Trinity: Neural Network Adaptive Distributed Parallel Training Method Based on Reinforcement Learning.
Yan ZengJiyang WuJilin ZhangYongjian RenYunquan ZhangPublished in: Algorithms (2022)
Keyphrases
- neural network
- training process
- reinforcement learning
- cost function
- preprocessing
- training algorithm
- parallel implementation
- significant improvement
- dynamic programming
- pairwise
- high accuracy
- objective function
- clustering method
- fuzzy neural network
- feedforward neural networks
- learning capabilities
- parallel computing
- neural network is trained
- segmentation method
- optimal policy
- detection method
- training samples