RadiaSoft has launched a new toolbox in Sirepo for machine learning called Activait. In this webinar, Senior Research Scientist Jon Edelen will demonstrate the construction, testing, and deployment of ML models and classifiers in this new toolbox.
Starting with a contextual overview of how ML models and classifiers can help particle accelerator operations, Jon will then do a live demo of our machine learning tools in Activait focused on these two applications.
First, he’ll train a neural network to predict RMS beam sizes in a drift tube linac. Second, he’ll test different classification techniques for differentiating between injector settings with high losses and low losses in a photocathode RF gun. Then, he’ll wrap up by showing you how to export code from Activait and execute it on our Jupyter servers for further analysis and customization.