Beyond optimization—supervised learning applications in relativistic laser-plasma experiments

We explore the applications of machine learning techniques in relativistic laser-plasma experiments beyond optimization purposes. predict beam charge electrons produced a laser wakefield accelerator given wavefront change caused by deformable mirror. Machine enables feature analysis merely searching for an optimal charge, showing that specific aberrations are favored generating higher charges. Supervised models allow characterizing measured data quality as well recognizing irreproducible and potential outliers. also include virtual measurement errors experimental to examine model robustness under these conditions. This work demonstrates how methods can benefit physics interpretation highly nonlinear problem interaction.