Abstract: The presentation starts with examples of material design for nuclear fuels, battery electrodes, and computer memory gates. The methodology involves experiments and computer simulations that operate at various length and time scales, such as ab initio molecular dynamics, phase field, and finite elements. The experimental and computational data are analyzed by using machine learning algorithms and a Bayesian method that delivers uncertainty intervals.
The second part of the talk is focused on using artificial intelligence to design experiments and optimize in real time complex synthesis processes, using as an example the flame spray pyrolysis setup at MERF. The methodology includes machine learning, active learning, and computational fluid dynamics.
The presentation ends with a discussion of the impact of artificial intelligence on science and technology.