Active Predicting Coding: Brain-Inspired Reinforcement Learning for Sparse Reward Robotic Control Problems.

Alexander Ororbia Ankur Arjun Mali

Published in: CoRR (2022)

Keyphrases

control problems
reinforcement learning
state space
coding scheme
real robot
real time
function approximation
optimal control

queueing systems
learning algorithm
temporal difference
model free
reinforcement learning algorithms
human brain
adaptive control
eligibility traces

optimal policy
machine learning
learning problems
dynamic programming
robotic systems
reward function
state action
multi agent

approximate dynamic programming
action selection
partially observable
action space
reinforcement learning methods
continuous state spaces
stochastic control