History edit
Looking at previous generations of Artificial Intelligence languages, there are two features that are clearly missing: easy readability; a solid epistemological foundation for definitions and context.
The readability problem is clearly evident in the Semantic Web environment, and is typified by the contrast between XML[1] and N3[2] (aka Notation 3). When writing formal documents, XML is the usual choice. But when people are working together informally, and striving for ease of understanding, N3 is a frequent choice.
The epistemology problem is not as clearly evident, but it has been a thorn in the side of Artificial Intelligence researchers for decades. Terry Winograd, one of the successful early researchers in Natural Language processing, said[3]:
Language is a process of communication between people, and is inextricably enmeshed in the knowledge that those people have about the world. That knowledge is not a neat collection of definitions and axioms, complete, concise and consistent. Rather it is a collection of concepts designed to manipulate ideas. It is in fact incomplete, highly redundant, and often inconsistent. There is no self-contained set of "primitives" from which everything else can be defined. Definitions are circular, with the meaning of each concept depending on the other concepts.
Conceptual Graphs[4][5] appeared to provide a good foundation for definitions. However, visual graphs are just too simple; a written language is needed to express the complexities of the real world. A foundation for context was even more elusive; researchers could not agree on a definition of context:[6]
..the fields of knowledge representation and natural language.. In both fields, one observes a huge spectrum of answers to an important question in the technical agenda: "What is context?" This broad range of answers reflects both the confusion about context and the enormous difficulties in handling it.
The mKR language[7] combines a restricted natural language, Simple English, with a strong epistemology[8] to provide readability, definitions and context.
In 2002, the developers of mKR and RDF compared the two languages in a W3C email forum[9]. This forum produced a better understanding of both languages, but did not lead to any significant changes in either language. The developers later compared the mKR and OWL languages; this time a significant change was made in the OWL language. Property Restrictions were added to emulate the genus-differentia definitions of the mKR language.
At the suggestion of the RDF/OWL developers[10][11], a practical mKR language interface was developed for the Stanford University TAP knowledge base and the OpenCyc knowledge base. A simple mKR language interface was also developed for Amazon.com and Google. mKE (my Knowledge Explorer) was enhanced to read RDF files[12].
The most recent changes in mKE (my Knowledge Explorer) provide command-line options to initialize the knowledge base with concepts from a language chosen by the user. Language options include RDF, OWL, OpenCyc, TAP, Amazon, Google.
References edit
- ^ 199x, W3C, XML standard
- ^ 199x, Tim Berners-Lee, Notation 3
- ^ 1972, Terry Winograd, Understanding Natural Language, Academic Press, page 26
- ^ 1975, Roger C. Schank, Conceptual Information Processing, North Holland Publishing Company and American Elsevier Publishing Company, Inc.
- ^ 1991, John F. Sowa, Editor, Principles of Semantic Networks, Morgan Kaufmann Publishers, Inc.
- ^ 1998, Lucja Iwanska, Context in Knowledge Representation and Natural Language, The 1997 Fall Symposia Reports, AI Magazine, Volume 19, Number 2, Summer 1998, page 125.
- ^ 1997, Richard H. McCullough, Knowledge Explorer, The Icon Newsletter No. 52, page 6.
- ^ 1990, Ayn Rand, Introduction to Objectivist Epistemology, Expanded Second Edition, Meridian.
- ^ http://lists.w3.org/Archives/Public/www-rdf-interest/
- ^ 2002, Danny Ayers, RE: KR & W3C (was KR & Issue/bug tracking terms in RDFS?), 21 Dec 2002.
- ^ 2002, Seth Russell, RE: CycL vs. KR, 29 Nov 2002.
- ^ David Beckett, Resource Description Framework (RDF) Resource Guide, RDF Editors and Tools, McCullough Knowledge Explorer.