The Foundations and Impact of Neural Networks in Computational Intelligence

This review paper explores the foundations and impact of neural networks in computational intelligence, tracing their evolution from the perceptron to multi-layer perceptrons (MLPs) and radial basis function networks (RBFNs). It examines MLPs’ role as universal function approximators, regularization techniques (e.g., early stopping, weight decay) to prevent overfitting, and unsupervised learning paradigms. Aimed at students and researchers, it clarifies neural network principles, applications (e.g., pattern recognition, function approximation), and challenges, laying the groundwork for future advances in artificial intelligence.