Backprop-Free Reinforcement Learning with Active Neural Generative Coding.
Alexander G. Ororbia IIAnkur Arjun MaliPublished in: AAAI (2022)
Keyphrases
- reinforcement learning
- fitted q iteration
- generative model
- coding scheme
- network architecture
- neural network
- coding method
- bio inspired
- reinforcement learning algorithms
- function approximation
- neural fuzzy
- neural model
- state space
- learning algorithm
- optimal control
- machine learning
- supervised learning
- artificial neural networks
- markov decision processes
- unsupervised learning
- learning rules
- model free
- multiscale
- dynamic programming
- discriminative learning
- multiresolution
- robotic control
- learning process