My current research interests mainly focus on deep representation learning, weakly/semi-supervised learning, transfer learning and deep structured prediction, with their applications to vision and robotics problems. I am broadly interested in: (1) general scene and video understanding, (2) bottom-up grouping and mid-level representation, (3) robust representation for cross-domain/cross-task/open set generalization and adaptation, (4) lifelong learning with interactive/weak/self-supervisions. My research goal is to is to leverage the abundant domain knowledge, priors and structured information to eliminate uncertainties with minimum human supervision and uncover the secret towards true visual intelligence.