Deep learning has achieved great success as evidenced by many challenging applications. However, deep learning developed so far has some inherent limitations. In particular, deep learning is not yet adaptable to different domains and cannot handle small data. In this talk, I will give an overview of how transfer learning can help alleviate these problems. In particular, I will survey some recent progress on integrating deep learning and transfer learning together and show some interesting applications in sentiment analysis, image processing and dialog systems.