[month] [year]

Prof. Andrew Zisserman

Prof. Andrew Zisserman, one of the principal architects of modern computer vision presented a KCIS Distinguished Lecture on Learning from Sight and Sound on 2 January.

Prof. Zisserman described self-supervised learning from videos with sound. He divided his talk into two parts, in the first part he  described self-supervised learning from the visual stream alone, and showed the possibility of learning powerful embeddings for tasks such as facial attribute prediction and human action recognition. In the second part he explored multi-modal self-supervised learning from video and audio, and then investigated two proxy loss functions, synchronization and correspondence, to link the modalities.

Andrew Zisserman is one of the principal architects of modern computer vision. He is best known for his leading role during the 1990s in establishing the computational theory of multiple view reconstruction and the development of practical algorithms that are widely in use today. He is the only person to have been awarded the Marr Prize three times.

 

  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •