Background
Technology Overview
Applications to Knowledge Management and Knowledge Discovery
Specialized WebMindä Products (Market Analytics and Enterprise Analysis)

WebMindä is a Java-based software system which evolves its own "digital intuition," and using this intuition, poses and answers questions regarding information. It deals with textual and numerical information on an equal footing, freely making generalizations spanning different types of data. Unlike any other AI system, it forms its own intuitions regarding concepts by dynamically "grounding" linguistic concepts in terms of its own nonlinguistic experience. It is the first AI system able to create its own concepts that apply across different contexts.

Derived from a sophisticated, complex-systems-based mathematical theory of intelligence, WebMindä 's architecture is that of a massively parallel network, a population of many different static and dynamic agents continually recomputing their relationships with other agents, and acting on other agents in accordance with these relationships

The initial release of WebMindä will be tuned for applications in information retrieval, data mining and knowledge management. Specialized versions of WebMindä for financial analytics and enterprise analysis will be released later. "Killer applications" include the use of textual information to help predict numerical series (financial markets, productivity of a division of a firm, etc.); and the ability to track the evolution of a concept over time based on contextual analysis of a broad pool of information.

WebMindä runs efficiently on powerful stand-alone computers, but is most powerful when run over a network of computers, in which case its sophisticated server-server communication methods allow its internal network structure to harmonize with the connectivity structure of the computer network.

WebMindä does not aim to emulate human intelligence in an artificial way, but rather to be a natural intelligence for the digital world -- solving human problems with digital intuition.

The discipline of Artificial Intelligence is an extremely ambitious one, defined in the Artificial Intelligence Dictionary as "a multi-disciplinary field encompassing computer science, neuroscience, philosophy, psychology, robotics, and linguistics; and devoted to the reproduction of the methods or results of human reasoning and brain activity." The underlying long-range goal of AI is nothing less than the construction of thinking machines, machines that think as well as human beings, or better. In spite of these lofty ambitions, however, the reality of AI has up to this point been severely disappointing.

This is where WebMindä comes in. WebMindä utilizes an original complex-systems model of intelligence, combined with the power of multithreaded, agents-based Java network computing, to overcome these historical limitations of AI technology. Whereas previous AI technologies focus on specific mental qualities such as knowledge representation, perception, reasoning, or learning, WebMindä is based on a comprehensive mathematical model of the mind as a whole. It perceives information, analyzes information, remembers information and acts on information in an integrated way. It adapts to its own environment much as human beings adapt to the physical world. It evolves its own spontaneous structures, rather than relying too heavily on the intuitions of human programmers.

WebMindä does not have a general intuitive understanding of our world, but it does have a general intuitive understanding of its own world, the data stored on the Internet, and the interests and responses of human Internet users. This is the qualitative difference between WebMindä and previous AI systems; and this is why WebMindä has the potential to bring AI beyond the role of a specialized adjunct to unintelligent computer systems, and to make AI the heart of computing itself as we move on into the new era of super-powerful digital networks.

The WebMindä software architecture is founded on 7 years of research in theoretical, cognitive science by IntelliGenesis founder Dr. Ben Goertzel published in 4 technical books and a series of research articles. The culmination of this research is a mathematical model of mind called the "Psynet model," which draws on the insights of cybernetic and complex systems science, and represents mind as a collection of emergent structures arising from a self-organizing physical system. WebMindä is an implementation of the Psynet model of mind, which is efficient in terms of contemporary computer architectures and effective in terms of contemporary business computing goals.

Underlying the Psynet model is a vision of intelligence as "the capability to achieve complex goals in complex environments." WebMindä 's complex goals are firstly posed to it by human users and secondarily evolved by itself; its complex environment consists of the human beings, machines and networks with which it interacts.

According to the Psynet model, the achievement of intelligence does not depend on the implementation of particular mechanisms, but rather on the implementation of mechanisms that give rise to appropriate emergent structures. The unique achievement of the WebMindä design is to cause these emergent structures to emerge from current computers and networks in a computationally efficient and humanly useful way.

Although WebMindä has some neural-nettish features, it is not a "neural network" model in the traditional sense. The constraint of reasonable efficiency on networks of serial processors requires numerous differences from the design of the human brain.

These structures are coded into WebMindä , but in a highly flexible way, so that each structure may co-adapt with the others, the details of each structure thus ultimately being emergent from the details of all the other structures.

