MEET BOYUAN CHEN, ROBOTICS RESEARCHER

I HAVE LEARNED TO PLAY THE PIANO for more than 20 years. This process to me is nothing too different than building robots. You’re trying to understand the fundamental principles of how music works. In this case, we’re trying to understand how engineering principles work. And you use this knowledge to compose and make novel form of robots or a novel form of music.

Bouyan Chen, I’m an Assistant Professor at Duke University, focusing on robotics and AI research. We’re at the General Robotics Lab at Duke University, where I'm also the director of the lab.

When I look at robots, I don’t see them as a mechanical machine. I actually see them as almost a new species that are also evolving on both their embodiment form of structures and also how they think and how they behave.

What we are doing now is trying to lay out a foundation, how we’re going to evolve this new species. Throughout that trajectory, I’ve been exploring many aspects of engineering. Both the biology aspects of intelligent machines learning from biochemistry, biophysics, neuroscience, and also the engineering aspects of it, such as the building machines, and constructing mechanical designs, electrical circuits, and programming.

And lots of challenges that we were not able to tackle before are now possible with the third wave of AI growth, where we started bringing AI into a cognitive level capabilities by leveraging lots of data, advancements of computing, and new neural network architectures.

I think this is perhaps the greatest innovation that we have so far in the entire humanity. There was a moment around 2015, 2016 where, you actually see huge development of AI, started getting to robotics world. And this is the moment where I see myself to really pick up this new wave of development AI, but, put a body with it.

And I think a lot of development back then focused on virtual, a digital data like images, videos, or text languages. Not so much discussion about how to make use of the physical embodiment and very little discussion on how the embodiment form of robots will inform the learning process.

And in fact, I think this is the fundamental aspects of robotics the embodiments. I think body in some way shapes how we learn, and also shapes how we think about what we can and cannot do. I think what will be really essential for future engineers to be successful is this full stack mindset.

I think being able to understand how mechanical structures, sensors work from the hardware perspective, all the way to how you can understand the software, the control algorithms, the planning algorithms, the learning algorithms can be very helpful. So these co-designing or co-thinking from body mind interactions and structures will make you a very unique assets in the robotics community. And I think we’re at a turning point where there is more and more demands from industry, academic, research labs, institutions, all demand talents who can understand both sides of things.