The Machine Learning Center at Georgia Tech (ML@GT) is home to many talented students from across campus, representing all six of Georgia Tech’s colleges and the Georgia Tech Research Institute (GTRI).
These students have diverse backgrounds and a wide variety of interests both inside and outside of the classroom. Today, we’d like you to meet Greg Canal, a Ph.D. student who enjoys playing jazz piano and hopes to work in a data science lab after graduation.
Hometown: Middletown, New Jersey
Advisor: Chris Rozell
Current Georgia Tech degree program: Ph.D. in Electrical and Computer Engineering
Other degrees earned: BSE in Electrical and Computer Engineering, minor in Music, from Duke University
Tell us about your research:
My research lies at the intersection of information and coding theory with interactive machine learning. Broadly speaking, coding theory is the study of how to get information from point A to point B as efficiently as possible, by encoding this information in two stages: “source” coding, where we want to compress the information into as few bits as possible, and “channel” coding, where we want to add redundancy to protect this information from corruption by noise and errors.
My goal is to leverage these concepts in designing optimal interactive machine learning systems involving humans; the human has possession of some knowledge (e.g., how to classify animal images into different categories), which is the information at point A, and we want to encode this information through labeled examples to teach a machine learning algorithm at point B. Just as in coding theory, we want to design adaptive learning algorithms that use as few labeled examples as possible (source coding) but also making sure to account for the fact that the human may have inconsistent answers with errors (channel coding).
Favorite Conference and why:
My favorite conference is the International Conference on Machine Learning (ICML), mostly because I presented my work at it this year and was a great opportunity to network in the field and see a variety of interesting work.
Tell us about some of your hobbies:
I really enjoy studying and playing jazz piano. I love that improvisation, playing by ear, and performing with others are all core components of the genre. In college I played in a jazz ensemble, and since then I’ve played in a Georgia Tech jazz combo as well as at jam sessions around Atlanta.
Favorite Georgia Tech experience:
I lived in the Graduate Living Center my first year, where I got lucky with randomly assigned roommates. They have ended up being some of my closest friends during graduate school.
Who is someone that inspires you and why?
Donald Glover, since he is extremely successful in so many different creative areas.
What is your proudest accomplishment?
At this point in time I would have to say getting a paper into ICML this year and presenting it (obligatory paper plug). It was a culmination of ideas and techniques that I spent a lot of time learning and helping to develop.
If you were told you only had one week left to live what would you do?
Get a second opinion.
What are your plans for after graduation?
I’d like to work in a data science research lab somewhere in the industry.
If you could time travel, what period of time would you go to and why?
I would go to a point in the future where I’d be able to travel through space, just because that’d be awesome.
What drew you to machine learning?
I became interested in machine learning through my study of signal processing, which shares many tools with ML. I was drawn to signal processing due to its versatility and power to solve a wide range of interesting problems by using a core set of techniques.
Podcast, movie, tv show, or book? Why?
I enjoy books because I can get completely absorbed in reading.
What is your favorite place to study?
I’m a big fan of Kavarna Coffee Shop. It’s a space where I can both focus on work and relax at the same time.
What are you most looking forward to in the next ten years and why?
I’m looking forward to traveling as much as I can – in particular I’d like to travel to Ireland, where two of my grandparents emigrated from.
What is one thing you wish people understood about machine learning and artificial intelligence?
It’s not a panacea. Before interpreting or acting upon the results of ML/AI systems deployed in the real world, it’s important to understand the limitations of both the algorithm used and the data it was trained on.