for anomaly detection in smart industrial networks

Key Points

• Anomaly detection accuracy in the smart industrial networks is significantly improved using federated learning techniques.
• Results indicate a privacy preservation rate of over 90% when applying the adaptive framework.
• Assessment utilizes a privacy-preserving mechanism within a federated learning structure to enhance efficiency.
• The framework suggests major improvements in data privacy, yet further testing in diverse environments is necessary.