Analysis and design of materials with machine learning: from probabilistic methods to quantum computing opportunities – Dr Miguel Bessa, Materials Science and Engineering, Delft University of Technology

Analysis and design of materials can be significantly empowered by machine learning. This talk discusses how three novel machine learning methods can be used to design new materials and analyze history-dependent physics. In the first example, Bayesian machine learning is shown to guide the design of a new lightweight, recoverable and super-compressible metamaterial achieving more than 90% compressive strain without damage. The second example focuses on how deep learning can predict path-dependent material plasticity. And the final example presents a new quantum machine learning algorithm that can break the curse of dimensionality that plagues Gaussian processes.