A comparison of classical and variational autoencoders for anomaly detection

TL;DR

This paper compares classical and variational autoencoders for anomaly detection by analyzing their architecture, properties, and performance on a simple line reconstruction task.

Contribution

It provides a detailed comparison of classical and variational autoencoders, highlighting their differences in architecture and effectiveness for anomaly detection.

Findings

Variational autoencoders better capture data variability.

Classical autoencoders are simpler but less flexible.

Performance varies depending on the reconstruction task.

Abstract

This paper analyzes and compares a classical and a variational autoencoder in the context of anomaly detection. To better understand their architecture and functioning, describe their properties and compare their performance, it explores how they address a simple problem: reconstructing a line with a slope.