Fast multi-geometry calorimeter simulation with conditional self-attention variational autoencoders

Dylan Smith, Aishik Ghosh, Junze Liu, Pierre Baldi, Daniel Whiteson

arXiv:2411.05996·hep-ex·Published 2024-11-08

The simulation of detector response is a vital aspect of data analysis in particle physics, but current Monte Carlo methods are computationally expensive. Machine learning methods, which learn a mapping from incident particle to detector response, are much faster but require a model for every detector element with unique geometry. Complex geometries may require many models, each with their own training samples and hyperparameter tuning tasks. A promising approach is the use of geometry-aware models, which condition the response on the geometry, but current efforts typically require cumbersome full geometry specification. We present a geometry-aware model that takes advantage of the regularity of detector segments, requiring only the definition of cell sizes across regular segments. This model outperforms the current state of the art by over 70% across several key metrics including the Wasserstein distance metric.

TopicsParticle & High Energy Physics

Tagsparticle-physics

arXiv categorieshep-ex

arXiv abstract pagePDF