Assessment of few-hits machine learning classification algorithms for low energy physics in liquid argon detectors
This work addresses classification bottlenecks for physicists analyzing low-energy data in neutrino experiments like DUNE, though it is incremental as it applies existing ML methods to a specific domain problem.
The paper tackled the challenge of classifying few-hits events in low-energy physics using liquid argon detectors, demonstrating that Convolutional Neural Networks and Transformer-Encoder methods outperform conventional deterministic algorithms, with specific gains in distinguishing single- versus double-beta events.
The physics potential of massive liquid argon TPCs in the low-energy regime is still to be fully reaped because few-hits events encode information that can hardly be exploited by conventional classification algorithms. Machine learning (ML) techniques give their best in these types of classification problems. In this paper, we evaluate their performance against conventional (deterministic) algorithms. We demonstrate that both Convolutional Neural Networks (CNN) and Transformer-Encoder methods outperform deterministic algorithms in one of the most challenging classification problems of low-energy physics (single- versus double-beta events). We discuss the advantages and pitfalls of Transformer-Encoder methods versus CNN and employ these methods to optimize the detector parameters, with an emphasis on the DUNE Phase II detectors ("Module of Opportunity").