WebMindä provides a practical software framework within which this vision of "the network is the computer is the mind" is made practical and useful, in a business context. The machines interconnected on an Intranet are leveraged together into a single thinking WebMindä unit, controlled by a "cortex" of one or more dedicated WebMindä machines. And machines interconnected by slower Internet connections are able to interact much like different human beings, exchanging questions and answers and chunks of relevant knowledge.

WebMindä has been designed using rigorous object-oriented methodology, and is being programmed in Java. The value of Java for WebMindä is multifold:

With the advent of ultra-efficient compilers such as IBM's Formula, the speed gap between C++ and Java is narrowing; our own tests show a ratio of 2.5:1. The WebMindä server will also be able to exploit the added power of the Java chip, to be introduced in early 1999. Finally, the much-publicized "Java browser wars" have no significant effect on WebMindä , as the WebMindä server is not browser-based, and the WebMindä client can be run either as an applet in a Java-1.2 compatible browser, or as a freestanding application.

A single WebMindä consists of a collection of one or more WebMindä servers, each one generally running on its own machine. The intelligent core of a WebMindä server is a collection of intelligent agents called the Psynet. A Psynet, in turn, consists of multiple pods, each pod containing a number of agents. A collection of WebMindä servers linked together as a single WebMindä may be said to have one collective Psynet, consisting of the Psynets of all the individual WebMindä servers acting as an emergent whole.

At the present time, the boundaries of a single Psynet are the boundaries of the intranet: Ethernet connections are fast enough that WebMindä servers thus connected can act as an emergent whole; whereas Internet connections are too slow. Thus, one has a model of the intranet as the WebMindä individual; and the Internet as a population of WebMindä s interacting in a manner best described as "social", exchanging information and questions and answers but not participating in collective trains of thought on a real-time basis. This of course may change as bandwidth increases over the next decade; at some point the Internet will become as fast as today's Ethernet, and a single global Psynet will become achievable.

Agents in WebMindä fall into several classes. First there are static agents or nodes, which have a relatively persistent state. Nodes may contain information relating to data such as texts or database files or internal information such as categories or other patterns recognized among other nodes. Nodes are not truly static, they update themselves continually, but they represent the backbone of WebMindä 's intelligent internal network. Most of the information contained in a node actually consists of a record of the node's relationship to other agents in the Psynet. There are many different types of node, pertaining to different types of information.

In addition to nodes, there are relational agents or links, which establish and retain relationships among nodes. These relationships may fall into many different categories, and may be parameterized by qualities such as intensity, uncertainty, and so forth. Finally, relational agents are formed by the third kind of agent, mobile agents which move around within the internal relational geometry of an individual Psynet, potentially from one WebMindä server to another, actively seeking new relationships to embody in relational agents. Mobile agents can also move from one WebMindä to another, across the Net, but this requires a different brokering mechanism than the free flow of agents within the same WebMindä server, or between different WebMindä servers on the same Intranet.

Every WebMindä server comes with a standard collection of pods: pods for standard data types such as text and numerical data, several pods for natural language understanding, a pod for categorization, a PsyContext pod for real-time information processing, a pod for query processing, and other pods devoted to such tasks as creative brainstorming, self-reflection, socialization, etc. Other pods will be application-specific: e.g., a WebMindä server aimed at financial applications would contain a pod of historical financial data; an enterprise application would contain a pod of transient e-mail data; etc. Additional pods may be loaded and saved by the user.

There is one Pod, the PsyContext pod that bears a special relation to the outside world. This pod is WebMindä 's short-term memory, containing information regarding what WebMindä has recently experienced, and other things judged to be associated to these recent experiences. WebMindä 's "lineal awareness" system consists of a loop that takes in new data, processes it in conjunction with the agents in the PsyContext, and then places it in the PsyContext, removing some obsolete information.

WebMindä does not have a separate "natural language system" as such. Language processing is carried out by the Lineal Awareness system, and the category nodes, in a way that is thoroughly integrated with the Psynet. To disentangle language from thought is to remove most of what is interesting and important about language.

The first part of WebMindä 's language understanding is linguistic categorization, which is carried out by category nodes according to the normal dynamics of the Psynet. WebMindä 's self-organizing categorization is capable of determining parts of speech for words, and of determining semantic categories such as "house", "action", etc. as well.

Once straightforward categorization has been used to create semantic and syntactic categories, attractor neural network methods are used to refine semantic links between words, based on contextual information. Words are then joined by relational agents if they are related closely in the bodies of information that WebMindä has been exposed to.

