Fast Particle-based Anomaly Detection Algorithm with Variational Autoencoder

TL;DR

This paper introduces Set-VAE, a particle-based variational autoencoder for anomaly detection that improves signal efficiency and reduces inference latency, with potential applications in physics trigger systems.

Contribution

The paper presents Set-VAE and CLIP-VAE, novel particle-based VAE algorithms that enhance anomaly detection efficiency and speed for physics applications.

Findings

2x signal efficiency gain over traditional methods

CLIP-VAE reduces inference latency by 2x

Lower caching requirements for real-time deployment

Abstract

Model-agnostic anomaly detection is one of the promising approaches in the search for new beyond the standard model physics. In this paper, we present Set-VAE, a particle-based variational autoencoder (VAE) anomaly detection algorithm. We demonstrate a 2x signal efficiency gain compared with traditional subjettiness-based jet selection. Furthermore, with an eye to the future deployment to trigger systems, we propose the CLIP-VAE, which reduces the inference-time cost of anomaly detection by using the KL-divergence loss as the anomaly score, resulting in a 2x acceleration in latency and reducing the caching requirement.