I’m pleased to bring GoRobotics an exclusive interview with Angelica Lim of Kyoto University. When I first started writing here at GoRobotics, one of my goals that I stated was to bring more compelling academic research to the general public and enthusiasts because behind lots of jargon and hidden in some grad student’s lab somewhere is a robot waiting for it’s chance in the spotlight.

Let’s get right into things with Angelica.

How did you end up a roboticist? Was it a childhood dream?

I had no idea I wanted to be a roboticist when I was a kid. It started when I was on exchange in France, doing a year of Computer Science classes at the University of Nice. One of our projects was to pick amongst research topics proposed by faculty members, and “Build a Data Server for an Autonomous Underwater Vehicle (AUV)” was one of them. I ended up choosing that on a whim, and our team did a pretty good job coding it up in C++ under her specs. I got called back the next year to help integrate it with a real “live” AUV for a competition in England, and I was hooked. I liked it so much that I put together the robotics team back home in Canada. That was my second robotics competition – hopefully not my last!

How did you end up in Japan working on robots?
The main reason I wanted to come to Japan was simply because the hardware is much more advanced and easy to acquire. Full-size humanoid research platforms have been out in Japan for almost a decade. Only now are companies like Willow Garage starting to gain traction in North America.

On a more personal level, I also felt like my research options would be limited in North America. In the US, robotics research is heavily funded by the military, and therefore it seemed to me, at least that my research would have to conform to very serious and grave goals in order to gain funding. In Japan, robotics applications sound less like “Big Dog” and more like “RIBA Nurse Robot” and “Fan Dancing Robots” . I prefer the Japanese outlook on a future with robots. Does that make sense?

More after the jump …

It does. It sounds like you want the application of robots to be a wholesome augmentation of society rather than something that has the potential to destroy our planet any further than we have already.

To more specifically answer your question as to how I managed to get to Japan – I found a scholarship that sends Canadians to Japan to do research, and boom! Here I am.

What lab are you with and what kind of research does it work on?
I’m with Kyoto University’s Speech and Media Processing lab. Nowhere in the title does it say “robots”, but I can assure you that over half the group works with them in some way. One sub-team works on Human-Robot Interaction, doing research such as robot learning using neural networks. Another does Speech Processing and Robot Audition: giving robots speech comprehension and production abilities. The third group does music research, such as taking a CD recording, extracting parts and synthesizing new instruments. My group is a fusion of the music team and the robot team. We do research on music robots.

What project are you on specifically?
We’re developing a theremin-playing robot that has adaptive abilities. It can play with humans and understand gestures to play on command, and that is the project I am currently hands-on with. The robot can also listen to another player (for example, a drummer) and adapt its playing to their rhythm. It is a music robot that can watch and listen to the people it’s playing with. This, I have to say, is an ability that many human musicians still have yet to learn.

If you can get a robot doing that before the kids in band class, I’m going to be impressed.

What is your favorite thing about your current project?

I’m a flute player, and I have a pile of duet music that I’ve been dying to try out with someone. My favorite thing about my current project is that late at night I can head into the robot lab and jam with the robot.

What conferences have you presented your work on this project to, and how was the reception?
I presented my gesture recognition work last month in Tokyo at a Japan-wide conference on technology. The attendees seemed to really like the idea of using gesture recognition to guide the music robot. We don’t often realize the unconscious physical gestures that we use, not just in music, but in every day life. One attendee noted that it’s not just robotics or music research we touch on here, but a more fundamental, ethnographic question – how do humans communicate? To give robots this gesture recognition capability, we must first understand what exactly what we’re doing ourselves.

How do you think what you’re working on will become useful to the world? What do you hope it will be used for?
At conferences, our work typically falls under the category of “entertainment robots”. At first glance, it seems like our work is somehow less impressive for that reason, as intuitively this does not seem immediately useful to most people. Why would we want robots to entertain us? Wouldn’t you rather have a human play music instead of a robot? There’s no way they could play as well, etc. Well, I’d like to think on a longer scale and about the bigger picture as this field develops.

In Japan, the idea of having “robot companions” is very exciting, especially for the elderly population. There just aren’t enough young volunteers or workers to care for them on more than a just basic level. The brain needs exercise and stimulating interaction, particularly for those recovering from neurological impairment or deterioration, such as in the case of a stroke. A robot that can play chess, perform Bach, or do magic tricks wouldn’t just be an “entertainment robot”, it could actually be a tireless therapeutic device.

If you had to pick other research than your own lab, what research or robot out there do you think is really cool? What do you think Go Robotics should be keeping an eye on?
I’m really interested in the programming by demonstration research paradigm.

For example, Aude Billard’s group out at Ecole Polytechnique Federale de Lausanne in Switzerland has made a robot that learns how to cook by imitation as well as a robot that can play ping pong.

The start-up Aldebaran Robotics in France is also taking leaps forward with their miniature robot system named Nao. Eventually, once they work out the bugs on the small-scale version, they’re planning on developing a human-sized version with the same functionality!

What do you think is the most common misconception people have about robots?
“Robots will take over the world. Stop giving the robots intelligence; they’ll use it against us!”

I think that people are fooled into thinking how intelligent robots are based on on the great YouTube videos that come out every other week. The reality is, it looks real, and we are trying very hard to have robots pass the Turing Test (ie. when we can’t tell the difference between a robot and a human.) But, looking and seeming intelligent and actually being intelligent are very different. Even more compelling is looking and seeming conscious compared to the idea of being conscious. My tip is: don’t be fooled! They are about as conscious as your car or microwave.

Thanks again to Angelica for taking the time from her busy schedule to talk to us about her world of robots and especially for giving us some compelling developments to watch going forward. Pass this interview along to your friends, because you’re not going to find it anywhere else.