I have been interested in genetic algorithms, neural networks, and artificial intelligence/machine learning since I first heard about them in the 1980s. Although they have been around for a long time, they have had limited applicability for a number of reasons -- neural networks are hard to train and are very sensitive to input variation, genetic algorithms require very clear criteria for measuring success, nearly everything relies on the implementer having a good sense of the problem domain, etc. Recently there has been renewed interest and huge advances in machine learning—it is practically a household word today thanks to the talents of our new friends Siri, Alexa and Cortana (to name a few). New frameworks and approaches have made the methods and algorithms more accessible than ever!
This talk will present a general overview of some of the concepts and challenges in this space, and show the results in progress of my own attempts at implementing a neural network and using a genetic approach to evolve its skill.