Trigger Optimization and Event Classification for Dark Matter Searches in the CYGNO Experiment Using Machine Learning

TL;DR

This paper presents machine learning methods for real-time data reduction and event classification in the CYGNO dark matter search experiment, improving trigger efficiency and background discrimination using autoencoders and weakly supervised classifiers.

Contribution

It introduces a novel unsupervised autoencoder-based anomaly detection and a weakly supervised classification approach for dark matter event identification in optical TPC data.

Findings

Autoencoder retains 93% of signal while discarding 97.8% of image area.

Autoencoder inference takes ~25 ms per frame on a consumer GPU.

Classifier approaches the theoretical performance limit for neutron recoil identification.

Abstract

The CYGNO experiment employs an optical-readout Time Projection Chamber (TPC) to search for rare low-energy interactions using finely resolved scintillation images. While the optical readout provides rich topological information, it produces large, sparse megapixel images that challenge real-time triggering, data reduction, and background discrimination. We summarize two complementary machine-learning approaches developed within CYGNO. First, we present a fast and fully unsupervised strategy for online data reduction based on reconstruction-based anomaly detection. A convolutional autoencoder trained exclusively on pedestal images (i.e. frames acquired with GEM amplification disabled) learns the detector noise morphology and highlights particle-induced structures through localized reconstruction residuals, from which compact Regions of Interest (ROIs) are extracted. On real prototype data, the selected configuration retains (93.0 +/- 0.2)% of reconstructed signal intensity while discarding (97.8 +/- 0.1)% of the image area, with ~25 ms per-frame inference time on a consumer GPU. Second, we report a weakly supervised application of the Classification Without Labels (CWoLa) framework to data acquired with an Americium--Beryllium neutron source. Using only mixed AmBe and standard datasets (no event-level labels), a convolutional classifier learns to identify nuclear-recoil-like topologies. The achieved performance approaches the theoretical limit imposed by the mixture composition and isolates a high-score population with compact, approximately circular morphologies consistent with nuclear recoils.