Speaker
Description
The study of multi-Higgs production is essential for probing the Higgs boson’s self-interactions and exploring potential physics beyond the Standard Model. In this work, we investigate Tri-Higgs events decaying into six b-quarks and propose a novel machine learning approach to enhance experimental sensitivity. Our method utilizes the Symmetry Preserving Attention Network (SPANet). This neural network architecture is specifically designed to incorporate inherent symmetries in particle reconstruction tasks, improving jet pairing and event classification. By leveraging SPANet, we achieve a significant performance gain over conventional Dense Neural Network-based analyses. The results highlight the potential of this approach in improving search sensitivity, paving the way for more efficient analyses in future high-energy collider experiments.
