Introduction
Neural networks aren't only for images and text. They can predict numbers — house prices, delivery times, energy demand, the position of an object in an image. This is regression, just as in the ML track, but now the model is a multilayer perceptron that can learn complex, non-linear relationships on its own.
The good news: building a regression MLP mostly means getting three choices right — the output layer, the loss function, and the scaling of inputs and targets. The hidden layers are the same as for any other MLP.