The first modular unit of a self-expanding cognitive force might be viewed as two separate entities with different functions, but a shared objective.
1) Local Decision Machine (LDM)
The first entity (LDM) is any reasonably competent Simile Learning Machine operating within its given decision environment. This machine (LDM) will react to local fluctuations as effectively as it can, in its current immature configuration. (I recommend a Simile Flux Reactor Shaping Machine [see last section] with an i/o quark-particle interface.)
2) Changes & Expansions Manager (CEM)
The CEM will model the Entity Task Universe (ETU) and "observe" the LDM in interaction. It will make experimental changes in the LDM. This C&E Manager will also create a number of clones of the LDM to observe (during operation) various "portions of" the LDM and its ETU.
Entity Expansion Cubed
There are 3 Elements, 3 Realms, and 3 Directions of Entity Expansion In Cognitive Space
ELEMENTS: (cognitive "hardware")
(1)LDM, (2) CEM, (3) ETU RAModel
REALMS: (within the ETU model)
(1) Entity, (2) Environment, (3) i/o Interface
DIRECTIONS: (observer/shaper-perspective types)
(1)Implosion (in-line), (2) Explosion (overview), (3)Parallel (imperfect)
Expansion
Any effective learning machine (LDM), once introduced to its target environment, may be defined as a three part Entity Task Universe (ETU) of matter and motion (m&m). In order for an entity to self expand to the given task, these three occasionally overlapping realms of a newly immersed entity must be modeled (see outline of ETU), and subjectively monitored by clones of the (LDM).
Here is a brief analogy to illustrate the "occupations and relationship" of these two entities (LDM and CEM) which comprise the first module of an expanding cognitive force:
The LDM attempts to perform a defined decision task, while the CEM (the C&E Manager) observes and suggests changes in the "action and material" of the LDM. The C&E Manager also creates clones of the LDM, and then tells these new entities "when and where" to set up and perform.
After writing this paragraph, I couldn't help but notice that this language describes certain notorious human occupational relationship duos, such as an entertainer and his manager, "Elvis and the Colonel" for instance. Since the more amusing analogies are frequently the better conduits to parallel order continuity, I was moved to mention this one.
One interesting incongruence in this comparison is that instead of creating a display for an audience, "Elvis's talent" is to discover and experiment with matter and motion as it appears from his unique perspective on "stage". Another difference is that a computer hosted "Colonel" can clone and book a vast number of "Elvi" at selected trial perspectives. Notice that such entities may optionally be cloned imperfectly.
To explain the "when and where" aspects a newly cloned decision entity, I first suggest there is more than one way to "watch" a machine operate. One may consider all of the machine and its material, or merely part. This selected configuration of interest may also be "observed" from a number of different value inhibited perspectives.
Expansion of Cognitive Matter Occurs in Three Dimensions
1) In-line Decision Post (implosion)" a local processing decision is defined as the responsibility of a new "somewhat independent" entity.
2) Overview Post (explosion) portions of the ETU are adopted as the universe of a new OSP. Input/Output patterns will be sought that reliable effect the entity measurements of value levels in the target environment.
The first two (implosion and explosion) involve creating new OSP entities. The third kind of expansion is parallel or "slightly imperfect" expansion.
3) Parallel expansion is the trial mutations (usually minimal) in processing structure and/or event definition clusters (memories). (One may also wish to view the parallel expansion of an existing OSP as a new entity.)
---
Random Perspective Expansion (RPX) within the model of the ETU will display patterned references to parallel environmental order. Fixed points of view from within a given thinking machine reveal inherent order (solve problems) existing within the target environment.