Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning.
Felipe Petroski SuchVashisht MadhavanEdoardo ContiJoel LehmanKenneth O. StanleyJeff ClunePublished in: CoRR (2017)
Keyphrases
- neural network
- genetic algorithm
- artificial neural networks
- reinforcement learning
- fuzzy logic
- training process
- deep architectures
- multi layer perceptron
- training algorithm
- function approximators
- modular neural networks
- feedforward neural networks
- learning classifier systems
- pattern recognition
- simulated annealing
- genetic programming
- multi agent
- neural network structure
- deep learning
- training set
- training phase
- function approximation
- particle swarm optimization
- neural nets
- radial basis function
- markov decision processes
- active learning
- learning capabilities
- fault diagnosis
- fitness function
- metaheuristic
- back propagation
- support vector
- training samples