Faculty Spotlight: Kevin Gold
We are welcoming our initial cohort of undergraduate students transferring into the BS program in Data Science from other BU units this fall. With many data science courses on the books for Fall 2021, we sat down with newly hired Associate Professor of the Practice Kevin Gold to learn more about his background, interests, and passions. Please join us in welcoming Kevin to the BU community!
Q: What were you doing before joining BU?
Kevin Gold: I've been at Northeastern for six years as a teaching professor, with AI and algorithms my favorite classes to teach. Long ago, around 2008 when I got my PhD in computer science, I focused on real-time robotics and real-time AI for games, which are similar except that a robot has much more uncertainty about what is going on. But when I went into industry, I did whatever they needed related to AI, which was on a bigger scale than anything I had done in academia, and was interesting precisely because of that scale. I worked a bit with MIT Lincoln Laboratory on humanlike Internet traffic patterns, with YouTube on finding ISPs that didn't deliver on their speed promises, and with Google's search team on some details involving question answering. I tend to find a lot of machine learning or AI-related technologies interesting, especially when they're deployed in the wild - like at Google.
Q: What motivated you to pursue computing and data science?
A: I wandered around majors quite a bit in college, and majored in Cognitive Neuroscience and Math before Computer Science (CS). Three things struck me about CS: one, it was about making things in a way my other classes were not. Two, the classes I had taken up to that point worked fine as CS electives, remarkably, and that was not only convenient but spoke to the discipline's versatility. And three, my time as a young elementary school student trying to copy programs from books actually paid off, because I was still left with a knack for programming, even if my early exposure was both superficial and full of frustrated tears. But it wasn't until I met a passionate AI professor in grad school that I chose to go deep into AI; throughout undergrad, CS was just fun, and AI just one more way to make stuff.
I don't see my shift to data science as being too much of a shift, because I've always studied machine learning from graduate school onward, and I've always been interested in applying it to lots of different things. The emphasis on big data means you can't create "microworlds" where the AI only works because the environment is artificial.
Q: Can you talk about your teaching philosophy? What excites you about BU now offering a BS in Data Science?
A: I've had a longstanding hobby interest in writing and in games, and a lot of my mentality comes from advice I read for one or the other of those:
Start with a hook. Focus attention on the most important stuff. Show the payoff down the road. I also get a lot of mileage out of the advice in Chip and Dan Heath's book Made to Stick, as I'm constantly trying to follow their advice to make lectures more well-structured, unexpected, concrete, and credible.
The existence of a new program also means a chance to rethink what is important in 2021, and that means a shift toward more cutting-edge technology - and that is exciting.
Q: How do you approach teaching data science, compared to what you’ve taught before?
A: Data science feels more about coding and visualizing data - it feels more hands-on, less about theory (although there's still plenty of theory). I'm doing many more lectures that involve coding in some way as a result of the change.
Another challenge is that while data science can solve real-world problems, the real world is messy, and education wants neat examples for pedagogy. I feel that tension more with this discipline.
Q: You’ve designed three choose-your-own-adventure web games. When did you start making games and why?
A: This was the result of knowing one head of the game-making company Choice of Games through an annual games convention, AnonyCon. He liked the tabletop roleplaying games I had run for fun once a year every December, and thought I should try writing for his new company. I worked on those games every weekday for an hour every day, a thousand words a session, doing it little-by-little so as not to interfere with my day job (which was MITLL, and then Google). When Choice of Robots was finally done, they were surprised I was still working on it - it was years later. It has rather little to do with teaching or data science, but I like telling my students that you don't have to be just one thing. You can be multiple things at one time, or you can be multiple things over time. Lately, I've had no time for writing those games because of my two-year-old, but I hope to again some day.
Q: What sorts of things do you look forward to doing at the Faculty of Computing & Data Sciences?
A: I hope we can create a community where we really get to know our students and each other, and create a really wonderful learning environment as a result.