Stocco

AAAI Conferences 

This paper proposes a reinterpretation of selective attention as a form of control of working memory based on self-generated reward signals and model-free reinforcement learning.