Mark Hudson Beale
President

Home > A.I. Technolgy > Neural Networks

Neural networks are software tools that learn from examples, in a way loosely modeled on the behavior of real biological neurons. Neural networks are capable of solving almost any problem, given enough examples and training time.

For instance, by training a neural network on old loan applications and their repayment histories, the neural network can learn to predict the credit worthiness of new loan applications.

If a neural network is trained on sounds recorded from a submarine, and their actual sources (another submarine, whales, etc.) the neural network can then be used to continuously listen for sounds and report the source it hears.

The Neural Network Toolbox for MATLAB provides a set of tools for training neural networks, and deploying the solution.

The Neural Network Design textbook explains the mathematics behind neural networks, and includes interactive visualization software that makes the important concepts intuitive and easy to understand.

We are available for courses on neural networks and consulting on neural network problems.