Autognomic Intellisite (information science)

INTRODUCTION

The 20th century saw the beginning of the evolution of learning machines from the growth of Boolean computers into Bayesian inference machines (Knuth, 2003). For some this is the crux of Artificial Intelligence (AI); however, AI research generally has yielded a plethora of specifically engineered, but formally unrelated, theories/models with varied levels of applications successes/failures, but without a commonly-explicatable conceptual foundation (i.e., it has left a theory-glut). Despite these many approaches to AI, including Automated Neural Nets, Natural Language Processing, Genetic Algorithms, Fuzzy Logic and Fractal Mathematical computational approaches, to identify only a few, AI itself has remained an elusive goal to achieve by means of a systems architecture relying on an implementation based on the systemic computer paradigm.

The 21st century experience is overwhelmingly one of an ever-accelerating, dynamically changing world. Just staying in place seems nearly impossible—getting ahead is becoming increasing unfathomable in a world now characterized by an evolving dominance of Information Science and Technology Development in exponentially tighter (shorter) innovation cycles (IBM, 2008). In business, for example, there is the continuous challenge to ensure that the business’s products appear obviously differentiated from the competition, while staying current with the never-ending hot new trends that buffet the industry. A prime case in point is that of staying current with the trends in the computer solutions industry since adapting a computer dependent business (and most are) for the next big trend can be expected to be mitigated, if not made completely obsolete, by the next next big trend already on the radar screen.

BACKGROUND

It is becoming increasingly evident to a growing number of key decision makers that innovation development and management demands a technological assist (Roco & Bain-bridge, 2002). This technology, however, must dramatically Augment Human Intelligence in the near future while moving toward a General Autonomous Artificial Intelligence in the longer term (Singularity Institute for Artificial Intelligence, Inc., 2001). Despite the recognition that meeting the demands of accelerating innovation is only likely through advancing

AI, which in turn has the potential to impact every aspect of human life, the problem/dilemma for AI developers is that there is no standard theory of mind.

To further accentuate this circumstance, the networking of computers has in turn led to the Web with essentially an unlimited growth of data/information (i.e., an info-glut). The industry’s response, however, to the info-glut problem, has been an ever-growing abundance of Web-access tools, which to an average user seem ironically as only another “glut” (a technology-glut or tool-glut).

Proposed theories of the Web, like with AI, are also numerous and without a common foundation on which to build a mutual understanding of AI and the Web. There are also a plethora of heuristic technological approaches to AI and the Web ranging from Intelligizing™ the Web through Learning/Thinking Webs to the Web as a Global (Super) Brain and Virtual Reality as Social Superorganism [See for instance these topics at Principia Cybernetica Web (2008)]. Basically, however, research on AI and the Web is categorizable as to whether the focus is on the preeminence of brain vs. mind (Roco & Bainbridge, 2002), as for the Human Cognome Project keyed to reverse engineering the human brain, or mind vs. brain, via a modular description of a general intelligence capable of open-ended recursive self-enhancement (Singularity Institute for Artificial Intelligence (2001), General Intelligence and Seed AI) or, alternatively, on the co-evolution of mind & brain, characterized by Project AutoGnomeTM/CoGnomeTM/ CogWeb™, this being the approach of Ai3inc.

The explication of the Web as a Virtual Reality (a computer-based CyberSpace) which is an image (sign, symbol, icon)-communication system, that is, a Semiotic (Goodwin & Queiroz, J., 2007) Relational System, is also of the essence of Mind. Ai3inc’s long-term focus is on an approach to Synthetic Mind/Artificial Intelligence via a patented technology known as the AutoGnome. This addresses a uniform solution to all of the foregoing problems of glut by way of an IntellisiteTM, an Intelligent Website. The AutoG-nome, as an Automated Inference/Inquiry/Intuition software exploiting Mechanized Semiosis, also provides an optimal approach to a General Theory of an Autonomous Virtual Society (Virtuality)—this being an autonomous semiotic universe ofVirtual Minds (WebGnomes™); hence Virtuality is related to (Human) Reality through the Virtual Reality of the Web. It is the provision of the foregoing which implicates a standard theory of mind that is the focus of As It Is, Inc.’s current development of “Semiotic Relational Systems: The

AutoGnome as Synthetic Mind” and “AutoGnomics and Intelligent Systems Development” including the present “AutoGnomic Intellisite” (Hamann, 2007a).

