The Machines...They're Getting Smarter

Graduate Student Mentors: Daniel Cicala and Brandon Coya

Suppose you wanted to draft a championship baseball team on a budget, but have no idea how. You could get a larger budget by buying the next hot stock. Or you could do both. “But I don’t know about home runs or the stock market”, you think.

Suppose you had the most exciting job in the world: an international spy! Your mission: pick out a potential villain from a massive crowd. Or maybe your job was a bit less exciting but crucial for a modern functioning society: sorting mail. How are you going to separate all of these letters into their zip codes?

Grab your laptop and head down to your local cafe, because it will take care of everything once we teach it how to learn for itself. All of these problems are within the purview of machine learning. One major success of machine learning is the development of the computer program AlphaGo. It made headlines when it defeated the world’s best players of the game “Go”. This was never accomplished with programming techniques because there are too many possibilities to account for.

We will learn basics of the machine learning paradigm known as neural networks and the math behind the scenes. By the end of this project, we will design a computer program that can recognize digits written by a human and understand the underlying mathematical principals. No prior coding experience needed.

References:

[1] Neilson, Michael. http://neuralnetworksanddeeplearning.com/

[2] Ian Goodfellow, Yoshua Bengio, Aaron Courville. Deep Learning.