Georgia Institute of Technology
My research focuses on Learning from Demonstration (LfD) algorithms which provide intuitive ways to empower end-users to teach robots new skills and specify what they want the robot to help with. I currently have two research directions: personalized learning from demonstration and safe learning from demonstration. In personalized LfD work, I aim to provide a personalized service to each user while not requiring a single user to provide massive demonstration data for the machine learning algorithm by sharing task commonalities across users. In safe LfD work, I aim to allow users to specify safe and dangerous configurations of the robot such that the robot can carry out the learning process and the final execution of the learned behavior safely with human partners.