Understanding the current state of ML
To understand why the benefits derived from the application of FL can outweigh the increased complexity of this approach, it is necessary to understand how ML is currently practiced and the associated limitations. The goal of this section is to provide you with this context.
What is a model?
The term “model” finds usage across numerous different disciplines; however, the generalized definition we are interested in can be narrowed down to a working representation of the dynamics within some desired system. Simply put, we develop a model B of some phenomenon A as a means of better understanding A through the increased interactability offered by B. Consider the phenomenon of an object being dropped from some point in a vacuum. Using kinematic equations, we can compute exactly how long it will take for the object to hit the ground – this is a model of the aforementioned phenomenon.
The power of this approach is the ability...
