Optimization, automation and Machine Learning for Accelerator Physics and Photon Science.

chaired by Tom Mertens (HZB)
Monday, 11 February 2019 from to (Europe/Berlin)
at HZB ( 13.10 kino )
Magnus Strasse 2, 12489 Berlin
Description
As systems increase in complexity, producing larger and complexer datasets,
modern optimization and analysis techniques combined with some form of Artificial Intelligence or Machine Learning become necessary and gain in popularity fast. This is no different in the field of accelerator physics and photon science. 

The workshop is intended to bring together accelerator physicists, beamline scientists and experimentalists to join efforts, develop new ideas and share experience. The forward-looking goal is to exchange best practices, nurture collaborations and explore capabilities to satisfy the growing interests and needs in the scientific community. 
Participants Gregor Hartmann; Simon Hirlaender; Tom Mertens; Roland Müller; Gianluca Valentino; Luis Vera Ramirez
Material
Support tom.mertens@helmholtz-berlin.de
Go to day
  • Monday, 11 February 2019
    • 10:00 - 11:15 Machine Learning tutorial
      Machine Learning tutorial in Python by Prof. Gianluca Valentino. This tutorial will cover the whole process from feature engineering, selection and dimensionality reduction to supervised & unsupervised learning (classifiers, clustering and regression) and then model evaluation. Google colaboratory will be used during this tutorial.
      Convener: Dr. Tom Mertens (HZB)
      • 10:00 ML problem design 1h15'
        Speaker: Prof. Gianluca Valentino (University of Malta)
        Material: Slides powerpoint file}
    • 11:15 - 12:00 ML in photon science 45'
      Machine learning in photon science.
      Speaker: Mr. Gregor Hartmann (Uni Kassel)
    • 12:00 - 13:30 Lunch
    • 13:30 - 14:15 First steps towards reinforcement learning supported operation. 45'
      Speaker: Mr. Simon Hirlaender (CERN)
    • 14:15 - 15:00 Automation efforts at the European XFEL 45'
      Speaker: Dr. Sergey Tomin (DESY)
    • 15:00 - 15:20 Coffee Break
    • 15:20 - 15:45 LHC beam mode classification using beam orbit and beam loss data 25'
      Speaker: Dr. Gianluca Valentino (Univeristy of Malta)
      Material: Slides powerpoint file}
    • 15:45 - 16:15 Machine learning techniques for measurement prediction at BESSYII 30'
      Speaker: Mr. Luis Vera Ramirez (HZB)