Using swarm intelligence to optimize neural network hyperparameters: comparative analysis on MNIST and CIFAR-10

Abstract

Swarm Intelligence offers powerful methods for solving optimization problems used in configuring hyperparameters of neural networks. This article examines the performance of the particle swarm optimization algorithm compared to Grid Search on two different datasets: MNIST and CIFAR-10. Experimental results show that the effectiveness of optimization methods varies depending on the complexity of the task and the data.

Keywords

• swarm intelligence, hyperparameter optimization, neural networks, particle swarm optimization, grid search, MNIST, CIFAR-10, algorithm efficiency, machine learning methods.