~/giac-gmle · study ▊
practice sheet · all objectives switch to smart practice →
GIAC Machine Learning Engineer
0/3 checked
A security analyst needs to build a system that flags anomalies in log data without relying on any predefined signatures or human tuning of detection rules. Which deep learning architecture, as taught in the course, is specifically applied to this signature-free log anomaly detection task?
Explanation
The course applies autoencoder knowledge—latent representation and reconstruction loss functions—to build log anomaly detection that requires no signatures or human intervention [2]. Section 5 explicitly covers reconstruction loss functions for signature-free anomaly detection [1].
During Section 4, students build a real-time network protocol classification system and then add anomaly detection to it. Why is this described as unusual relative to typical practice?
Explanation
Section 4 focuses on supervised learning, and students implement anomaly detection within the classification system—explicitly noted as a task typically reserved for unsupervised approaches [2][1].
Locked question — unlock this exam to view it.
locked — unlock to review