Speaker
Description
The integration of advanced artificial intelligence (AI) techniques into astroparticle experiments marks a transformative step in both data analysis and experimental design. As space missions grow increasingly complex, the adoption of AI technologies becomes critical for optimizing performance and achieving robust scientific outcomes. In this context, we explore two innovative AI-driven approaches tailored for space-based calorimetric and tracking systems.
Firstly, we propose a fully custom-designed Transformer-based model for particle identification in space calorimeters. A key challenge in these experiments is the distinction between particle types, such as electrons and protons, based on energy deposition patterns. By capturing long-range dependencies across thousands of input channels, Transformers offer a powerful framework for robust classification. Our approach aims to enhance both the accuracy and reliability of particle identification, with the potential to extend classification capabilities across a wide energy spectrum, from 1 GeV to 100 TeV.
Secondly, we address the challenge of tracking in noisy environments by introducing a Graph Neural Network (GNN)-based solution. Tracking detectors in space are often affected by high levels of noise, including backscattering hits from the calorimeter and electronic noise, which complicate the reconstruction of primary particle trajectories. Leveraging the graph-based structure of tracking systems—where nodes represent energy deposits (hits) and edges encode their relationships—our GNN model performs node-level classification to distinguish signal hits from noise. This enables efficient and accurate reconstruction of the primary tracks and their parameters.
By addressing these complementary challenges in calorimetry and tracking, our work demonstrates the impact of state-of-the-art AI methodologies in advancing the capabilities of future space-based astroparticle experiments.
Details
N/A
| Internet talk | No |
|---|---|
| Is this an abstract from experimental collaboration? | No |
| Name of experiment and experimental site | N/A |
| Is the speaker for that presentation defined? | Yes |