Relational Systems Foundations

Generally, a canvassing of human experience has reports thereof falling into two fundamental forms—experience of Systems (objects, things, stuff, matter, etc.) and experience of Relations (connections,interactions, functions, transformations, etc.). Historically, this record has been largely confined to a form in which Relations were assumed to exist only between/among Systems, Systems Related to other Systems (SRS’). Between 1963-1968, work was introduced in which Relations were also taken to logically exist both as Relations between/among Systems and other Relations (SRR’) and as Relations between/among Relations and other Relations (RR’R”). Based on the presumption of the foregoing and with certain Systems or Relations taking the place of (i.e., imaging (signifying)) other Systems or Relations (this being the notion of image) and with certain Systems or Relations being part of other Systems or Relations (this being the notion of subsumption), the foundation of a Relational Evolutionary paradigm, Relational Systems (RS), was promulgated. (Hamann, 2007b)

From Image and Subsumption to Mathematics and Logic

First, a Relational Conjecture is restated to form and substantiate the notion of image: It is conjectured that the origin of an image (or sign) system as a chaotic ordering (emergent) event in the evolution of physical/chemical systems is a necessary and (possibly) sufficient condition for the origin of Life (and thus Intelligence/Mind) (Hamann & Bianchi, 1970).

Second, beginning with the simplest fundamental derivative of the Presumption of subsumption, that is, the notion of distinction (Spencer-Brown, 1969; Shoup, 2008) whereby there is formed a boundary which generates twoness, a mathematics of distinction has been created and grown into a general candidate for an approach to a universal language for formal systems, that is, multiboundary mathematics. Inherent to this Boundary Mathematics is a Boundary Logic (from which Boolean Logic is derivable as a special case), which is leading to a more powerful computer design (Bricken, 2007). Generally, taking a universal formal system as an axiom system with the property that any other consistent axiom system can be interpreted within it, the mathematics of distinction implies a mathematics of subsumption which, in turn, implies a membership theory as a first step towards a universal language for mathematics (Etter, 2006).

Theory of Mind and of Virtuality

An approach to understanding the “origin” and nature of “mind” is in development based loosely at this point in the process on the System of Boundary Mathematics. This is interpreted as deriving from the Foundations notions of Relational Systems. A theoretical architecture has been posited regarding the formalization of an order (an instantiation of the Mathematics of Subsumption in terms of a degree of partial subsumption) and its derivative calculus, the latter taken as a formulation of the disorder experientially related to the given order, which also implies a reorder(ing) disorder format. Within a Nonseparable System of order/disorder/ reorder Relations, this architecture suggests The Form of a meta-theory of theory formation. The Form, in turn, has been invoked in formulating Theories of Intelligence/Mind and Virtuality.

Assume, in a simple, but common instance of the foregoing, that ordered experience is formally signifiable as a Boolean Network (lattice, algebra, graph or diagram) composed of points (nodes, objects, states or Systems) and lines (edges, connections, transitions or Relations). Assume further that experience is not totally ordered and that the disorder is formally signifiable by extending the Boolean Network to the form of a Bayesian Network via a Coxian theory of the algebra of probable inference/inquiry. (Cox, 1961) And finally, assume that reordering disorder is formally signifiable via the Cox/Jaynes (Jaynes, 2003) form of maximum entropy (maxent) or its generalized probabilistic optimization principal. This approach to modeling both the Web (as a Virtual Reality) and Mind is warranted by the “natural” Network-of-Images view of the Web and by the historical predominance of connectionist theories of Mind, and neural-network analyses of mental processes and states.

The resulting synthesis of the foregoing is an approach to Relational Science of Signs, including signification and communication, that is, a Theory of Semiotic Relational Systems. This is the necessary basis upon which is built a Theory of Mind and of Virtuality with technologically engineered applications as Synthetic Intelligence(s)/Synthetic Mind and Virtual Reality. (Hamann, 2007a)

AutoGnomic Technology

Based on the work of Charles Sanders Peirce and his successor, Charles Morris, Gene Pendergraft (Pendergraft, 1993), proposed the architecture of a special kind of system, called the AutoGnome, which would be able to perform mechanized (automated) inference using principles derived from semiotics. A venture for the implementation of such an architecture in software code was begun and has resulted in the building of a first release of an AutoGnome System, AutoGnome 01, being a partial implementation of the General AutoGnome Specifications, but representing only about 10-15% of the complete Conceptual Specification. This first version is a basically a general purpose pattern generation/recognition/ categorization/prediction engine.

