ALEXANDRIA, Va., July 21 -- United States Patent no. 12,688,427, issued on July 21, was assigned to GDM Holding LLC (Mountain View, Calif.).
"Training action selection neural networks using hindsight modelling" was invented by Arthur Clement Guez (London), Fabio Viola (Montreal, Canada), Theophane Guillaume Weber (London), Lars Buesing (Letchworth Garden City, Great Britain) and Nicolas Manfred Otto Heess (London).
According to the abstract* released by the U.S. Patent & Trademark Office: "A reinforcement learning method and system that selects actions to be performed by a reinforcement learning agent interacting with an environment. A causal model is implemented by a hindsight model neural network and trained using hindsight i.e. using f...