Differentiable programming for particle physics simulations

Abstract

We describe how to apply adjoint sensitivity methods to backward Monte-Carlo schemes arising from simulations of particles passing through matter. Relying on this, we demonstrate derivative based techniques for solving inverse problems for such systems without approximations to underlying transport dynamics. We are implementing those algorithms for various scenarios within a general purpose differentiable programming C++17 library NOA (github.com/grinisrit/noa).

arXiv comment12 pages, 3 figures, presented at QUARKS online workshops 2021, initial version, comments welcome

References in corpus (4)

• Explicit high-order symplectic integrators for charged particles in general electromagnetic fields
• A General Metric for Riemannian Manifold Hamiltonian Monte Carlo
• Backward Monte-Carlo applied to muon transport
• Introducing an Explicit Symplectic Integration Scheme for Riemannian Manifold Hamiltonian Monte Carlo

Cited by in corpus (2)

• A differentiable programming framework for spin models
• Deep-learning-based decomposition of overlapping-sparse images: application at the vertex of neutrino interactions