AIB1: AI/ML Bootcamp

Presentation Abstract
This bootcamp will present the basics of AI/machine learning (ML) and their applications to microwave engineering. It is intended for engineers who want to learn the basics of AI/ML or are interested in using AI/ML for microwave applications, marketing and sales professionals who are interested in understanding the basics and relevance of AI/ML for microwaves, professionals with AI/ML expertise seeking to explore potential applications to MHz-to-THz technologies, and university students who like to acquire the basic knowledge of AI/ML. To this end, the bootcamp includes introductory presentations on the fundamentals of AI/ML, covering supervised, unsupervised and reinforcement learning. Moreover, we will introduce common types of neural networks such as fully connected artificial neural networks (ANNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), long-short term memory networks (LSTMs), generative adversarial networks (GANs). We will discuss their function, training cost, relative advantages and limitations, as well as their suitability to various applications. We will also introduce concepts such as generalizability (i.e. the accuracy of neural networks to cases outside their training set) and overfitting (when the network learns training data well, but fails to generalize to new cases). Examples of applications of AI/ML to microwave engineering to be presented include: electromagnetic modeling and optimization, microwave filter modeling/design, GaN HEMT modeling, Doppler radar based human motion recognition, gesture recognition and object identification, radio coverage prediction and design optimization of reconfigurable intelligent surfaces. The course will provide ample opportunities for audience interaction and Q&A.