IBM Research Knowledge Structures Group Hawthorne, NY 10532 Phone: (914) 784-7055 Fax: (914) 784-6912 Email: welty@watson.ibm.com URL: http://www.research.ibm.com/people/w/welty/
Chris Welty is a Research Scientist at the IBM T.J. Watson Research Center in New York. Previously, he taught Computer Science at Vassar College, taught at and received his Ph.D. from Rensselaer Polytechnic Institute, and accumulated over 14 years of teaching experience before moving to industrial research. His principal area of research is Knowledge Representation, specifically ontologies and the semantic web, and he spends most of his time applying this technology to Information Retrieval and, in the past, Software Engineering. He serves on the steering committees of the Formal Ontology in Information Systems and Automated Software Engineering Conferences, as program chair of the Knowledge Representation Conference, on the editorial boards of AI Magazine and The Journal of Web Semantics, and is an editor in the W3C Web Ontology Working Group. His work on ontologies and ontology methodology has appeared in CACM, and numerous other publications.
Dr. Welty also was the first ACM SIGART Information Director, and was responsible for establishing the SIGART Electronic Information Service. He was the editor in chief of the former ACM SIGART Intelligence Magazine, from its initial planning in 1996 through to its ultimate demise in 2001, and has been an active ACM member since 1990.
Ontology is a word that dates back to the 17th century, and describes a practice that dates back to Aristotle, yet many computer scientists consider it a new field. This talk covers what Ontology was, what it has become, and some of the important problems in computer science, such as information integration and the semantic web, that it addresses.
This is a general talk suitable for undergraduates in Computer Science, as well as general audiences with interest in the semantic web.
The Computer Science field of Ontology has been progressing in a scatter-shot manner, and there has been little attempt to formalize the techniques used and lessons learned from research. OntoClean was one of the first attempts to build a formal foundation for ontological analysis, which helped identify the most common logical modeling mistakes and the justifications for why they are, in fact, mistakes. In this talk, I present an overview of the OntoClean methodology for ontological analysis, and highlight the common modeling problems with examples.
This talk requires some background in first-order logic, and some knowledge of database design or AI is often helpful. It is suitable for advanced undergraduates in Computer Science as well as Information Systems students with a database or AI background.
The OntoClean methodology has received much attention and use in research and practice. In this talk, I present the formal foundation of OntoClean, explaining advanced notions such as identity criteria, rigidity, and unity. The talk presents formalizations using modal and temporal logics, and the problems we have encountered with these formalizations.
This talk requires considerable background in logic and at least some exposure to modal and temporal representations. It is suitable for graudate students and researchers in AI.
The highest level in the so-called "layer cake" of the semantic web is OWL, the Web Ontology Language. The expressiveness of OWL is based on a cross between RDF triples and Description Logics. This talk covers the history of OWL, including description logics and RDF, and then goes through the OWL language itself with examples of how it is intended to be used.
This talk is suitable for a general semantic web audience, however some understanding of logic and knowledge representation is helpful.
Information Extraction techniques focus on shallow linguistic processing to recognize mentions of, e.g., names of people, places, organizations, etc., in text. Inherent in this processing has been the belief that this extracted "knowledge" can be used in structured systems such as databases and, more recently, knowledge-bases. In our research, we have found there to be a significant divide between the semantics of extracted data and the ontologies that knowledge based systems typically rely on, which leads to problems that are greatly exacerbated by reasoning. In this talk, I discuss the nature of this divide, the kinds of problems it causes, and how we are attempting to "cross" it.
This talk deals with advanced issues in Information Extraction and Knowledge Representation. It is suitable for graduate students and researchers in these areas.