The genetic algorithm (GA) is used in several places within WebMindä , for the purpose of creating new agents satisfying dynamically-arising requirements. For this purpose, a highly generic approach to GA evolution of Java objects has been created. Any node that is to be evolved must implement the interface Evolvable, which contains three methods: mutate(), which mutates the object; cross(), which crosses the object with another Evolvable object; fitness(), which reports how fit the object is in its current situation.

The relation between the genetic algorithm and the brain is somewhat uncertain. Gerald Edelman's "Neural Darwinism" theory argues that the brain can be viewed as a genetic algorithm with mutation only; however, a case can also be made for crossover between "neuronal maps." At very least, we may view the use of the GA in WebMindä as a more efficient way of accomplishing, using crossover and mutation, what the brain may accomplish with mutation only.

In order for an intelligence to understand the meaning of a word, concept or other entity, this entity must be "grounded" in the system's own experience. Consider, for example, the position of WebMindä on encountering the word "two." In order to deal with "two" in an intelligent way, the system must somehow understand that "two" is not only a pattern in texts, it also refers to the two computer users who are currently logged onto it, the two kinds of data file that it reads in, etc. These are all patterns associated with "two" in various situations. Associating all these patterns with the "two" category node is what fully integrates the "two" node with the rest of the network, guaranteeing that the behavior of the "two" node is in harmony with its surroundings.

Symbol grounding is connected in a fascinating way with introspection. In WebMindä , each agent has the ability to look inside itself and report various aspects of its current state. Symbol grounding is achieved by special mobile agents attached to individual nodes representing concepts (e.g. the node for two"). These agents are evolved by the genetic algorithm; they survey the introspections of other nodes and combine the results of these introspections to form "groundings" of the node to which they are attached.

WebMindä 's query processing is integrated with the rest of its mental activity, just as for a human being, question-answering is not so different from purely internal thought processing.

The distinction between activity due to a particular query and activity due to general Psynet thought processes or other queries is carried out via an innovative, proprietary technique of "parallel thought processes," which allows WebMindä to do one thing the human mind cannot: carry out hundreds or thousands of simultaneous trains of thought, and keep them all straight!

An individual WebMindä server running in a single Java process is, potentially, an autonomous WebMindä . On a multiprocessor machine with a large amount of RAM, one can achieve quite a powerful WebMindä in this way. Given the current state of computer hardware, however, the most cost-effective way to achieve a very smart WebMindä will be to network a number of WebMindä servers together into a single multi-machine Psynet.

To understand the implications of these hardware restrictions, which push toward a network implementation, one must consider them in a broader context. They overlap with fundamental system-theoretic requirements in an interesting way. Hardware leads us to distinguish three levels of WebMindä system:

We use the term an "individual WebMindä " or simply "a WebMindä " to refer to an individual WebMindä intelligence; and the term "global WebMindä " or "WebMindä society" to refer to a collection of WebMindä intelligences. At the present time, the most natural mapping from social/individual concepts to computer hardware is

Social interactions between WebMindä s are as follows:

Intra-mental interactions encompass social interactions but add an additional components. For example, nodes within a single WebMindä (currently, a single intranet) will reach out to one another when activated and reinforce the relational agents intertwining them; nodes separated by a slower Internet connection will not.

Like the human mind, WebMindä has a great deal of emergent complexity behind the scenes; but this complexity is not visible to the outside observer. From the user's point of view, WebMindä functionality will be very simple -- or at least, simple in the sense that communicating with a human being is simple. There will be subtleties to be mastered, but these will involve getting an intuitive feel for how the system thinks, rather than mastering a complex notation or user interface.

The WebMindä client is installed on a user's machine; the user clicks on the WebMindä icon and the client window comes up. The user then selects one from among a number of query forms, fills out the blanks, and clicks "submit." For easy queries, an answer appears almost immediately, in verbal or graphical format. For harder queries, the WebMindä client gives a signal when the query is done, and the user goes to the Results menu to cause the results to appear on the screen. Very simple! The presupposition is that the client is talking to a WebMindä server, on the intranet or the Internet, that has understood a body of information relevant to the user's queries.

