Jake Peterson

Hello all, my name is Jake Peterson. I am currently majoring in Computer Science with a minor in Mathematics here at DePaul University. My research areas are in computer vision, machine learning, and data science. Outside of research, I am the president of the Computer Science Society at DePaul, an avid painter, and an experienced barista. After graduation, I plan to work as a software engineer while preparing for graduate school applications. My academic goals include, but are not limited to, a Juris Doctor degree and a graduate degree in computer science. 

BS in Computer Science

Jacob Furst

jake@peterson.com

Project: Hybrid End-to-End Convolutional K-Nearest Neighbors Classifier

Our goal is to create a novel end-to-end machine learning model that replaces the fully-connected layers of a Convolutional Neural Network (CNN) with a differentiable K-Nearest Neighbor (KNN) classifier. Due to the large number of weights that need to be learned in a fully connected layer, fully connected neural networks are computationally expensive to train, require a large amount of data to train accurately, and are difficult to interpret. By replacing the fully connected layers of a traditional CNN with a differentiable KNN, our hybrid Convolutional K-Nearest Neighbors (C-KNN) classifier overcomes the limitations of fully connected layers without a reduction in classification accuracy.