OntologyBased FreeForm Query Processing for the Semantic Web

1
Ontology-Based Free-Form Query Processing for
the Semantic Web

• Thesis proposal
• by
• Mark Vickers

2
The Problem

• Searching the web for an answer to a question is
  hard.
• Returns documents (usually too many)
• Can it instead return just the right answers?
• Semantic web
• Proposed ontology-based framework for making
  information machine-readable
• Better access to information
• How should semantic web be searched?

3
Solution AskOntos a Query System for the
Semantic Web

• Allows free-form queries (unconstrained natural
  language queries)
• Semantically interprets queries using information
  extraction via extraction ontologies
• Returns database-like tables of extracted values

4
Thesis Statement

• We can build a system that will show how
  extraction ontologies can be used to execute
  conjunctive, free-form queries over semantically
  annotated web pages.

5
Extraction Ontologies
Object sets Relationship sets Participation
constraints Lexical Non-lexical Primary object
set Aggregation Generalization/Specialization
6
Extraction Ontologies
Data Frame
Internal Representation float
Value Phrase
Value Expression \s\s(\d1,3)(\.\d2)?
Left Context
Key Word Phrase
Key Word Expression (Pprice)(Ccost)
Operation Phrase
Operator gt
Expression (more\sthan)(more\scostly)
7
AskOntos Overview
Extraction Ontology Repository
Query
AskOntos
Ontology Matching
Extracted Values
Extracted Values
Extracted Values
Extracted Query Values
Extracted Values
Form XQuery
Answer
Web Pages
8
Step 1. Parse Query
Find me the and of all
s I want a

price
mileage
red
Nissan
1998
or newer
gt Operator
9
Step 2. Find Corresponding Ontology
Find me the price and mileage of all red Nissans
I want a 1998 or newer
gt Operator
Similarity value 6
Similarity value 2
10
Step 3. Formulate XQuery Expression

• Conjunctive queries run over selected
  ontologys extracted values

11
Step 3. Formulate XQuery Expression

• Value-phrase-matching words determine conditions
• Conditions
• Color red
• Make Nissan
• Year gt 1998

gt Operator
12
Step 3. Formulate XQuery Expression
1 for doc in document("file///c/ontos/owlLib
/Car.OWL")/rdfRDF 2 for Record in
doc/owlThing 3 4 let id
substring-after(xsstring(Record/_at_rdfabout),
"CarIns") 5 let Color doc/carColor_at_rdfID
concat("ColorIns", id)/carColorValue/text() 6
let Make doc/carMake_at_rdfIDconcat("MakeI
ns", id)/carMakeValue/text() 7 let Year
doc/carYear_at_rdfIDconcat("YearIns",
id)/carYearValue/text() 8 let Price
doc/carPrice_at_rdfIDconcat("PriceIns",
id)/carPriceValue/text() 9 let Mileage
doc/carMileage_at_rdfIDconcat("MileageIns",
id)/carMileageValue/text() 10 11
where(Color"red" or empty(Color)) and 12
(Make"Nissan" or empty(Make)) and 13
(Yeargt"1998" or empty(Year)) 14 return
ltRecord ID"id"gt 15 ltPricegtPricelt/Pric
egt 16 ltMileagegtMileagelt/Mileagegt 17
ltColorgtColorlt/Colorgt 18
ltMakegtMakelt/Makegt 19 ltYeargtYearlt/Yeargt
20 lt/Recordgt
For each owlThing
Get the instance ID and extracted values
Check conditions
Return values
13
Step 4. Run XQuery Expression Over
Ontologys Extracted Data

• Uses Qexo 1.7, GNUs XQuery engine for Java
• Use XSLT to transform results to HTML table

14
Evaluation of AskOntos

• Measure success by
• Ability to match query to correct ontology
• Ability to translate free-form queries into
  formal queries
• We will create
• Extraction ontologies for car ads, diamonds,
  ski resorts, real estate, theatre schedules
• Queries
• 20 Conjunctive queries for each domain
• Queries gathered from research group members

15
Query Translation Metrics
Find me the price and mileage of all red Nissans
I want a 1998 or newer.
Human conversion
PROJECT Price, Mileage,Color, Make,
Year SELECT (Color,,red),
(Make,,Nissan),
(Year,gt,1998)
16
Query Translation Metrics
Find me the price and mileage of all red Nissans
I want a 1998 or newer.
for doc in document("file///.../Car.OWL")/rd
fRDF for Record in doc/owlThing
where(Color"red" or empty(Color)) and
(Make"Nissan" or empty(Make)) and
(Year"1998" or empty(Year)) return ltRecord
ID"id"gt ltPricegtPricelt/Pricegt
ltColorgtColorlt/Colorgt
ltMakegtMakelt/Makegt
ltYeargtYearlt/Yeargt lt/Recordgt
Human conversion
PROJECT Price, Mileage,Color, Make,
Year SELECT (Color,,red),
(Make,,Nissan),
(Year,gt,1998)
Automated conversion
PROJECT Price,Color, Make, Year SELECT
(Color,,red),
(Make,,Nissan),
(Year,,1998)
17
Conclusion/Contributions

• AskOntos
• Is a free-form query system for the semantic web
• Applies information extraction for query
  processing
• Answers questions with extracted data values
• Contributions
• Web queries that use semantic annotations
• Web queries returning answers from extracted data
• Processing free-form queries using ontologies

18
END
19
TREC 2004 QA Question Topics
20
Related Research
21
Evaluating the Framework

• Input
• Obituaries ontology
• 25 obituaries from two newspapers

Four of eighteen object sets shown above. Data
from Salt Lake Tribune and Arizona Daily Star
22
Scaling to the Web

• Ontologies crawl and harvest web pages
• Ontologies extract values from pages
• Ontologies indexed (??)
• Queries extracted by relevant ontologies
• Rely on Google-like technology