Information Retrieval and Databases: Synergies and Syntheses

1
Information Retrieval and Databases Synergies
and Syntheses

• Peter Clark
• Mathematics and Computing Technology
• Boeing Phantom Works
• The Boeing Company

2
Information Retrieval and Databases Synergies
and Syntheses

• Boeing faces huge challenges with mixed data
• parts, dimensions, locations, costs, schedules,
• procedures, functions, protocols, design
  principles,
• Documentation for a 747
• would not fit inside the aircraft
• even if it could, the aircraft would be too heavy
  to fly
• (and, every 747 is different!)

3
Key Challenges in Integrating IR and DB Approaches
Premise

• Synergy requires explicit, shared semantics

Structure chasm ? Semantics chasm?
Challenges

• Extracting the semantics of unstructured data
• Encoding the semantics of structured data
• Defining useful, shared ontologies
• Machine inference
• Leveraging common-sense knowledge

4
Integrating IR and DB approaches Two Snapshots
within Boeing

• ESDS
• An expert system for standards knowledge
• ? deductive database canned text paragraphs
• Synthesizes custom page on the fly
• Resource Locator
• Texts tagged with concepts (from a Thesaurus)
• Retrieval based on semantic distance

5
Summary
Structure Chasm
Semantics Chasm