When a lot of background knowledge is available about a topic, why not use it when making decisions? This is called a knowledge-based system. Inference engines in expert systems and unification, as done in logic solvers, are examples of this.
Another way to retrieve knowledge when making decisions is based on representing knowledge in a graph. Every node in the graph represents a concept, while every edge represents a relationship. Both can be embedded and represented as numerical features that express their location with respect to the other elements of the graph.
In this recipe, we'll go through two examples for each of these possibilities.