Together with my mentor and CTO of Michigan Aerospace Corporation, Dr. Matthew Lewis, I researched a novel unsupervised clustering method that we intended to use in a Department of Defense project we were working on at the time. The algorithm revolved around quantifying the idea of the algorithms "surprise". We did a substantial amount of work, but ended up having to spend our time on different projects. Source code can be found at my github, though the repository is fairly derelict and illegible at the moment.