Speaker
Description
We report on new developments of the PIConGPU code with regards to coupling laser plasma accelerator simulations to large-scale AI models, ML-assisted date reduction, ML-guided optimization and surrogate modeling.
We report research highlights in laser acceleration of electrons and adjacent fields to showcase the breadth of applications covered by PIConGPU. Specifically, we report on extending the Traveling Wave Electron Acceleration Scheme towards staging, diagnostics integration for beam reconstruction and optimization for applications such as compact FELs.
We discuss how the integration of AI with HPC fosters progress in these applications beyond laser-driven lepton acceleration, e.g. for laser ion acceleration. We discuss the importance of common input/output standards for this and conclude by an outlook towards coupling simulation and experiment.
| Working group | WG1 |
|---|