The distinction between a "WebMindä server" and a "WebMindä client" is crucial. A WebMindä server is a Java process embodying a self-organizing intelligent body of information. A WebMindä client is a Java process designed to interact over a network with one or more WebMindä servers. The client runs on any 32-bit machine; the server requires an OS supporting robust multithreading, such as NT or Unix. The WebMindä client is actually a simplified version of the WebMindä server, and is designed so as to be able to carry out limited self-organizing intelligent operations on local data when its host machine has been idle for a period of time.

In essence, the distinction between a WebMindä client and WebMindä server resides in the following qualities:

An individual "WebMindä " is a collection of WebMindä servers, networked together via high-bandwidth cable. Generally the core of a WebMindä will be one or more computers dedicated exclusively to WebMindä , with perhaps other systems dedicated partly to WebMindä on the periphery. The machines dedicated to WebMindä will be supplied with the maximal hardware affordable: e.g. 512 Meg of RAM, and ideally, multiple boards.

The absolute simplest application of WebMindä is to the increasingly onerous task of extracting useful information from the vast pools of online data. Web search engines represent the lowest level of information retrieval technology. More sophisticated products, aimed at intranets, document archives, etc., involve natural language technology and automatic statistical categorization. But even these systems are only minimally successful, because they lack more than a rudimentary understanding of context. WebMindä , operating from a large pool of nodes stored in the RAM of one or more powerful machines, uses this body of information to interpret queries in the context of everything it knows, and hence to provide on-target responses to requests for information -- something all the more necessary each month, as more and more information is put online.

Standard search engines, based on key word and Boolean search, fall badly short on both accounts: they miss a great number of relevant documents (poor recall), and they provide a great number of uninteresting documents (poor precision). More sophisticated applications developed by such firms as Semio, Perspecta, Inquizit, etc. provide improved recall and especially precision, by doing sophisticated statistical analysis of the sentences in texts. However, these techniques use relatively rigid analysis methods, which are optimized for particular types of information, and which take into account only certain types of contextual data. WebMindä 's much more flexible architecture allows the analysis of semantic content to vary based on the particular nature of the data in a particular application, and allows the incorporation of all types of contextual information in the response to a query. This is because it achieves intelligent information retrieval as a side-effect of general, adaptive, self-organizing intelligence, rather than as a consequence of highly specialized techniques for analysis of semantic expressions in text.

The comparison with alternate methods for document retrieval provides an excellent window into WebMindä 's approach to natural language. WebMindä does involve specialized methods for syntactic and semantic analysis of natural language expressions, presented in user queries and in texts, but these methods are integrated with the Psynet. There is a continual feedback between the specialized language parsing agents, and the cognitive processes of the Psynet, which results in semantic analysis that draws not only on the statistics of words and phrases in texts, but on the whole body of relationships that WebMindä has inferred between various types of data, including texts, database records, user preferences and responses, trends over time, etc. In short, WebMindä 's treatment of language, in information retrieval and otherwise, is based on the integration of language with the broad context of reality as processed by the Psynet, whereas the current crop of semantic-content-analyzing information retrieval tools are based on analyzing semantics as an insular, self-contained universe.

The effectiveness of responses to search queries relies on the meaningfulness of the relations already recognized by WebMindä and embodied in its structure; and this meaningfulness is ensured by WebMindä 's ongoing, autonomous cogitative activity. The Psynet may be thought of as database of nodes referring to documents, concepts, words, phrases and other information, but it is not static like an ordinary database; it is not a fixed fund of knowledge, which sits there passively waiting for queries to set it into motion. Instead, it is constantly refining its structure -- and its structure is, like the structure of the human mind/brain, inseparable from its complex dynamics. The concepts that it forms are not defined by mere statistical clustering, but by the outcome of self-organizing, reflexive dynamics.

From a software point of view, "data mining" consists of recognizing patterns within individual data tables, and emergent among various data tables. WebMindä has something unique to contribute to both of these aspects of data mining, but its originality here is primarily focused on the emergent level. A particular talent of WebMindä 's is its ability to recognize patterns emergent among data and text, "text-numerical data mining."

Standard data mining packages primarily provide functions falling into two categories: categorization/clustering, and prediction. WebMindä provides both of these functions, and others besides. Clustering and categorization are performed by the generic category nodes mentioned above, applicable equally to nodes referring to numerical data or to text. Prediction, on the other hand, is performed by a collection of "hard-wired" prediction routines, notably polynomial prediction and simplex-based prediction; and by a proprietary method for expressing standard nonlinear time series analysis algorithms as interactions of associative and hierarchical relational agents.

