Neural Network Learning: Theoretical Foundations

The authors explain the role of scale-sensitive versions of the Vapnik Chervonenkis dimension in large margin classification, and in real prediction, and discuss the computational complexity of neural network learning.