Evolving large-scale neural networks for vision-based reinforcement learning.
Jan KoutníkGiuseppe CuccuJürgen SchmidhuberFaustino J. GomezPublished in: GECCO (2013)
Keyphrases
- neural network
- reinforcement learning
- function approximators
- pattern recognition
- function approximation
- genetic algorithm
- state space
- connectionist systems
- real world
- small scale
- neural network model
- real life
- real time
- fuzzy systems
- human computer interaction
- learning capabilities
- computer vision
- learning algorithm
- fuzzy logic
- artificial neural networks
- augmented reality
- transfer learning
- markov decision processes
- multi layer
- temporal difference
- learning agent
- supervised learning
- model free
- vision system
- evolutionary algorithm
- multi agent
- decision making
- data sets