Hands-on Lab

Introduction

In this lab, students will explore how to detect anomalies in network traffic using a Variational Autoencoder (VAE). Anomalies indicate unusual or suspicious patterns that could suggest a security breach or network attack.

Objectives

- Understand how VAEs work for anomaly detection.

- Train the VAE on network traffic data.

- Detect and analyze anomalies in a given dataset.

Lab Steps

1. Dataset Preparation:

   - Load the network traffic dataset.

   - Preprocess the data (normalize features and encode labels).

2. Train the VAE:

   - Train the VAE model using the provided dataset.

   - Plot the training and reconstruction loss.

3. Anomaly Detection:

   - Use the trained VAE to calculate anomaly scores.

   - Compare anomaly scores to a threshold to classify anomalies.

4. Evaluation:

   - Analyze the percentage of anomalies detected.

   - Visualize results using graphs.

Visualization of expected results: