EC Container 5

Components of an Intelligent Machine

Mon, Jul 12 2010

Components of an Intelligent Machine

Inspired by Asimov to build an intelligent machine, but never found instructions? Here we cover the basics.
Subscribe iTunes Store

Show Length: 10:35

A Stored Purpose system is based upon a native intelligence derived from two architectures, the existence model, called Ema and the construction model, called Mica. Those two architectural models, when superimposed, make complex intelligent entities possible.

Two Architectures Superimposed to Create Complex Intelligent Systems

The two designs that make Stored Purpose intelligent machines possible are the:

Existence Model Architecture (Ema) - Ema defines a fundamental logical model of intelligent existence and can be used to create complete “single cell” intelligent entities. Each cell, or Ema entity is an atomic intelligence unit, possessing the minimum requirements for intelligence.
Multi-level Intelligent Cellular Architecture (Mica) - Mica defines physical construction for complex Stored Purpose entities. It combines multiple Ema cells, as a single cluster or as multiple clusters, to make systems of varying complexity.

Existence Model Architecture

What separates intelligent systems is that they Exist, i.e. they process a function to sustain a definition of their existence. That function of existence requires two parts:

A definition of the Self which we call Identity, and
Processes to instantiate and maintain the Self … even when the entity and it’s surroundings, called “context” change.

For intelligent systems, we define a third requirement, … an ability to execute a Purpose other than maintain Self. Yet meeting requirement, and thus serving some additional Purpose in the ecosystem … seems an unavoidable and automatic consequence of simply executing the first two parts of Existence. As such, there is no need for additional design components to perform this function.

We developed a method to define Identity, that we call the Platonic Form. We also created an algorithm that that enables a defined entity to interact with a physical world via technology, and align it’s measured and planned states with its Identity that we call the General Intelligence Architecture, or Gia.

Together, the Platonic Form and Gia, when processed in a special type of container, called a Soma, make possible an intelligent entity.

The Platonic Form is created in a format we call the purpose hypergraph. It combines contextual fabric with form subgraphs to make possible the storage of the “essence” of a symbol. Platonic Forms combine other Platonic Forms to define symbols of great complexity. A Form that defines a “plank”, can be used as the basis of a Form that defines a “door.” A rectangle can be used to define a box which can in turn be used to define a room which can be used to define a bathroom.

1. Organize Forms, 2. Compose Forms to Plan, 3. Act at Time Now - Reality, 4. Sense in Context, 5. Pursue Goals, 6. Understand, 7. Learn

The General Intelligence Algorithm is the outcome of seven processes. This processes constantly compose the entity’s Identity definition (#1) into a type of memory called Plan (#2). Plan memory can be located in one physical device or propagated across the many physical devices of a metacomputer. Once Forms are in Plan, an active or conscious system will continuously process the Plan data, just before time Now, to prepare for Reality. At time Now, which is Reality, the system will sense (#4) and control (#3) all that has been planned with its states already pre-set to the values planned. This means that if a Stored Purpose system plans right, it doesn’t really need to do anything at time Now, because by executing Plan, it has already loaded the data to be sensed and executed the shape of motive effort need to influence the physical world.

When there is a difference found between Plan and measured data, then outcomes five, six and seven of Gia kick in. These are Goal Pursuit (#5), Understanding (#6), and Learning (#7). Goal Pursuit is simply the re-alignment of the entity’s measured Self with Identity. Understanding is simply a measure of the difference or impact to Plan caused by an event measured or in Plan, caused by Goal Pursuit. Finally Learning is simply readjusting the system’s Forms as might me needed due to prediction failure, or due to new symbols gathered in it’s processing.

Multi-component Intelligent Cellular Architecture

Ema makes simple intelligent entities possible, but sometimes an entity’s Purpose requires more capability, i.e. an entity with multiple component parts that enable it to perform diverse functions simultaneously. Mica makes this possible by combining individual Ema entities, which we call Cells into more complex entities. Each cell is based upon the same Identity fabric as all others, but will have a different role defined at a specific location in the contextual fabric of the Identity.

As cells, called Instruments in mechanical systems, are combined, each cluster of cells can have its own contextual purpose and thus complex entities can be created from many diverse “subsystems.” Each subsystem is based upon the same Identity, drawing it’s Purpose from a unique position in the context of the Identities purpose hypergraph. The Mica design provides for intelligent designs with limitless sophistication, yet efficient, due to the fact that all subsystems are based upon the same script, that is the overall system’s identity.

Terminology

While intelligent mechanical technology (mechatech) has many similarities to bio-technical (biotech) systems, we use different terminology for the base, Level 1 system (cell, instrument) and the complex Level 4 systems (body, metacomputer). The following table illustrates additional terminology differences and provides examples of systems at each level.

Multi-Level Intelligent Cellular Architecture