My latest work under Koyote Science, LLC, can be found under Deep Learning, where I outline an approach to resolving an unsolved problem in music informatics — how can we identify chords from an audio stream? It turns out the problems yields nicely to the latest ML techniques, although there’s still more work to go! The dream is to augment other musicology efforts like HookTheory.com and other work presented on this website.
In the Document Surfer, I demonstrate an innovative way to surf unlinked documents just like we surf the web. You can try it out on email publicly released from Enron (Enron Email Surfer). It ties into the Creativity Engine, a research wiki I designed to help navigate an organically growing set of research pages and forge useful links based on data-mining techniques.
In Musicology, I present the principles behind a MusicWiki, which permits community input of musical data, analytics, and visualizations. This work has recently earned me a quotation in The Boston Globe.
l’ve recently join the Data Science, Advanced Analytics team at Twitter. This will be a great opportunity to develop the ideas from the Email Surfer to solve real-world problems that affect hundreds of millions of people. Feel free to check back in for updates!
My professional activities can be followed in Professional Summary, and you can reach me at douglas [d0t] mason [ @T ] gmail [d0t] com. You can also follow my antics using my Twitter handle @CaliCoyote.