Below is a link to a fascinating article about Karl Friston, whose research on the “free energy principle” (also known as active inference) tries to explain how biological (or even artificial) systems maintain or even increase their organization. Without much explanation, here are some details. The model is a system with four main variables:

• Sense (s)
• Action (a)
• Internal States (r or μ)
• Hidden States (ψ)

Sense and action divide the internal states of the system from the hidden states external to it. This division is called a “Markov Blanket”. I’ve tried to show the equations between the variables correctly but I’m not too sure if I have them right. They seem to change from paper to paper. F is an expression of the free energy, but I’m not sure what f is at the moment, except for being some sort of “flow”.

Further Reading:

https://www.wired.com/story/karl-friston-free-energy-principle-artificial-intelligence/

https://en.wikipedia.org/wiki/Free_energy_principle

https://en.wikipedia.org/wiki/Markov_blanket

http://rsif.royalsocietypublishing.org/content/15/138/20170792

[*11.6]

<>