In our Uppsala social robotics lab, we aim to answer a fundamental question: How could robots learn like humans ? We achieve this by researching and developing new robot learning models using machine learning theories. The key of our research is to understand what interactive elements make the learning process different and how could we build efficient learning models for human robot interaction. Our research is multi-disciplinary and it requires knowledge of different fields. Except machine leanring algorithms, we also investigate on computational psychology models. We conduct both fundamental and applied research together with other collabrators around the world.