Deep reinforcement learning

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Deep reinforcement learning (DRL) uses deep learning and reinforcement learning principles in order to create efficient algorithms that can be applied on areas like robotics, video games, finance, healthcare.[1] Implementing deep learning architecture (deep neural networks or etc) with reinforcement learning algorithm (Q-learning, actor critic or etc), a powerful model (DRL) can be created that is capable to scale to problems that were previously unsolvable.[2] That is because DRL usually uses raw sensor or image signals as input as can be seen in DQN for ATARI games[3], and can receive the benefit of end-to-end reinforcement learning as well as that of convolutional neural network.


  1. ^ Francois-Lavet, Vincent; Henderson, Peter; Islam, Riashat; Bellemare, Marc G.; Pineau, Joelle (2018). "An Introduction to Deep Reinforcement Learning". Foundations and Trends in Machine Learning. 11 (3–4): 219–354. doi:10.1561/2200000071. ISSN 1935-8237.
  2. ^ Arulkumaran, K.; Deisenroth, M. P.; Brundage, M.; Bharath, A. A. (November 2017). "Deep Reinforcement Learning: A Brief Survey". IEEE Signal Processing Magazine. 34 (6): 26–38. doi:10.1109/MSP.2017.2743240. ISSN 1053-5888.
  3. ^ Mnih, Volodymyr; et al. (December 2013). Playing Atari with Deep Reinforcement Learning (PDF). NIPS Deep Learning Workshop 2013.