Machine Learning

The idea is that robots will be able to learn how to recognize, classify, and even behave in certain situations by interacting with people like a child does: by watching and listening to what people do and say, and by trying to imitate them. My research focuses on a small part of this bigger picture: on the design of learning algorithms that help robots learn from people. Implementing a social robot requires designing a robot that can recognize and classify people, and then interact with them in a meaningful way. To do this, we need to develop learning algorithms for robots that can help them recognize and classify people, and then learn how to interact with them. The problem of learning from interaction is a hard problem. It is difficult because interactions are complex and dynamic. People are complex Robots need to interact with people, but people are complex. People have a rich set of physical, verbal, and non-verbal behaviors. People have a rich set of goals. People have different personalities. People ar