Classification of multiple classes that are not linearly separable from each other is a challenging task in machine learning. Describe the fundamental differences between the ways that these two different methods, SVM and NN, address nonlinear separation problems and compare their advantages and disadvantages.