It's Like Déjà Vu All Over Again
"You could probably waste an entire day on the preceding links alone. But why take chances? We also give you Paul Snively..." — John Wiseman, lemonodor
OpenCyc.org. Via Simon :OpenCyc is the open source version of the Cyc technology, the world's largest and most complete general knowledge base and commonsense reasoning engine. Cycorp, the builders of Cyc, have set up an independent organization, OpenCyc.org, to disseminate and administer OpenCyc, and have committed to a pipeline through which all current and future Cyc technology will flow into ResearchCyc (available for R&D in academia and industry) and then OpenCyc.
- I guess this is AI stuff!?! And it's applicable to the Web and XML because it could help structure the web. [Roland Tanglao's Weblog]
Commercial Cyc is the result of a multi-decade program to implement common-sense reasoning for computers, so it absolutely falls within the definition of classic AI. The principal researcher is Doug Lenat, a highly-regarded master in the field. It's interesting to note that an early collaborator with Dr. Lenat was R. V. Guha, who later went to Apple, where he created MCF or the Meta-Content Framework, and still later went to Netscape, where he created RDF. So RDF very much has its roots in the knowledge representation wing of the AI community.
But Cyc has a more expressive language, CycL, than RDF is, and an extremely sophisticated inference engine customized and optimized for CycL. Cyc's biggest contribution to the rest of the world, however, may be its Upper Ontology: the set of concepts and relations from which all other concepts and relations derive, for the practical purpose of common-sense human reasoning. The upshot is that if you can successfully categorize whatever specific concepts and relations according to Cyc's upper ontology, you can expect an inference engine using the ontology to deliver high-quality results within the constraints imposed by the engine itself.
This is indeed potentially of enormous significance to the Semantic Web, and for that matter, to mankind in general. We've never before had, in relatively portable and widespread form, the groundwork upon which to attempt to build automated rationality. Every prior attempt was either excruciatingly limited in scope (dealing with toy "block world" domains) or utterly proprietary and closed, and typically both. The Cyc project marks a "back to the future" milestone: a return to the traditional AI ideal of implementing all of human common-sense knowledge and reasoning ability, but the first project to actually release anything approaching such breadth and depth.
Now all we need are builds of the inferencing and editing tools for non-Intel-Linux platforms... in the meantime, expect the activity around CycL and the Upper Ontology to heat up.
7:54:53 PM Google It!
