26.07.2025 –, Workshop 2 (2. OG)
Sprache: English
Could we put everything we know about materials properties and the periodic table into a neural network, and finally get the structure of a room temperature superconductor? There has been a massive amount of funding influx for scientific research using AI/ML to solve global issues. One use case is applying ML to design new materials for applications in energy and sustainability (e.g. lithium-free batteries, rare-earth magnet replacements for windmills). To explore this, first I will outline the kinds of problems materials scientists are solving. Then I will critically examine what "sustainability" even means in the context of materials science through a case study of a German chemical company. Lastly I will discuss my favorite research examples (from academia to my life) outlining the key ways we can use AI/ML to accelerate materials discovery. I will highlight the huge potential I see as well as some pitfalls, and quickly share my favorite open-access materials science related software packages / databases.
If there is time / interest after questions I can give a short introduction into materials simulation - how you can download structure information online and use packages to build materials and predict some basic properties.
Trying to get the periodic table to do what I want using material simulations. Interested in open-source tool development, making science more accessible, and innovating more sustainable technologies to fight climate change. M.Sc. student in Materials Science at the TU Darmstadt / INP Grenoble.