Reinforcement Learning Improvements

Very interesting item.   Quite technical.  I like the hint of using pools or teams of agents to solve reinforcement problems.   Brings to mind the idea of process design and bringing multiple resources to bear.    Can it be more directly linked to process optimization processes?  Checking it out.

Google Brain and DeepMind researchers attack reinforcement learning efficiency   In VenturebeatBy Kyle Wiggers

Reinforcement learning, which spurs AI to complete goals using rewards or punishments, is a form of training that’s led to gains in robotics, speech synthesis, and more. Unfortunately, it’s data-intensive, which motivated research teams — one from Google Brain (one of Google’s AI research divisions) and the other from Alphabet’s DeepMind — to prototype more efficient means of executing it. In a pair of preprint papers, the researchers propose  (technical paper) Adaptive Behavior Policy Sharing (ABPS)  , an algorithm that allows the sharing of experience adaptively selected from a pool of AI agents, and a framework — Universal Value Function Approximators (UVFA) — that simultaneously learns directed exploration policies with the same AI, with different trade-offs between exploration and exploitation.  .... "