Autoencoders on FPGAs for real-time, unsupervised new physics detection at 40 MHz at the Large Hadron Collider

TL;DR

This paper presents a method to implement deep autoencoder-based anomaly detection on FPGAs for real-time identification of new physics signatures at the LHC, achieving high sensitivity within strict latency and resource constraints.

Contribution

It introduces a novel FPGA implementation of autoencoders for real-time, unsupervised physics anomaly detection, enabling rapid and resource-efficient data filtering at the LHC.

Findings

Anomaly detection achieved in 80 ns on FPGA

Supports high-purity new physics dataset collection

Uses less than 3% of FPGA resources

Abstract

In this paper, we show how to adapt and deploy anomaly detection algorithms based on deep autoencoders, for the unsupervised detection of new physics signatures in the extremely challenging environment of a real-time event selection system at the Large Hadron Collider (LHC). We demonstrate that new physics signatures can be enhanced by three orders of magnitude, while staying within the strict latency and resource constraints of a typical LHC event filtering system. This would allow for collecting datasets potentially enriched with high-purity contributions from new physics processes. Through per-layer, highly parallel implementations of network layers, support for autoencoder-specific losses on FPGAs and latent space based inference, we demonstrate that anomaly detection can be performed in as little as $80\,$ns using less than 3% of the logic resources in the Xilinx Virtex VU9P FPGA. Opening the way to real-life applications of this idea during the next data-taking campaign of the LHC.