a Exploiting Thesaurus-Based Semantic Net for Knowledge-Based Search
Peter Clark John Thompson Lisbeth Duncan Heather Holmback Knowledge Systems Boeing, Mathematics and Computing Technology
Overview
Problem: searching for information in particular, for human experts Approach: Search using concepts, not words Use a thesaurus as the initial ontology Enhance it using simple AI techniques The Application: Two deployed "Expert Locator" applications
Overall Picture
"tube placement"
Query words
Search Engine
Web pages
Databases
Human Experts
Document repositories
...
Problems with word searches..
Words have many senses (polysemy) e.g. "plane" finds both airplanes and geometry Many words mean the same thing (synonymy) e.g. "tail fin" misses "vertical stabilizer" Lack of world knowledge e.g. "jet engine" misses "propulsion systems" Goal: organize search around concepts, not words Need a conceptual vocabulary ("ontology")
The Ontology Bottleneck
Massive up-front cost to build an ontology
The Approach
Use a technical thesaurus, enhanced with AI techniques Boeing's Thesaurus: Highly customized to aerospace and Boeing Massive knowledge repository
37,000 concepts, 18,000 synonyms 100,000 relationships (3 types)
Many person-years investment of effort
A (tiny) fragment of the ontology...
ignition Pneumatic equipment starting engines Engine starters Flame propagation Flame stability Combustion stability Burning rate Jet engines flameout combustion afterburning Jet spray spray Ramjet engines Hydrogen fuels Turbojet engines lift thrust Propulsion systems
Converting Words to Concepts
Search word: "jet"
ignition Pneumatic equipment starting engines Engine starters Flame propagation Flame stability Combustion stability Burning rate Jet engines flameout combustion afterburning spray Turbojet engines
?
lift
thrust Propulsion systems
? ?
Ramjet engines Hydrogen fuels
?
Jet spray
Matching Query and Target Concepts
Semantic distance between "ignition" and "jet engines"?
ignition Pneumatic equipment starting engines Engine starters Flame propagation Flame stability Combustion stability Burning rate Jet engines flameout combustion afterburning Jet spray spray Ramjet engines Hydrogen fuels Turbojet engines lift thrust Propulsion systems
Expert Locator Demo
(see end of this presentation for the demo in powerpoint form)
Enhancing the Thesaurus: 1. Increase connectivity using subsumption
100,000 links are not enough! 40% of concepts are "orphans" But: Many concept names are phrases Can add links by analyzing these phrases
Space Shuttle Main Engine generalization related-to Engine Space Shuttle
Subsumption Computation Algorithm
1. Compute all possible generalizations by "word chopping" and "word generalization"...
Engine Vehicle Engine Vehicle Main Space Space Engine Vehicle Main Engine Space Shuttle Engine Space Vehicle Main Engine Shuttle Space Vehicle Vehicle
Space Shuttle Main Engine
Space Shuttle Shuttle
Subsumption Main
Space Computation Algorithm
2. Identify existing Thesaurus concepts and links within these
Engine Vehicle Engine Vehicle Main Space Space Engine Vehicle Main Engine Space Shuttle Engine Space Vehicle Main Engine Shuttle Space Vehicle Vehicle
Space Shuttle Main Engine
Space Shuttle Main
Space Shuttle
Subsumption Computation Algorithm
3. Add missing connections to nearest existing concepts
Engine Vehicle Engine Vehicle Main Space Space Engine Vehicle Main Engine Space Shuttle Engine Space Vehicle Main Engine Shuttle Space Vehicle Vehicle
Space Shuttle Main Engine
Space Shuttle Main
Space Shuttle
Some Example Inferred Links
Measuring Instruments Optical Measuring Instruments Equipment Distance Measuring Equipment Range Finders Optical Range Finders Halogen Compounds Fourine Compounds Nitrogen Fourine Compounds Fourides
Nitrogen Flourides
21,000 generalization/specialization and 37,000 related-to links added Number of "orphans" down from 40% to 13%
Enhancing the Thesaurus: 2. Use NLP to refine the "related-to" links
Current: New: Metal Tube Metal Tube related to made-of Metal Metal
27 relationship types chosen (causes, location, ...) heuristic noun-noun rules selects relationship, e.g
For compound "X Y" (e.g. "metal tube"): IF X is a Material AND Y is a Physical-Object THEN Y made-of X
Can use relation type to help compute semantic distance
Definition: "Flap: A movable airfoil attached to an airplane's wing, and used to increase lift or drag."
NLP Flap isa: Airfoil attribute: Movable attached-to: Wing part-of: Airplane purpose: Increase object: Lift, Drag Movable attribute
purpose isa bt rt attached-to
Enhancing the Thesaurus: 3. Knowledge from Text
Airfoil
Increase Drag
Flap
object purpose
Airplane
part-of
Increase
object
Wing
Lift...