Deep learning has transformed the way in which we design machine learning systems. In this talk a brief overview of modern advances to the deep learning paradigm was provided. Starting off with deep reinforcement learning: DQN (Nature, 2015) and A3C (CoRR, 2015), motivate the role of generative modelling in the emerging research landscape and discuss several recent models, including Pixel CNNs (ICML, 2016), AIR (NIPS, 2016) and Conditional 2D->3D (NIPS, 2016).

Video : http://www.aparat.com/utacm

 Slides: http://acm.ut.ac.ir/acm/uploads/post_attachment/attachment/87/Intelligent_Perception.pd