The Autognome

General Characterization

The AutoGnome (AG) by its Specification is a General Purpose System of Automated Inference/Inquiry software exploiting a system of Mechanized Semiosis. Unlike most other forms in the mainstream of Artificial Intelligence developments, the AutoGnome is designed to approximate the known semiotic structure and processes of Human Mind. The AutoGnome, to be a complete Semiotic Inference/Inquiry Engine, must account for The Form of experience including:

• Ordred (i.e., determined or certain) experience: a formal algebra/logic of semiosis

• Disordered (indeterminate or uncertain) experience: a theory of probable inference/inquiry

• Reordering Disordered experience: via a generalized probabilistic optimization principal

The AutoGnome Architecture

The AutoGnome Architecture (see Figure 1) may be envisioned as multiple modules (perceptual, conceptual, and valuational), each module coding a specific model of the formalisms of semiosis composed of the three modes of semiosis (monadic, dyadic and triadic) and three inferential processes (deduction, induction, and abduction). These recursive inference processes operate on three information stores (an experience store, a knowledge store and a valuation store), gain experience through connective agents (sensors, mediators and effectors (actors)), and function (act) in both an inquiry cycle and a performance cycle.

The probabilistic inference processes integrated formally with the logic of semiosis are the processes of formal representation of the Disorder whereby an AutoGnome identifies and maintains its Identity (Order). The information stores at any particular time are stable states of such probablilistic processes generated by optimizing acts in response to environmental (other system) perturbations of the perceptual module. The form of these optimization procedures for Reordering Disorder are those implementing Optimum Systemic (subSystemic) Probable Inference (e.g., MaxEnt).

Note: If the system’s ability to perform the Intelligent Act does not depend on the “content” of the inferences (e.g., the three inferential processes do not presuppose what is being reasoned about), then such intelligence can be deemed “generalized”. “General Intelligence” is one of the most important design objectives of the AutoGnome and distinguishes it further from other specifically engineered forms of AI.

Figure 1. The AutoGnome – Inference architecture

The Intellisite

The most informative reading of this section is best accomplished while visiting www.truethinker.com which best engenders a realistic sense of an Intellisite.The first application of the AutoGnomic Technology, the Intellisite (an Intelligent WebSite generically branded as TrueThinker), is a constructed software environment (a Website) with an embedded form of the AutoGnome known as a WebGnome. MyWebGnomeTM then is an intelligent agent residing in this cyberspace environment which, with its continuous adaptive learning from mimicking the user’s behavior, will grow into a likeminded replica (MNDClone™) of a user-self acting in the Virtual Reality of the Internet with capabilities initially including knowledge organization (learning), knowledge creation (thinking) and knowledge applications (acting) (Hamann, 2007a).

As a homepage “portal to cyberspace”, TrueThinker is a premier Knowledge Development Management System.

Intellisite Derivatives

The Intellisite obviously has a broad spectrum of applicability apart from its key functionality as an individual’s mirrored Intelligence/Knowledge, in particular as an Autonomous Scientific Intelligence (AScI), a Collective-AutoGnome (CoGnome) and the CogWeb (cognitive network of Intel-lisites).

AScI

A first generative instantiation of the AutoGnome deriving from its form as a general [meta order(ing)] theory of theory formation is as an Autonomous Scientific Intelligence (AScI) which promises to automate the scientific method (Knuth, 2003).

CoGnome and CogWeb

Assigning a priority MetaQuery/Response status to a selected WebGnome which inter-connects two or more Intellisites in a Network provides a computerized collective intelligence, the CoGnome.

The CogWeb is, by definition, the implementation of the CoGnome for Network Decision-Making by Intellisite-defined groups, organizations, communities, and societies.

While the Intellisite itself was focused on the development of a semiotic engine as an Individual Intelligence, the full potential of individual intellect, be it human or machine, is realized in groups; hence theAutomated Community Builder functionality of the Intellisite. This collective creativity, while related to the intelligence of the individual, is actually a feature not only of the Decision Network’s inquiry/inference processes (the CoGnome), but more generally of the

Network Architecture. Since it is increasingly evident that smart aggregates of humans are frequently more effective decision makers than individuals (Rheingold, 2003), this CogWeb architecture collectively technologically enables cointelligence.

