553.790 Neural Networks and Feedback Control Systems
Spring 2022
This roundtable course is an introduction to two related areas: neural networks (NNs) and control systems based on the use of feedback. Artificial NNs have many important applications while feedback control is relevant to virtually all natural and human-made systems. NNs are applied in areas such as system modeling and control, classification, function approximation, time-series filtering/prediction/smoothing, and speech/image/signal processing. Topics to be covered for NNs include network architecture (perceptrons, feedforward, recurrent, deep NNs, etc.), learning algorithms (including deep learning), and applications. This course also provides an introduction to feedback control systems, including the role of feedback in regulating systems and in achieving stability in systems. We consider stochastic (noise) effects in feedback systems. We also consider the interface of NNs and control by discussing how NNs are used in building modern control systems in problems where standard methods are infeasible.
Prerequisites: Matrix theory, differential equations, and a graduate course in probability and statistics.