RL based hyper-parameters optimization algorithm (ROA) for convolutional neural network.
Fatma M. TalaatSamah A. GamelPublished in: J. Ambient Intell. Humaniz. Comput. (2023)
Keyphrases
- optimization algorithm
- convolutional neural network
- hyperparameters
- model selection
- reinforcement learning
- cross validation
- bayesian inference
- closed form
- random sampling
- support vector
- face detection
- bayesian framework
- em algorithm
- maximum likelihood
- multi objective
- prior information
- maximum a posteriori
- noise level
- incremental learning
- sample size
- incomplete data
- optimization method
- evolutionary multi objective
- missing values
- differential evolution
- particle swarm optimization pso
- parameter settings
- artificial bee colony
- hybrid optimization algorithm
- neural network
- state space
- parameter space
- object detection
- multi class
- data sets
- expectation maximization
- probabilistic model
- prior knowledge
- learning process
- machine learning