Chapter 6
Nonlinear Networks
6.1 Advantages of Nonlinearity
Linear systems (both neural networks and other linear devices in general) are well known for their predictable nature and use of simple mathematical descriptions to lead easily to solutions. You may wonder why we leave the comfortable area of linear networks and move on to complex nonlinear networks. The most important reason is that networks of nonlinear neurons are able to solve more varieties of problems. Linear networks can solve only a single class of problem—one in which the correlation between output and input signals is linear. Nonlinear networks are far more flexible.
If you are adept at mathematics, you will understand what the problem is, but don’t worry if mathematics is ...
Get Exploring Neural Networks with C# now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
