Media coverage of robotics tends to be a bit one-sided, and we've been flooded with images of robots that perform effortlessly because they've been designed and trained for a specific task: cleaning a floor, kicking a basketball, even running like a dog. The idea that a brand new, cutting-edge robot would have trouble picking up an everyday object sitting on a table is strange to us, and Joe Palca of National Public Radio's All Tech Considered talks to Brown CS Assistant Professor Stefanie Tellex to find out why and learn about her team's crowdsourcing experiment designed to change this learning curve.
The audio interview, a text article, and multiple videos of robots in action are available here. You can also watch PhD student John Oberlin and Stefanie's "Bandit-based Adaptation for Robotic Grasping" video by clicking the image above and read their paper here.
For more information, please click the link that follows to contact Brown CS Communications Outreach Specialist Jesse C. Polhemus.