Machine learning phase and glass transitions by Prof. Arun Yethiraj скачать в хорошем качестве

Machine learning phase and glass transitions by Prof. Arun Yethiraj

Abstract : " Artificial intelligence and machine learning is all the rage. AI is taking over entry-level software engineering and many other jobs. Can ML provide insight that will help computational physicists and chemists become more effective? Will it replace human learning and put us out of business? In this talk I will discuss efforts in my group to apply ML methods to two problems: The phase behavior of complex fluid mixtures and the dynamics of supercooled liquids. I will describe unsupervised methods that are quite successful for these problems, i.e., they are accurate and provide significant physical insight. They do require insightful application, however, and are not primed to put us out of business just yet. "