Christoph Koch

Humboldt-Universität zu Berlin, DE



Christoph Koch received his PhD at Arizona State University in 2002 under the supervision of Prof. John Spence. He then went on to a postdoctoral position at the Max Planck Institute for Metals Research in Stuttgart, Germany. In 2011 he accepted a professorship, endowed by the Carl Zeiss Foundation at Ulm University in Germany, and since 2015 he has a full professorship at Humboldt University of Berlin, where he heads the structure research and electron microscopy group. His group operates several electron microscopes, and focuses on the development of novel imaging, diffraction, and spectroscopy techniques and their application to relevant problems in materials science and solid state physics. Christoph is co-spokesperson of FAIRmat, a Germany-wide consortium for establishing a national research data infrastructure in the area of materials science.


Machine learning in electron microscopy and spectroscopy


The application of Machine Learning (ML) has a long history in the analysis of experimental transmission electron microscopy (TEM) data, which includes images, diffraction patterns, spectra and multidimensional collections of them. One family of ML-methods with a long tradition is the collection of different multivariate statistical analysis (MSA) techniques, which include principal and independent component analysis (PCA & ICA), mainly for denoising or finding distinct chemical compositions or structure in experimental data sets. However, also deep learning techniques are being applied more and more, e.g. for automating the interpretation of experimental data or for producing more easily interpretable reduced representations of it. In this combination of lecture, demonstration and hands-on tutorial, an overview of a range of different ML-methods being applied in the field of TEM will be given, the potential that the novel materials discovery laboratory (NOMAD) and its online tools have for both computational and experimental materials research will be demonstrated, and a hands-on tutorial, where you will be given a chance to use these tools, will be given. 


  • Presentation about various applications of ML in the field of electron microscopy

  • Demonstration of the NOMAD platform for computational and experimental materials data

  • Hands-on tutorial of using NOMAD and the NOMAD Remote Tools Hub (NORTH) for training and using an artificial neural network to interpret experimental data.
Scroll to Top