Numerical Methods for Algorithmic Systems and Neural Networks

Abstract

These lecture notes are devoted to numerical concepts and solution of algorithmic systems and neural networks. The course is divided into four parts: traditional AI (artificial intelligence), deep learning in neural networks, applications to (and with) differential equations, and project work. Throughout this course an emphasis is on mathematical ingredients from which several are rigorously proven. In the project work, the participants usually form groups and work together on a given problem to train themselves on mathematical modeling, design of algorithms, implementation, and analysis and intepretation of the simulation results.