This New Atari

Date: June 15, 2017 Published by

  • Not long after that Atari work came out,
  • Vicarious’ Schema Networks, on the other hand, think more like humans do—at least according to Phoenix.
  • “You observe the entire scene in Atari games.
  • In each scenario, Schema Networks outplayed the deep reinforcement learning networks’ highest scores.
  • The problem with deep reinforcement learning networks, he says, is they are essentially trial and error.