Department of Computer Science & Engineering
University of California, San Diego

CSE-291: Ontologies in Data Integration, Spring 2003

Logistics

Overview

Ontologies play an increasing role in data and information integration, as evidenced by new workshops, special tracks at conferences, new journals, and activities such as the "Semantic Web".

Within computer science, ontologies have traditionally been studied in the context of AI and logic-based knowledge representation. The database community has also (re-)discovered this topic and is now increasingly active in this area. Application areas range from the business world (Enterprise Information Integration) to information integration for scientific data (e.g., Gene Ontology, Unified Medical Language System, ...). The latter is the focus of this course. In particular, we will study ontologies from different perspectives, and will address questions such as:

The seminar will include introductory presentations by the instructor, possibly guest lectures by other faculty, and presentations by students based on the literature or practical exercises.

More specifically, we will

Students will have the choice between "theory" and "practice" studies, where the former is typically a presentation of one or more selected articles, and the latter is a "modeling experiment" that involves applying an ontology tool or formalism to a specific domain (e.g., ecology, geosciences, biology, ...) or problem.

Grading

Grading will be primarily based on the quality of the report and presentation that the student will prepare and give during one of the last meetings of the class. The presentation will be based on one or more papers, and may include a system demonstration (for the ``hands-on'' topics).

The difference in the number of units is reflected in the level of effort: e.g., for one unit, a typical presentation and report will cover 1-2 papers (for a theory topic), or a modeling exercise with a single tool (for hands-on topics). For four units, a comparative study with 3-4 papers is typical (or a very detailed analysis of 1-2 papers), or a comparative modeling exercise with several tools (for hands-on topics).

Presentations will be ca. 45 minutes (independent of units taken). Comparative studies can be conducted in teams of two students and presented jointly (30-45 minutes each student).

Topics and References

Schedule

  1. April 4: Course overview, data/information integration, mediator architecture, what is an ontology?, examples, ... [slides.ppt]
  2. April 11: Ontologies (cont'd), example ontologies, introduction to topic maps, first-order logic primer, introduction to description logic [slides.ppt]
  3. April 18 Guest lectures by ...
  4. April 25
  5. May 2 Introduction to description logics, and reasoning in FO: tableaux calculus examples [slides.ppt]
  6. May 9 Tableaux calculus II, reasoning about concepts with LeanTAP, definitorial cycles [slides.ppt]
  7. May 16 (9am-10am!):
  8. May 23: Student presentations
  9. May 30: Student presentations
  10. June 6: Student presentations

Bertram Ludaescher
Last modified: Tue Jun 17 13:24:21 PDT 2003