In the Loop: Anind Dey

Former college sprinter Anind Dey is teaching his 8-year-old daughter the rudiments of running.

In June 2014, Anind K. Dey was named the Charles M. Geschke Director of the Human-Computer Interaction Institute.

A graduate of Simon Fraser University with a bachelor's degree in computer engineering, Dey also earned master's degrees in aerospace engineering and computer science, as well as a Ph.D. in computer science, from Georgia Tech. Prior to joining the CMU faculty in 2005, he was a senior researcher at Intel Research Berkeley and an adjunct assistant professor in the Electrical Engineering and Computer Science Department at the University of California at Berkeley. He spoke to Link Editor Jason Togyer.

What was your first technology experience?

My brother had a Texas Instruments device that was somewhere between a calculator and a computer. He’s seven years older than me, and was very interested in computing. It was the first device that he taught me how to program.

Did he end up in math or computer science?

He did a master’s degree in mechanical engineering. He was a roboticist for a while and now he’s in the IT industry doing computing.

Was it your intent to become an engineer?

It was. Partway through my undergraduate degree, I visited a company that made six-degree-of-freedom motion simulators for aircraft, and I thought that was what I wanted to do for the rest of my life. I said, “What do I need to do to get a job here?” They said “Leave Canada, go to the U.S., get a Ph.D. in aerospace engineering and then come back.”

Why didn’t you stay in aerospace engineering?

It wasn’t for me. But I was talking to some of my mentors at the time and they said that it seemed like based on my interests, I could do really well in computer science. I stuck it out for one more year to get a master’s degree in aerospace engineering—so now I get to tell my kids that I’m a rocket scientist! Then I switched into computer science, and ended up getting another master’s degree and a Ph.D. in computer science.

Where is the line between computer engineering and computer science?

It’s extremely fuzzy. It used to be that if you were doing software, you were doing computer science, and if you were doing hardware, you were doing computer engineering. Now I have students in ECE who would fit into our department in HCII. To me, the line is so fuzzy as to be non-existent.

Talk to me about the maker movement. What, in your mind, is triggering it?

A couple of things. Technology has become super-cheap, and really accessible. Before, you needed a strong skill set to use some of these technologies. Now, the tools are so supportive, even for novices, that someone can quickly put together software and hardware, and do things that they wouldn’t have been able to do before.

Some people look at Arduino or Raspberry Pi and say, “They’re just toys.” Are they a gateway to bigger science?

I think they are—particularly for young kids. Our kids at home started off just doing some very simple things in Scratch. We tried to keep them away from the technology, and they asked for more. Now, my 8-year-old daughter is working on an Arduino board with my wife and trying to build an MP3 player. It’s so accessible that kids are able to use their creativity to build some really interesting things. They’re not just sitting in front of the computer and playing games any more. I really think this is a great time. There’s this idea that everyone should be able to think in a computational way, and the maker movement is embracing that.

Will we see a payoff in 20 or 30 years?

I think we’ll see a huge payoff. You’ll have a population that has an innate understanding of how to use technology, and how to apply it, that I never grew up with.

Is there still a danger that people will become too dependent on technology—mere consumers rather than creators?

That’s a millennia-old debate. I do have this worry that sometime in the distant future we may reach a time when we’re totally dependent on technology, but I don’t think we’re anywhere close to that. And the human desire to be creative and to build and invent—I don’t see how it would ever go away.

How did you become interested in computer interfaces and HCI?

It was all because of my greatest mentor, my Ph.D. advisor at Georgia Tech, Gregory Abowd. He was a software engineer who had moved into HCI. He was the one who introduced me to it and helped me realize that it was an interesting field and that I could be successful in that field. Working with him on my Ph.D. was very much geared toward how people would interact with novel computing technologies.

We used to expect people to adapt to their computer. Now we expect the computer or interface to adapt to the people. That’s a major mind shift.

It is. I don’t think our technology is there yet. In order to really understand you, or be able to adapt to you, a system has to be able to understand your intentions. Not just what you did, but maybe why you did it, so it can be predictive and provide you the support you need, while at the same time being cognizant that you need to do some things for yourself. I think the idea of a system adapting to you is actually a lot more complex than we make it out to be sometimes.

What are your research interests now, broadly described?

The thing I’ve gotten most excited about recently is the idea that we can learn peoples’ routines over time. There are technologies we can use to understand your behavior, and that’s everything from your smart phone to the physical tracking device you might wear on your wrist. My particular interest now is trying to identify what is routine behavior, and what is not. This might be interesting, for instance, if you have someone who is at risk for cognitive decline. You can monitor routine behavior, and as soon as you start noting anomalous behavior, you might say this is the time when an occupational therapist should come and make a visit, or this person should go see their PCP right now.

Similarly we’re working on some projects where we’re trying to automatically identify good drivers from bad drivers. What if you could understand what normal, good behavior was, and what anomalous, poor behavior was and try to adapt these or make these assessments more personalized? Our goal is not just to perform an assessment, but to actually use the assessment to be a teaching tool later.

How are we protecting people against invasions of privacy or loss of security when we’re collecting all of this data on their habits?

Our take on it, in our research group, has been that the data never leaves the device. If you’re always doing the processing locally to figure out if something is routine or not routine, then it’s just a collection of applications that can use that data, which hopefully are applications that you’ve installed and that you’ve given permission. There are applications that are malicious and can steal that data, and certainly privacy and security are questions that need to be addressed. Fortunately, we have a great group here at CMU that works on those kinds of problems.

What outside interests do you have?

My kids are my main hobby right now. We have two, ages 8 and 10, and I’m teaching my 8-year-old how to run—I used to be a sprinter in college. I also love to cook with my family—it’s one of my favorite activities. We spend a lot of time traveling together. I love to listen to music—although I’m probably the least musical person in the family. And I’m also a huge hockey fan.

For More Information: 

Jason Togyer | 412-268-8721 | jt3y@cs.cmu.edu