Backprop-Free Reinforcement Learning with Active Neural Generative Coding.
Alexander OrorbiaAnkur Arjun MaliPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- fitted q iteration
- network architecture
- coding scheme
- generative model
- function approximation
- neural network
- coding method
- bio inspired
- markov decision processes
- optimal control
- reinforcement learning algorithms
- temporal difference
- inter frame
- state space
- optimal policy
- model free
- policy search
- transfer learning
- supervised learning
- action selection
- dynamic programming
- reward function
- learning algorithm
- neural model
- reinforcement learning methods
- robotic control
- machine learning