Just like a Brain - Convolutional Neural Networks

“Just like a brain” is the phrase used to describe Neural Networks; although an exaggeration they are rather abstract high dimensional functions capable of learning things that would be otherwise impossible to program. Dedicated for complex problems that don’t yield to traditional algorithmic approaches Neural Networks are simple to construct and often appear to be a good approach to certain classes of problems, such as pattern recognition. 

 They use different layers of mathematical processing to make ever more sense of the information they are fed, from human speech to digital image. Essentially, they learn and change over time. That’s why they provide computers with a more intelligent and nuanced understanding of what confronts them. 

 Herein we construct two such neural networks – an Artificial Neural network (ANN) to provide churn modeling of customers of an institution - And a Convolutional Neural Network (CNN) that learns to predict animal species from a huge dataset of images.

Click Here To Download The Paper