Self-driving cars see the road, its markings, and pedestrians crossing it to steer themselves safely; Siri uses complex audio processing and speech recognition algorithms to pick up phrases and even whole sentences; music apps like Pandora and Spotify learn your tastes in music to the point where they can suggest new music that you're likely to enjoy. They are not magic. The technologies mentioned above all utilize capabilities that humans instinctively possess: the ability to make intelligent judgments based on our perceptual senses. Recently, machines have become much better at making these same sort of perceptual judgments, and many times their design is either inspired by or coincidentally shares common elements with how humans accomplish the same thing using our senses and our brains. In this class, we explore how our senses are suited to solving perceptual problems, how we design machines to solve very similar problems, and how the algorithms for the two compare. Our study will span the engineering disciplines of computer vision, speech recognition, signal processing, and artificial intelligence; and the scientific fields of biology and physics. The goal is to look at all the difficult perceptual problems both nature and engineers have encountered (vision, hearing, learning), look at how they've tried to solve them, and ultimately to try and make some "sense" out of them. A word of advice: do not let any of the big words I may have used deter you from signing up, though at the same time, do be aware that this class will have a lot of math because that is the tool we use to understand complex things we do not yet understand.

We may take our favorite sense for granted, but computer scientists have struggled with the problem of replicating the power of human vision for decades. While nature, over the course of hundreds of million years, has evolved several good designs for vision in animals, only recently have the smart engineers at places like Google and Vecna been able to start cracking the problem of efficient and reliable machine vision. Sign up to learn about the process that involves 60% of the human brain and how we implement it in machines, because when it comes to vision, there's more than meets the eye.

Parsers are incredibly handy -- from writing interpreters for a toy programming language to doing natural language processing, learning how to write parsers well is a useful skill for any application. They are also interesting to study, because they give insight into the structure of languages as well as the nature of computation itself. In this course, we will use Haskell to study and learn how to write parsers, with hands-on real-world examples. In addiiton, we will learn about the relevant areas of linguistics and theoretical computer science. No knowledge of Haskell beforehand is necessarily, as we will also be learning the principles of functional programming along the way.

Interested in the theoretical foundations of computing? Want to learn about what makes a computer a computer? Wondering what the heck functional programming is, anyway? This class covers everything from the Church-Turing thesis to the fix-point combinator, with some hands-on Haskell programming as a motivator. This is a primarily a math class aimed at understanding how computers work at the theoretical level.

So you know how to code, and you've always wanted to team up with a few others to build something incredible, but you've never gotten the chance to do a hackathon before. This class is an opportunity to do just that. We begin with a few tips for hackathon success (including how to host your own hackathon, which can be just as fun), and cover some useful API's. Later, we split into teams and try to hack our own projects, and we finish by showing off our creations to the class.