Workshop: Design of Experiment

Designing experiments efficiently is a key challenge in research, particularly when time, materials, or measurement capacity are limited. Design of Experiments (DoE) offers a structured approach to planning experiments more effectively. This workshop will cover the basics of DoE and demonstrate how Bayesian Optimization can further improve experimental design. In a hands-on session, we will explore BayBE (https://github.com/emdgroup/baybe) - an open-source Python library for Bayesian Optimization developed by Merck KGaA. Through examples from practical applications, you will see how BayBE supports scientists in navigating complex experimental spaces, accelerating discovery, and improving outcomes even when data is scarce. 

A technical documentation can be found at https://emdgroup.github.io/baybe/stable/index.html.

Registration

Register for the workshop during the registration process for the PhD seminar. Number of participants is limited to 25. An additional fee is taken for the workshop.

Schedule

25.09.2026 (full day)

Start is at 9 a.m. (same room as used during the PhD seminar).

End is at 3 p.m. latest.

Prerequisites

There are no must-have prerequisites. Potential installations of Python packages can be done at the beginning of the workshop. A basic understanding of Python is desirable to run and follow the script and to get the most out of this workshop. Please bring your own laptop.

You can choose to run the script via Google Colab or locally. Google Colab is the preferred choice as it requires minimal setup. A Google account is needed in order to run the scripts, which can be created for free upfront.

If you choose to work locally, you should already be familiar with Python and have an up-to-date Python installation and an IDE of your choice. This workshop does not cover basic Python or IDE setup.

The link to the Python script will be provided in the workshop.

To easily connect and work with our Raspberry Pi equipment, we will use Thonny (https://thonny.org/). This can be downloaded and installed upfront.