VVN Talk: "Physics, AI, and the route to new materials" by Michael Sluydts скачать в хорошем качестве

VVN Talk: "Physics, AI, and the route to new materials" by Michael Sluydts

Technological advancements go time and time again hand-in-hand with the discovery of new materials. A good example was the necessity for improved steel at the start of the second industrial revolution at the end of the 19th century, and the development of semi-conductors we now use every day in our computers, tablets, and smartphones. But the demand for technological advancements, despite their development at an exponential rate, remains a necessity for all fields of science and industry. The material industry, for example, holds the promise for increasingly efficient solar cells and more sustainable batteries, but also more exotic technologies like superconductors and thermoelectrics. The search for new materials isn’t always smooth sailing. There are many different materials and only a select few are promising in terms of usability in the industry. The combination between experience and intuition can guide us to select the most promising materials, but in practice physics often falls short through the complex properties of said materials. In order to solve this, we can probe for solutions using artificial intelligence since modern AI can cross-reference incredibly large chunks of data and find complex relationships between them. At the same time, on the quantum side of things, we can use quantum mechanical simulations in order to predict the inherent properties of many different materials to learn about them. This knowledge can be used by AI to develop a new kind of ‘physical intuition’, which makes it possible to find new interesting materials thousands of times faster than by running simulations ourselves or through experimental research. This lecture is given to us by Dr. Ir. Michael Sluydts. He works on the development of new materials for energy applications at the Center for Molecular Modeling (CMM) of the UGent in collaboration with Umicore, and he simulates the uses for AI, specifically developed for the science engineering industry, with his company ‘ePotentia’.