Localized anomaly detection with a calorimeter normalizing flow
Abstract
Normalizing flows trained as fast calorimeter simulators also give the likelihood of each shower, which can serve as an anomaly score. Because this likelihood sums over every voxel, a signal confined to a few voxels can be lost in the fluctuations of the rest of the shower. For example, in a public benchmark dataset in which a long-lived neutral particle decays to two photons, the CaloFlow likelihood used in previous work selects less than 5% of the signal at a background efficiency of 10−3 when the particle has the largest simulated mass and decays at or just behind the front face of the calorimeter. In this work, we define a regional anomaly score by summing the squared latent coordinates of the same flow within small detector regions around the point where the particle enters the calorimeter, with the threshold set using background showers alone. At the same background efficiency, this regional anomaly score selects more than 95% of the signal. The per-coordinate likelihood terms, summed over the same regions, perform equally well, but the squared latent coordinates have a known distribution for background showers and clearly illustrate when a shower departs from it. Therefore, normalizing flows may be sensitive to new physics, even if the global likelihood is not.
Related
- Normalizing Flow Emulator for Gravitational Wave Detection Probability in the LIGO/Virgo/KAGRA O4 Observing Run
- Knowledge Distillation of a Normalising Flow for Real-Time Anomaly Detection at LHC Level-1 Trigger
- Fast and Flow-rious: Gravitational Wave Background Bayesian Model Comparison using Normalizing Flows and Nested Sampling
- Normalizing Flow-Based Bayesian Parameter Estimation for Noisy Quantum States
- Conditional Normalizing Flow for Gas-Surface Scattering from Thermal to Hypersonic Velocities
BibTeX
@misc{mendiratta2026localized,
title = {{Localized anomaly detection with a calorimeter normalizing flow}},
author = {Raunav Mendiratta and Jason Weitz and Javier Duarte},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.00234},
eprint = {2611.00234},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.00234},
}
RIS
TY - UNPB
TI - Localized anomaly detection with a calorimeter normalizing flow
AU - Raunav Mendiratta
AU - Jason Weitz
AU - Javier Duarte
PY - 2026
DA - 2026/11/01
PB - arXiv
DO - 10.48550/arxiv.2611.00234
UR - https://arxiv.org/abs/2611.00234
AB - Normalizing flows trained as fast calorimeter simulators also give the likelihood of each shower, which can serve as an anomaly score. Because this likelihood sums over every voxel, a signal confined to a few voxels can be lost in the fluctuations of the rest of the shower. For example, in a public benchmark dataset in which a long-lived neutral particle decays to two photons, the CaloFlow likelihood used in previous work selects less than 5% of the signal at a background efficiency of 10^−3 when the particle has the largest simulated mass and decays at or just behind the front face of the calorimeter. In this work, we define a regional anomaly score by summing the squared latent coordinates of the same flow within small detector regions around the point where the particle enters the calorimeter, with the threshold set using background showers alone. At the same background efficiency, this regional anomaly score selects more than 95% of the signal. The per-coordinate likelihood terms, summed over the same regions, perform equally well, but the squared latent coordinates have a known distribution for background showers and clearly illustrate when a shower departs from it. Therefore, normalizing flows may be sensitive to new physics, even if the global likelihood is not.
ER -