For any cybernetic system e.g., a regulator (referred to also as a monitor-and-act unit), it is necessary to know which variable to monitor, which variable to affect, how to affect them, etc. Thus, in order to perform regulation successfully, the system must be built such that it does the job well. The hardware and the software of the systems, all of its components, are relevant for how it will do the job. We refer to these parts of the system collectively as the system’s cybernetic knowledge.
For example, a heating system that regulates room temperature, consists of a number of components tuned such to make it possible to keep the space comfortably warm. It has a heating device; and another device to measure the room temperature; and a decision machinery to choose whether to heat or not; and also possibly a machinery to decide how strongly to heat, etc. This entire structure, combined, contains the cybernetic knowledge of the system. These components jointly satisfy the requirement for the system to be a good regulator.
Practopoietic theory is concerned with the question of how this machinery comes about. How is cybernetic knowledge being created within adaptive systems? Or a better to ask: How is this knowledge being extracted? Following the Good Regulator Theorem, it becomes clear that, to be successful, the system must be a model of its environment. So the question is then: How is the necessary information extracted from the environment and implemented into the cybernetic system? How does our body know what to do? How do our reflexes become a model of the surrounding world? How do we as persons know what the right action in a given situation is? How do mental processes (perception, thought, consciousness) contribute to the process of extracting cybernetic knowledge from the environment and eventually, guiding our behavioral actions in an intelligent and adaptive way?
These questions are addressed by the theory of practopoiesis.
In machine learning, cybernetic knowledge can be referred to as an optimal policy.
Back to: Key concepts of practopoietic theory