HZB Machine Learning Summer School 2020

HZB Machine Learning Summer School 2020

Montag, 10. August 2020 - Freitag, 14. August 2020

Helmholtz-Zentrum Berlin für Materialien und Energie

Logo: Helmholtz-Zentrum Berlin (HZB)

We are happy to announce the first HZB Machine Learning Summer School & Hackathon 2020, which will be held virtual.

We want to bring together master and PhD students from computer science/informatics with early career scientists (Ph.D. to early Postdoc) from large-scale facility-based research centers.

Use cases range from multi-parameter optimization of highly complex state-of-the-art research instruments up to applications of high-intensity free-electron lasers.

The Summer School has the following format:

  • Two-day lecture series on machine learning basics, including techniques of deep learning and reinforcement learning. Participants should finally be able to develop a "hands-on AI method competence".
  • Three-day hackathon with projects from the area "AI at large research facilities", such as:

    • Optimized design of experiment infrastructure (e.g. Beamlines for the future, highly optimized synchrotron radiation facility BESSY III)
    • Optimized use of the experiment infrastructure
      (e.g. Reinforcement Learning based on archived and live data of the storage ring or its archive or injector gun optimization based on “learned” simulations)
    • "Near-realtime-analysis"/Quick evaluation of experiments

The target group should already have strong competence and hands-on experience in computer science. We will favor applications that already have contacts across "research domains".

Solid knowledge of python is necessary.An advantage is the basic understanding of Jupyter-notebooks and open-source machine learning packages (e.g. Sklearn, Pytorch or Tensorflow).

You will need a laptop with a running python (3.6+) environment.

If you want to join, please include a short motivation statement (max. ½ page) and your CV (max. 1 page) in your application. In case you have an compelling use case for our hackathon, you may include a short abstract (max. 1 page) describing your idea. Please include directions how to access your input data. We will check the input data carefully in advance to have only feasible project ideas for the hackathon.

The summer school, as well as the hackathon, is free of charge for all participants.

Deadline forapplications is June 15, 2020 (GMT+2, midnight).

Participants will be selected based on their motivation statement, CV, and the optional abstract. You will receive notice from us before June 21, 2020.

As the number of places is limited, we still need a binding registration from you. The deadline for your final confirmation of the participation is due on June 30, 2020 (midnight, GMT+2). In case of application withdrawal, we will immediately fill up vacant positions from the waiting list.