Abstract: This talk will be an introduction to Artificial Intelligence through an interactive exploration of one of the most important algorithmic models driving current AI progress: neural networks. Popular tech like Chat-GPT and Midjourney show the remarkable ability of AI to produce surprisingly complex responses to prompts by human users, and generate images and video that look almost as if they were actually captured on film or created by an artist. The "intelligence" of these Artificial Intelligence systems comes from statistical patterns and mathematical algorithms, series of rules that tell a computer how to transform input of some kind into an appropriate output. Neural networks are one implementation of these rules that is especially flexible and scalable, which is a major reason their use has risen alongside data production and collection and remain so relevant today.

Neural networks are surprisingly intuitive, and can be seen as a jumping-off point into the study of AI more broadly. With little background in topics like programming, linear algebra, or differential calculus, you learn how these models work through visual representations and manual manipulation of simple networks. This talk will focus on how neural networks learn patterns from data and generate predictions and new data based on what they have learned. Given the growing number of people interacting with AI systems and their products, it is useful to understand how these systems work for non-experts too. Even a basic understanding of the logic of these AI systems can help to shape expectations of new technologies, and allow more people to follow and contribute to conversations about how they are being used.

Bio: Patrick grew up in Penn Hills, outside Pittsburgh, and came back to the city after graduate school. He received his bachelors' degrees in Cognitive Science and Computer Science from Johns Hopkins University in 2017, and a Ph.D in Computer Science from Lehigh University in 2022, advised by Eric P.S. Baumer. His research interests lie broadly in the field of internet privacy, and responsible use of AI and data. He currently teaches courses in machine learning and computer programming.