Kalesha Bullard was teaching math as part of Teach for America when a summer research program introduced her to computer science. It changed her career path and eventually brought her to Tech. Now Bullard teaches only one student: Curi, a talking robot. The more it asks, the more it learns.

“I used to focus on how to motivate and engage high school students to help them learn. Now I’m teaching a robot how to be an active learner in order to achieve learning goals. Ultimately, I’d like to put a robot in a classroom — a robotic teaching assistant that learns from the human teacher and uses that information to help instruct and engage students.”

Common sense and situated knowledge — these are what drive people to reach for a drawer when looking for a fork, and reach for the right drawer when in a familiar kitchen. Learning and reasoning about such information is common for people, but not for robots. Sonia Chernova’s research seeks to give machines common sense as well. It’s impossible to pre-program a robot helper with all the knowledge it needs; our world is too diverse, our houses too different. So Chernova is finding ways for robots to learn this information, and to share it with each other.

“What if robots could share information? ‘In my kitchen, we put cutlery in a drawer next to the sink. What about yours?’ If they could pass along knowledge, every robot wouldn’t have to learn from scratch, helping to one day enable robots to perform useful tasks even in new environments. Parents raise children. We need to make sure they don’t have to raise robots too.”

Vivian Chu had no plans to get a Ph.D. when she went to work for IBM in 2009. She didn’t think she was smart enough, but everyone in her research lab insisted otherwise, and now she’s in her fourth year at Tech. Chu teaches robots basic concepts — such as the fact that cups hold liquid and boxes can be lifted — that enable them to complete complex tasks.

“When people find out I work with robots, they always say, ‘You’re so smart. I could never do that.’ That’s one of my pet peeves. ‘NO. NO. NO! YES, YOU COULD!’ I’m not a genius. But I’ve learned that if you work hard at something, you could do this too.”

Brittney English loves playing video games but doesn’t get to do it very often. She’s too busy designing them. Her games, and her custom-built exoskeletons that control them, are for stroke survivors with partial paralysis. The arm and wrist movements are the same ones used during the patients’ tedious, monotonous rehabilitation.

“Most people struggle with motivation while rehabbing at home. Doing the same exercises by playing games is much more enjoyable, and we can look at their scores to see if they’re doing it correctly. It’s one way to return some normalcy and fun to their lives.”

Cars will soon drive themselves on our highways. It will present new challenges for professors such as Karen Feigh. She designs computer algorithms and interfaces that allow autonomous vehicles to adjust to human preferences in ways that make drivers feel both safe and comfortable.

“Future cars will operate much more similarly to today’s commercial airplanes. The human driver won’t turn a steering wheel, change gears, or push brake pedals. Instead they’ll tell the car where to go and it will drive them. Between now and then, we will have cars that automate some functions – like maintaining distance from cars in front of them or maintaining location in a lane. These capabilities could improve safety, but often drivers feel uncomfortable using them because the car stops, accelerates, and keeps distances differently from what the driver is comfortable with."

What does a robot do once it learns something? That’s the question that guides Tesca Fitzgerald’s research. After she moves her robot's arm to teach it a task, the machine records what it sees and uses the information to solve related tasks in the future.

It’s one way Fitzgerald combines her two fields of study: robotics and human cognition.

“Robots are very good at doing specific things, but why should they use their brains to only do those tasks? People learn throughout our entire lives and build upon those experiences. Robots should do the same thing. They need to be dynamic — adjusting to problems and situations they’ve never seen before.”

Ayanna Howard has overseen nearly 50 projects in a dozen years at Tech. Her machines have taught children with autism to play Angry Birds and climbed through sand to replicate Martian exploration. Her next initiative will give personalities to telepresence robots. As one of the nation’s most recognized female roboticists, Howard is always searching for the next project with the potential to make a significant social impact.

“I’m always looking for a problem in need of a solution, then seeing if a robot can provide it. That often brings me to health care. It’s tough, because you have to prove that your solution isn’t just a fun gadget before it can be adopted. What’s success to me? When our robots are used by patients, then I know we have a great project."

Karen Liu isn’t afraid to watch her expensive machines topple on their heads when they’re pushed by her graduate students. Before robots can quickly walk or run in uncontrolled environments, they must first learn to fall. A lot.

“Even with the hardware constraints, there’s a lot of things we can do at the algorithmic level to improve the agility of robots. The most exciting part of my job is the freedom to continuously learn new things with my students. I see myself as a student trying to make sense in this mostly uncharted domain. No one is an expert. We’re all learning together.”