When conducting (deep) neural networks, we usually have a certain outcome in mind (e.g., identify handwritten digits). There are several ways for us humans to influence the way