WebMindä 's data mining capabilities may be accessed via natural language queries, or via OLAP-style graphical interface. Like standard OLAP applications, WebMindä will provide the interactive ability to query multi-dimensional databases in real-time, relating the data mined from databases with its internal store of information and providing analyses that take into account context as well as data-insular statistical and mathematical patterns.

"Information retrieval" and "data mining" are concepts that were created to describe the behavior of software much more limited in its behavioral repertoire than WebMindä . WebMindä , with its ability to understand context and to ground concepts in its own experience, also has the ability to do much more than any mere information-extraction tool. It is able to serve as a true knowledge management and knowledge creation system, integrating information from the document archives and databases of interest to a business, and allowing exploration of this information in all its facets, each piece considered in the context of the whole.

If WebMindä is installed on a company's intranet, then real-time queries regarding relationships between textual, numerical and other data to do with the enterprise may be posed by any employee with computer access at any time. The result is that WebMindä 's intelligence is integrated with the social intelligence of the organization, and the individual intelligence of the employees. Each time an employee accesses a document, he may ask WebMindä for related documents, and WebMindä will carry out this task with an understanding of the role of that employee in the company, that employee's particular needs and interests, etc. It will be able to make its own creative suggestions, based on its autonomous thought processes.

Furthermore, the social dynamics of the different WebMindä servers residing in different parts of the company's intranet (governed by WebMindä 's highly innovative, proprietary server-server communication methods) will grow to reflect the social dynamics of the individuals using those parts of the intranet. Each WebMindä server will respond most effectively and rapidly to queries involving information which it stores locally; but the information that a certain server stores locally may change over time, depending on user needs and internal Psynet dynamics. Thus, while providing easy access by all users to all information at all times, WebMindä will nevertheless nudge the information at the readiest disposal of individual humans and divisions in certain directions, based on its inferences and its own emergent understanding. WebMindä will do more than just provide an understanding of structures and processes; it will be a participant in processes, in the formation of emergent human and informational structures.

And, as various WebMindä units in various organizations exchange non-proprietary information, in the interest of increased mutual intelligence, WebMindä will be a participant in the formation of human and informational structures on the global scale. This is an exciting new vision of artificial intelligence, in the business context and beyond -- not AI as something separate from humanity, providing us with answers to our questions, but AI as something interacting symbiotically with humanity, participating in our communications, goals, and social structures and processes.

WebMindä is extremely well suited for the particular data mining tasks encountered in the world of finance. It has the capability to deal with derivatives and other more esoteric financial instruments with unprecedented fluidity; and it has an ability seen in no previous financial analysis software, namely the ability to integrate textual and numerical information to make financial decisions.

A special version of WebMindä , trained to effectively enact its intelligence in the context of real-time financial data, and enhanced with a specialized finance-oriented user interface, will be marketed as WebMindä Market Analytics.

The two most important unique aspects of WebMindä Market Analytics are its adaptive trading system, and its scheme for portfolio optimization, both of which are based on GA evolution of specialized agents in the context of the Psynet.

WebMindä 's trading system is based on several years of research and successful trading by IntelliGenesis co-founder Dr. Jeff Pressing. Trading is initially assumed to be done by applying a "trading rule" to a time series at a given point in time. The rule is a GA-evolved agent, combining the outputs of various predictors and indicators.

Next, the technique for using WebMindä for portfolio optimization is quite simple. Given a collection of interlinked nodes representing financial time series, one is faced with the following challenge: find a set of such nodes which maximizes the following criteria:

The WebMindä trading system is flexible enough that it can integrate any kind of predictor or financial indicator -- including those that are not purely numerical in scope. In one of WebMindä 's most innovative applications, textual analysis is used to generate new "textual indicators", which are then utilized in prediction in the same way as numerical indicators and predictors. The method used to create textual indicators is subtle and proprietary, involving GA evolution of specialized types of nodes.

Using its extremely flexible structures and dynamics, WebMindä can easily be trained to perform a variety of interesting "enterprise analysis" functions. A special version of WebMindä , trained on such applications and provided with specific tools for visualizing information flow through organizations, will be marketed as WebMindä Enterprise.

For example, WebMindä Enterprise will be able to:

Achieving these enterprise functionalities will not involve adding any new core structures to WebMindä , but only creating new types of query nodes, and tuning the categorization, visualization and concept-evolution processes to work on the type and volume of data found in enterprise applications.