duncan.bandojo

In the age of artificial intelligence, it is common to meet the term neural network, seeing diagrams of neurons connecting to other neurons, programmers training models, and so on. Here, we will discuss how neural networks are similar to plain mathematical functions (models), how they build upon traditional linear regression, and their application to visualize images with convolutional neural networks. The usual pedagogy for learning neural networks begins with the iconic diagram: columns of circles ("neurons") connecting to other circles across multiple layers, passing information from input to output. But what is this diagram trying to capture? And what really [...]