Join us at this week's meeting of the Philosophy Club virtually or in-person (Auerbach 320) this Wednesday, Sept. 30 from 1 p.m.—2 p.m. for our next University of Hartford Philosophy Club Meeting this Academic Year as longtime club member Dr. Clark Sexton presents “Knowledge Base Requirements for Non-Monotonic Reasoning in AI Systems."
TOPIC OVERVIEW
Knowledge base requirements are the essential features, data structures, and functional needs required to build a centralized, searchable library of information. Monotonic reasoning is a logical process in which adding new information to a knowledge base never invalidates or cancels out any previously derived conclusions This is the kind of reasoning typified in deductive or axiomatic systems, in which by design subsequent data or line entries cannot contradict prior ones. It is a sort of bottom-up reasoning in which the initial “data base” - very often a set of axioms - is never contradicted or challenged as it is utilized to deduce all subsequent lines.
In contrast, non-monotonic reasoning does not have this restriction and therefore has to be able to cope with the eventuality of subsequent data contradicting prior data or prior conclusions from that data.
While much of we have used computers for, especially in the earlier days, has in fact been monotonic in nature, in recent years, our turn toward newer developments in Artificial Intelligence have led decidedly and perhaps even definitively in the direction of non-monotonic reasoning. This, however, presents us with new and serious logical troubles, which are explored and expanded on in this presentation.
Non-Monotonic Reasoning is an essential feature of a truly intelligent AI System. In this presentation I will demonstrate a fundamental requirement of the Knowledge Base component of an intelligent system (whether natural or artificial). This, it turns out, will require "meta-information" about the information stored in the Knowledge Base, that will inform the decisions necessary to determine which propositions should be revised/retracted when new information would result in an inconsistent set of propositions in the Knowledge Base. Full document attached.
ABOUT CLARK
Clark Sexton of Fort Hays State University in Hays, Kansas, earned doctorates both in Computer Science from Kansas State University and in Philosophy from the University of Kansas. His research in Computer Science was in Artificial Intelligence, and, more specifically, Natural Language Processing. For this research, he implemented an NLP system that could parse a wide range of syntactic structures of English, perform type-checking to determine whether a sentence is meaningful, and disambiguate certain ambiguous expressions.
Clark continued his exploration of the relations of meanings in his dissertation in Philosophy, in which he presented a brief history of the analytic/synthetic distinction, replied to Quine's objections, and provided and presented arguments for his own account of the distinction.
Questions? Contact Brian Skelly at bskelly@hartford.edu or 413.273.2273.
An ongoing weekly tradition at the University since 2001, the University of Hartford Philosophy Club is a place where students, professors, and people from the community at large meet as peers. Sometimes presentations are given, followed by discussion. Other times, topics are hashed out by the whole group.
Presenters may be students, professors, or people from the community. Anyone can offer to present a topic. The mode of presentation may be as formal or informal as the presenter chooses.
Come live or join online! Invite friends. Suggest topics and activities. Take over the club! It belongs to you!