PSO-PS: Parameter Synchronization with Particle Swarm Optimization for Distributed Training of Deep Neural Networks.
Qing YeYuxuan HanYanan SunJiancheng LvPublished in: IJCNN (2020)
Keyphrases
- particle swarm optimization
- pso algorithm
- inertia weight
- neural network
- particle swarm optimization pso
- particle swarm optimization algorithm
- training process
- training algorithm
- genetic algorithm
- global optimization
- swarm intelligence
- particle swarm
- feedforward neural networks
- differential evolution
- particle swarm optimizer
- feed forward neural networks
- multi objective
- neural network training
- hybrid algorithm
- deep architectures
- particle swarm optimisation
- economic dispatch
- pattern recognition
- convergence speed
- shuffled frog leaping algorithm
- error back propagation
- global search
- back propagation
- ant colony optimization
- metaheuristic
- multi layer perceptron
- backpropagation algorithm
- distributed environment
- neural network structure
- hybrid particle swarm optimization
- bp neural network
- distributed systems
- multi agent
- artificial neural networks
- multi swarm
- recurrent networks
- population diversity
- recurrent neural networks
- premature convergence
- feed forward
- multi layer
- evolution strategy
- optimization algorithm
- hybrid optimization algorithm
- hybrid pso
- concurrent processes
- hidden layer
- evolutionary algorithm
- activation function
- neural nets
- genetic programming
- optimal parameters
- radial basis function network
- deep learning