While AI is changing our world and powering a novel industrial revolution, robots are expected to play an increasingly important role in our daily life. The overall goal of this research theme is to endow robots with human-like intelligence, e.g., learning and vision capabilities, so that robots can visually understand the real world and learn from their past experiences to adapt rapidly and autonomously to unknown novel environments. Applications of such AI-empowered autonomous and adaptable robots are uncountable: robotised picking and kitting in logistics, robot assistants for elderly and patients in hospitals, etc.  

Specifically, our current research themes are

• Transfer learning and domain adaptation
• Instance segmentation and object pose estimation
• Few-shot learning, meta-learning, incremental learning
• Data simulation for effective learning of robotised grasping