NeurosciencePLoS Computational BiologyPeer-reviewed
Recurrent neural networks that learn multi-step visual routines with reinforcement learning
Sami Mollard, Catherine Wacongne, Sander M. Bohté, Pieter R. Roelfsema
This paper trains a recurrent neural network with reinforcement learning to solve multi-step visual tasks, showing it can learn elemental operations and propagate enhanced activity between subroutines.
6 citations