Generative Methodology thread

September 8, 2005

Bead Threads sometimes merge and separate

ß the “Taos Discussion” . this thread on generative methodology à

Note from Dennis (Knowledge Foundations Inc)

Paul,

Thank you for your response. I am not sure if Dr. Ballard has time to respond due to his programming commitment and preparation for our up-coming presentation, so I will answer your questions from my business/technical perspective and allow him to follow-up with his science/physics perspective.

First, our way of thinking is based on Dick's definition of knowledge: Knowledge = Theory + Information. The human brain has not changed for at least 200,000 years. In terms of knowledge, the only difference between modern man and our ancestors are the number and complexity of theories we have in our brains. Theories are the rules, constraints and conditional relationships between our concepts, ideas and thought patterns that give them meaning. Theory represents 85% of knowledge content.

Theory answers “How?” “Why?” and “What if?” questions and is predictive because once learned, it allows us to "know" what the outcomes and consequences of situations and circumstances will be. We learn theory through enculturation, lessons-learned life experience and education. It is a priori - known before the fact.

Information represents anything that exists in time and space that can be processed by the senses, measured or counted. Information represents only 15% of knowledge content, and is the “Who?” “What?” “When?” “Where?” and “How Much?” facts that we learn after a situation or circumstance occurs - it is a posteriori.

Applications and databases leave 85% of knowledge content on the table, the theory, requiring costly human labor to make sense of their contents. They do this because their schema architectures fail when attempting to capture > n-3 relationships. Conventional database and application development tools are designed to inhibit programmer creativity beyond n-2. Semantic Web people (W3C) have achieved n-3 (triples) capacity.

What is interesting is that most of the social theories we use to understand our world today originated about 40,000 years ago. Estimations are that city organization and money matters developed 7,000 to 10,000 years ago. Concepts such as "buy low, sell high" or "interest on principal" are examples. These theories have not changed for centuries and are used today with equal results. Examples of modern theories might include the "wire tap" or wireless communication. Chester Gould, the author of the Dick Tracy comic strip conceived of the wire tap in 1948/49 and published the idea in his comic strip. The next year the FBI embraced the theory of wire taps as a law enforcement practice. Gould also conceived of the wireless wrist radio. This idea, or more exactly, the theory behind wireless communication, became an essential part of the Star Trek series. in 1977 the first cell phone was tested in Chicago.

By the time the cell phone came to market, most people in the modern world already had the theory of wireless communication in their brains, and their acceptance of the product was rapid. For the most part, theory evolves and changes slowly. Information changes constantly.

Coming from this perspective then, here are my answers to your thoughtful questions.

Your Question: "I do not know if it will capture all of the intent of the author - in those cases where the author was not very good at expressing his/her thoughts."

A: All meta-physical and physical knowledge can be captured, but the most valuable knowledge is based on well-justified theories that have been proven through empirical science or experience - axiomatic. We fully expect that homeopathic medicine, astrology, compiled advice columns, gardening, real estate salesmanship, philosophy and all academic subjects will be captured and distributed commercially.

I see these knowledgebases as a first step towards a world where people become power-learners because knowledgebases indirectly teach theory as they are used - whether for work or entertainment. There will be knowledgebase reviewers, knowledgebase of the month clubs and so forth. The deeply ingrained theories in our minds guarantees that we all know more than what we realize. Well engineered knowledgebases make explicit intent and nuance depending on desired granularity.

Your Question: "I have another concern, and that is one where there is not only one book but many books, and many viewpoints. How does the Mark II or Mark III manage inconstancies? Is it assumed that there is only one viewpoint, ultimately; and therefore this one "correct" viewpoint is the benchmark for all consistent "knowledge".

A: This is a key question. Some knowledge domains tend to be highly defined, such as aircraft maintenance which requires well defined requirements. In this case, every relevant perspective, language and jargon, mission and strategic objective, quality standard, policy, procedure, requirement, specification, function, task, timeline, time-history, cost, part, attribute, supplier, supplier part-history, terms of engagement and so forth are clearly defined by specification and procedure manuals, policy statements and other reference sources.

Less structured knowledge domains, such as those of corporate and government management pose a slightly different challenge. In this case, CEO's, CFO's, CIO's, CKO's, human resources, engineering, marketing, sales, service and legal perspectives that are constrained by missions, goals, policies, values, products/services, strategies, employees and so forth must be captured to develop a viable knowledgebase. Typically, executives and managers run "snapshot" reports and attend on-going meetings to determine the strategies and decisions that are most meaningful to the organization. The more complex the organization, the more reports and meetings.

Another type of knowledge domain is represented by a free flowing academic, national or international debate that supports a continuity of divergent viewpoints. In this case, effort to persuade is supported by research, reference or ideological reasoning.

These knowledge domain examples can be captured using Mark III (due to market in 2006), as long as the books, manuals, documents, databases, subject experts and other relevant knowledge content sources can be aggregated and modeled. The only limitations are time, money and available knowledge resources. Knowledge Engineers do not invent anything, their only job is to capture the knowledge that exists, or if their are knowledge gaps, to find the relevant knowledge necessary to fill-in the gaps. Every viewpoint and perspective is taken into consideration, every relevant relationship interconnected.

As users access the knowledgebase to answer questions, they key-in concepts and ideas into a special browser (K-Browser) that are central to their question. The browser takes them directly into the semantic web and placed them on an interactive reasoning path. As they move forward along the reasoning path, they gain "how" answers, as they move backward along the path, they learn "why" answers and as they move in multiple directions from the path, they learn "what if" alternatives. Because every viewpoint is taken into consideration, the user can explore their question from every viewpoint. Given this, a CEO, CFO, CIO, etc. can gain the same, or more precise answers from a knowledgebase than what they gain from meetings as long as all the knowledge of the organization is in the knowledgebase. Whether by machines or humans, users gain most strategic, mission trade-off, legally defensible or best practice answers to even their most complex questions within seconds and minutes.

Your Question: "There is a second aspect of a deep problem, the first aspect being this issue of a single universal truth. This second aspect has to do with "knowing" something that is just incorrect. What does one do with a computer that "knows" a body of knowledge that is incorrect. For that matter, what does society do with an economic/political system that has become "incorrect" if the nation is to have a democratic form of governance?

A: This is a deep question that has veracity, ethical, legal and moral implications that will be explored from every perspective as the knowledge society evolves. Our experience has been that conducting a focus group to define what questions the knowledgebase must answer is the first step. if the completed knowledgebase can answer those questions, then the knowledge engineering job is successful, if not, then the causes might include gaps in the knowledge content or a poorly constructed knowledgebase. (see case study #2 http://www.knowledgefoundations.com/Publications.html)

Your Question: "Perhaps it is I who knows something incorrectly. In any case, the book that I would write, that could be 100% encoded into a Mark III knowledge processor; would be what? A correct representation of an incorrect view of the world? Somehow I do not feel confident with a notion of a Mark III knowing what I know. I sense that you and Richard will have something to say about this, as would Tom Adi, John Sowa and others.

A: I do not see you writing your book on Mark III at this point either, though it could be captured and distributed as a knowledgebase - depending on its content. As the Mark III technology catches on, we expect first users will target the low hanging fruit - government and business.

OK Paul, I am hopeful I answered your questions with some level of success. I am leaving for vacation this weekend and will be back on the 20th. I look forward to hearing more about your book and your efforts in Taos.

Dennis