Understanding Deep Learning and Neural Semantics

I_01_Xiaogang Wang(Photo)
Xiaogang Wang
Chinese University of Hong Kong

Abstract :

Deep learning has achieved great success in computer vision. Many people believe that the success is due to employing a huge number of parameters to fit big training data. In this talk, I will show that neuron responses of deep models have clear semantic interpretation, which is supported by our research on multiple fields of face recognition, object tracking, human pose estimation, and crowd video analysis. In particular, the responses of neurons in the top layers have sparseness and strong selectiveness object classes, attributes and identities. Sparseness and selectiveness are strongly correlated. Such selectiveness is naturally obtained through large scale training without adding extra regularization during the training process. By understanding neural semantics, we are inspired to develop new network architectures and training strategies and they effectively improve a broad range of applications in face recognition, face detection, compressing neural networks, object tracking, learned structured feature representation in human pose estimation, and effectively learning dynamic feature representations of different semantic units in video understanding.


Professional Activities

  • 2015–present, Department of Electronic Engineering, the Chinese University of Hong Kong, Hong Kong, Associate Professor
  • 2014–present, NIVIDIA CUDA Research Center at CUHK, Director
  • 2014–present, Institute of Space and Earth Information Science(ISEIS), Executive Council Member
  • 2014 Organizer of ACCV International Workshop of Deep Learning on Visual Data 2014
  • Associate Editor of IEEE Transactions on Circuits and Systems for Video Technology
  • Associate Editor of Image and Visual Computing Journal
  • 2014 Area chair of European Conference on Computer Vision
  • 2014 Area chair of Asian Conference on Computer Vision
  • 2011–2012, Consultant for Hong Kong Applied Science and Technology Research Institute Company Limited (ASTRI) on Environment Adaptive Object Detection and Tracking
  • 2011 Area chair of IEEE International Conference on Computer Vision
  • 2011 Program committee member for the first IEEE workshop on modeling, simulation, and visual analysis of large crowds
  • 2011 Program committee member for the third Chinese Conference on Intelligent Visual Surveillance
  • 2011 Program committee member for the Workshop on Behavior Informatics
  • 2009–2015, Department of Electronic Engineering, the Chinese University of Hong Kong, Hong Kong, Assistant Professor


Awards & Honors

  • IEEE Member
  • 2016 PAMI Young Researcher Award 2016 Honorable Mention presented as CVPR
  • 2012 Hong Kong RGC Early Career Award
  • 2012 The Chinese University of Hong Kong Young Researcher Award
  • 2011 Outstanding Young Researcher in Automatic Human Behaviour Analysis award



  • 2009 Ph.D. Candidate in Computer Science, Massachusetts Institute of Technology
  • 2004 M.Phil. in Information Engineering, The Chinese University of Hong Kong
  • 2001 B.S. in Electrical Engineering and Information Science, University of Science and Technology of China