Natural intelligence as a perceptron – how to build an UAA-system

How to build a system with the characteristics of an UAA-system ? In the next, I'll describe an UAA-system as a perceptron, a li...

How to build a system with the characteristics of an UAA-system?

In the next, I'll describe an UAA-system as a perceptron, a limited size (has physical borders) cell automata. The system homogeneous, the cells have the same number of neighborhood cells and the connections to the other cells are the same for each cell. No geometrical changes, no physical movements occur between the cells in the cell space during the operation. The structural change achieved by the change of the cell’s status.

Description of the perceptron’s cell space and change-function

Cells

The cell-space contains identical cells. The cells have to be capable of storing and working according to the change-function. The cells classified into three groups by its functionality:

Sensor-cells: receives the stimuli sent by the sensors

Effector-cells: emits stimulus to the effectors

Cells belong in the third group placed between the sensor-cells and effector-cells. The secondary structure (described later) built up from these cells.

Primer structure

The cells' physical connections to each other are called primer structure. In the described model the primer structure is the same in any part of the cell space. It means any cell connects to the same number of other cells in the same manner. It’s called a cell-homogeneous structure.

This homogeneity applies only to the primer structure. The advantages of this method are easier to build the system, the primer structure does not change when the secondary structure changes, no geometrical changes, no physical movements occur in the cell space during operation. The other advantage is, that the change-function is the same for all cells.

The proper determination of the primer structure is crucial for an efficiently working system.

Secondary structure

During the operation necessary to connect the sensor-cells and effector-cells to each other through signal-carrying channels. These channels build-up from the primer structure utilizing the change-functions. Under the operation, the status of the cells (originally in base-state) changes. Based on these changes, a particular structure, signal-carrying channels develop, which can carry the stimulation provided by the sensor cells and preserving the stimulation’s specificity. The nature and the build-up of these channels described later.

Change function

The system has to accomplish mechanisms to build up the secondary structure. The system's operation based on the change-function and the change-function has to realize these mechanisms. Let see the operational mechanisms one by one and examine them on how to accomplish it by the change-function.

The determination of the primary critical stimuli

a./ At the sensor side, select cells, which connect to sensors. The role of these cells is to sense stimuli which are the primary critical stimuli determined by the builder of the system. If the primary critical stimulus appears in the environment, then the cell change to the active state and this activity spread across the cell-space changing the state of the connecting cells. This activity does not activate the effector-cells and stay as long as the stimulus exists.

b./ The primary critical stimulus can be based on the stimulus intensity limit rule. In this case, do not have to select specific cells to connect the primary critical stimulus sensors by the builder, but have to define a limit. If the sensor cell gets a stimulus in which intensity is over on the given limit, then the change-function value of this stimulus change to primary critical stimulus. This kind of stimulus travels through the whole system, through the boundary of the stimulus routes, but not activate the effector cells.

The primary critical stimuli only cause, but does not produce effector activity, not have to be specific but have to spread freely in the cell-space because they have to indicate the existence of a primary critical stimulus everywhere in the cell-space. Its role: a kind of cell-space activation which indicates intervention, effector activity needed. 

The primary critical stimuli start from the designated sensor cells, or from every sensor cells depending on the stimulus intensity as described before.

Spreading of the non-primary-critical-stimuli (neutral or secondary critical stimulus) in the cell-space

These stimuli have to be specific (not like the primary critical) does not spread freely in the cell-space. The spreading has to be realized in channels.

Those sensor-cells which connected to non-primary critical sensors need to prepare to build channels, need to set up channel-seeds in these places. The channels will grow from these seeds by the stimulation of the sensors.

Build up stimulus-transfer gates at the same time active contacted channels with resistance signal transfer

If stimulus spread in the channel, the border-cells of the channel become activity-signaling cells. If an activity signaling cell contacting to other channel's activity signaling cell (both pathways are active in the same time), then change the state of the contacted cells to gate-state (association gate, association connection), which later makes it possible that one of the channel's activity induces activity to the other one. The pass-through of the stimulus has to be with resistance. Otherwise all of the other, secondary and farther association-connected channels become active, won't prevail the stimulus intensity rule.

The resistance makes it possible for the stimulus to go through (a weaker pass-through) across the association gates compared to the original strength of the stimulus. It means that the stimulus can go through association gates proportionally its strength.

Connection of the sensor and effector cells

Necessary to connect these cells, when a primary critical stimulus is present in the cell-space, the system needs to issue an answer.

If an activity-signaling border-cell has an effector-cell and a primary-critical stimulus-signaling cell in its neighborhood, then create a signal-carry-over gate, which can transfer the signal of the active channel to the effector field's given part.

Affirmation

The cross-section of the channel is growing on the neutral cells account by the repeated stimulation of a channel. The numbers of the signal-carry-over gate-cells growing, when the necessary condition of its formation exists.

Forgetting 

Stable formations (channels, gates) transform into neutral cells based on random selection. The transformation processes are:

gate cell -------> border cell 

channel base-state cell -------> border cell 

border cell -------> neutral cell

The transformation should not create irrational structure, for example, do not let to form a border cell inside a channel. 

Inhibition

If the critical stimulus is next to the existing signal-carry-over gate-cells, then do not let the non-critical stimulus goes through over the gate (blocking).

The mechanism of how to count the power of the stimulus

If the power of the stimulus grows, the frequency of the sensor-cells to be in an active state grows (until the maximum frequency).


(a part of a diploma work of nineteen-eighty-eight)