Anomaly Detection With Conditional Variational Autoencoders

TL;DR

This paper introduces a novel conditional variational autoencoder approach with a specialized loss function and metric for hierarchical anomaly detection, demonstrating superior performance on benchmarks and CERN LHC data.

Contribution

It proposes an original CVAE-based framework with a new loss and metric tailored for hierarchical anomaly detection, addressing limitations of previous VAE methods.

Findings

Outperforms classical ML benchmarks in anomaly detection tasks.

Effective in real-world CERN LHC trigger system monitoring.

Demonstrates superiority over existing VAE-based methods.

Abstract

Exploiting the rapid advances in probabilistic inference, in particular variational Bayes and variational autoencoders (VAEs), for anomaly detection (AD) tasks remains an open research question. Previous works argued that training VAE models only with inliers is insufficient and the framework should be significantly modified in order to discriminate the anomalous instances. In this work, we exploit the deep conditional variational autoencoder (CVAE) and we define an original loss function together with a metric that targets hierarchically structured data AD. Our motivating application is a real world problem: monitoring the trigger system which is a basic component of many particle physics experiments at the CERN Large Hadron Collider (LHC). In the experiments we show the superior performance of this method for classical machine learning (ML) benchmarks and for our application.