Abstract by Spencer Reschke
An Introduction to Artificial Neural Networks
Artificial Neural Networks (ANNs) are among the most widely used and most successful algorithms used today in the field of machine learning. Because of their ability to model nonlinear functions, neural networks have been successfully implemented to solve problems in a variety of fields including medicine, face/image recognition, and computer vision. Due to the stunning complexity of ANNs, for many people it can be intimidating to begin learning about them. In this talk, I’ll introduce the basics of ANNs－forward propogation, activation functions, cost functions, and back propogation－all in a supervised learning setting. As a potential application of ANNs, I’ll then discuss how ANNs, Gershgorwin’s Circle Theorem, and isospectral reductions could be used to improve estimates of matrix spectra.