By Rupert Klein
Techniques of machine learning (ML) find a rapidly increasing range of applications touching upon many aspects of everyday life. They are also used with enthusiasm to close gaps in our scientific knowledge by data-based modeling. I have followed these developments with interest, concern, and mounting disappointment. When these technologies take over decisive functionality in safety-critical applications, we should know how to guarantee their compliance with pre-defined guardrails. Moreover, when they are utilized as building blocks in scientific research, it would violate scientific standards if these building blocks were used without a thorough understanding of their functionality, including inaccuracies, uncertainties, and other pitfalls. In this context, I will juxtapose (a subset of) deep neural network methods with the family of entropy-optimal ML techniques developed recently by Illia Horenko (RPTU Kaiserslautern-Landau) and colleagues.
By Matthew Juniper
By Sergey Gavrilyuk
By David Lannes
