Modern Artificial Intelligence for the Curious Layman
There is a quiet revolution happening in Machine Learning right now that is defying thousands of years of mathematical tradition, the gist of which can be understood without any background in ML at all. At the core of this revolution is a fundamental shift from representing mathematical relationships with equations to representing these relationships as structures. And this new approach is quickly proving its mettle; from IBM’s Watson to Google’s self-driving cars to Facebook’s facial recognition, the major players in Machine Learning are gaining a lot of value out of this new structural approach to AI.
In this session, I will provide a high-level introduction to the fundamental concepts behind Equational and Structural Mathematics, highlighting their relative strengths and weaknesses through examples that require no mathematical background to understand. Later on in the presentation, I’ll dive into an explanation of the inner workings of my own implementation of one of the most popular Structural algorithms – the unbounded Neural Network. At the end of the session, you will walk away with a high-level understanding of one of the most important and exciting breakthroughs in Artificial Intelligence since the field’s inception, as well as a deeper insight into just what Google, Facebook, IBM, and others are up to.
Patrick Abbs is a programmer with a passion for Machine Learning that predates his first ever “Hello World”. He began his career in the trial-by-fire world of NYC startups building Natural Language Processing tools for document categorization and recognition, before moving down to Austin for a more stable life as a Consultant for Headspring building custom applications for a variety of clients and domains. At work Patrick codes smart, sustainable websites, while at home he experiments with cutting-edge Machine Learning algorithms in what is definitely not an attempt to build Skynet.
- Austin .NET UG 52 Recordings
- AI 1 Recording