option 2: To identify malicious learning nodes