Research Projects
My research interests include computer vision, machine learning, and computer games. Currently, I am working on the following projects:

Knowledge-driven Few-shot Image Object Recogntion
The past decade have witnessed the progress of object recognition with the help of deep learning (such as CNN, Transformer). However, existing mehtods rely on huge costly labeled data which are difficult to be acquired. In many cases, data are rare, where existing deep-learning-based methods often fail. This project aims to developing new theories, methods and technologies for image-based object recognition under the few-shot setting. Our key idea is to utilize prior knowledge learned from mass auxiliary data to guide the training of neural network under few-shot settings.
Research topics: (1)Knowledge Graph, (2)Pre-trained Models and Tuning, (3)Model Generalization, etc.

Large scale topic model learning for social media data analysis
Our goal is to develop new machine learning tools for large scale social media data analysis. This project is supported by a NSFC-MSRA joint fund.