README.md

Attentive Neural Network Architecture

Note: not quite finished yet

ANNA is an attempt to make neural reinforcement learning efficient and scalable. ANNA is unique in that doesn't use backpropagation, and is instead based on a probabilistic graphical model.