SandwichNet is an interactive learning tool to explain the functioning of simple neural networks in a teaching context.
A visualized neural network can be manually weighted in order to classify input data. The input data is represented by different sandwich combinations, the output should classify whether a sandwich is edible or inedible.
By manually weighting the network, it is possible to understand at each level how individual neurons are combined with each other and what effect an activation function has.
The tool was programmed in P5 (https://p5js.org/)