ReLeQ : A Reinforcement Learning Approach for Automatic Deep Quantization of Neural Networks.
Ahmed T. ElthakebPrannoy PilligundlaFatemehsadat MireshghallahAmir YazdanbakhshHadi EsmaeilzadehPublished in: IEEE Micro (2020)
Keyphrases
- neural network
- reinforcement learning
- learning capabilities
- fuzzy logic
- pattern recognition
- function approximation
- function approximators
- multi agent
- state space
- fully automatic
- reinforcement learning algorithms
- multi layer
- learning algorithm
- learning classifier systems
- semi automatic
- data sets
- optimal policy
- supervised learning
- feed forward
- recurrent neural networks
- dynamic programming
- learning process
- artificial neural networks
- video sequences
- genetic algorithm
- deep learning
- multiscale