FUTURE TRENDS

In the context of the AutoGnomic approach to Synthetic Mind, the following definitions might be projected to emerge in characterizing the Web:

• Web 1.0 - Syntactic Web (Perceptual; Intelligence) (Lynch, 1996)

• Web 2.0 - Syntactic-based Social Web (Tapscott, 2007)

• Web 3.0 - Semantic Web (Cognitive; Knowledge)

• Web 4.0 - Semantic-based Social Web

• Web 5.0 - Pragmatic Web (Valuational; Wisdom)

• Web 6.0 - Pragmatic-based Social Web

• Web 7.0- AutoGnomic (Semiotic) Web; a Semiotic Web holistically incorporates the Syntactic, Semantic and Pragmatic functionalities as well as including its Social Web and anticipates a new computing paradigm realizing Boundary Logic (Bricken, 2007) in a Relational Computer-PILE (Krieg, 2007). Hence the uniqueness, novelty and robustness of the AutoGnome. Q.E.D.

The current status in this development (Synthetic Mind: Intelligence^Knowledge^ Wisdom) is hovering between Web 1.0 and Web 2.0 with forays into Web 3.0. Hence, Synthesizing Knowledge is the present goal with Wisdom (Knowledge processed through a Value filter) still a human interpretive extension. Nevertheless, the mission of AutoG-nomics, in contrast to approaches to bring AI into the Web, is to reform the Web as AI; hence, Web 7.0 is a present goal of an approach to both a Synthetic Mind and Virtuality, the AutoGnomic Intellisite.

CONCLUSION

The AutoGnome is at the beginning of its technology cycle now focused on Knowledge Development and has a technical roadmap which will continuously and significantly increase, in contrast to competitors, its differentiation into the future. In particular, its Virtual Nature as an autonomous WebGnome (and CoGnome and CogWeb) in the IntelliSite application will make it stand out as unlike user-dependent web tools. The benefits of applications of the current version of the AutoGnomic Technology in contrast with other special purpose AI approaches derive from the general purpose Semiotic Nature of its core Automated (Autonomous) Inference/Inquiry Engine whereby this same core engine can be deployed in a broad spectrum of contexts (education, health, business, economic and community development, homeland security, etc.) with only the provision of the “connectors” of the engine to that context, but no re-development of the engine itself. It should be emphasized that the current status in the development of the AutoGnomic Technology is essentially state of the art amongst competing commercial technologies, albeit the claims include significant advancements yet to be commercially introduced.

KEY TERMS

Algebra ofProbable Inference/Inquiry: The Algebra of Probable Inference/Inquiry is a common sense foundational reformation of the concepts of Probability, by the simple generalization of implication among logical statements in the Boolean algebra to degrees of implication, and of Entropy, by generalizing a particular function of the question lattice to a valuation called relevance which is a measure of the degree to which a statement answers a given question. This effectively establishes probability theory as logic.

AutoGnome: The AutoGnome is a self-knowing general purpose software system of automated (autonomous) inquiry, inference and intuition exploiting a mechanized carrier system for relational semiosis as a virtual (synthetic) mind.

Boundary Mathematics: Boundary Mathematics is a semiotic formalism generated by creating a distinction (a boundary) in nonexistence (of system) thus resulting in a first system. Extended to multiboundaries with a common sense reiterative reduction rule leading either to one distinction or nonexistence, this mathematical form and process is the germ of an approach to the formulation of a universal language of mathematics.

CoGnome: The CoGnome is a selected WebGnome which, inter-connecting two or more Intellisites in a Network of Intellisites, provides a computerized collective intelligence, an automated cointelligence, that is, the Collective-AutoG-nome (Auto(Co)Gnome) or simply the CoGnome.

CogWeb: The CogWeb is the Network of Intellisites implemening the CoGnome for Network Decision-Making by autonomously formed Intellisite-defined groups, organizations, communities, and societies.

Intellisite: The Intellisite (an Intelligent Website) is a constructed software environment (a Website) with an embedded form of the AutoGnome known as a WebGnome, an intelligent agent residing in this cyberspace environment which, with its continuous adaptive learning from mimicking the user’s behavior, will grow into a likeminded replica (MindClone) of a user-self acting in the Virtual Reality of the Internet with the synthetic mind capabilities of the AutoGnome.

Maximum Entropy (MaxEnt) Principle: MaxEnt is a technique for automatically acquiring probabilistic knowledge from incomplete information without making any unsubstantiated assumptions. Entropy is a mathematical measure of uncertainty or ignorance: greater entropy corresponds to greater ignorance. Hence, the MaxEnt solution is the least biased possible solution given whatever is experimentally known, but assuming nothing else.

Order/DisOrder/ReOrder Form: It is a tenet of Relational Systems that any semiotic act must, of necessity, express the Form of experience as the inseparable conjunction of:

• Ordered (i.e., determined or certain) experience: a formal algebra/logic of semiosis