Ontology-based information retrieval of scientific information Natalia V. Loukachevitch Laboratory of Information Resources Analysis Research Computing.

Ontology-based information retrieval of scientific information

Natalia V. LoukachevitchLaboratory of Information Resources Analysis

Research Computing Center of Moscow State University (MGU NIVC)

Thematic Search of Scientific Information

• Knowledge-based (ontology-based) search

• Use of synonyms

• Automatic query expansion

• Automatic analysis of query results

• Help in interactive search

Sociopolitical domain

Le

ve

ls o

f h

iera

rch

y

LawAccounting

Taxes

Banking

Bilingual Sociopolitical Thesaurus The thesaurus development is based on three methodologies:

• methods of construction of information-retrieval thesauri (information-retrieval context, analysis of terminology, terminology-based concepts, a small set of relation types)

• development of wordnets for various languages (word-based concepts,

detailed sets of synonyms, description of ambiguous text expressions)

• ontology and formal ontology research (strictness of relations description, necessity of many-step inference)

(33,000 concepts, 80,000 Russian terms, 85,000 English terms)

General Lexicon

Specific Lexicon

Специальная лексика

Socio-Political Domain vs. General Lexicon and Specific

Lexicons

Intermediate Zone

Information

Security

Aviation Ontology

Cul

tura

l H

erita

ge

Ontology on Natural

Sciences

and Technology

30,000 concepts; 70,000

terms

Thematic Structure tax; taxation system; tax payer;

finances; economy; tax legislation; VAT

legislation; law; draft law;

Taxation Code;

deputy minister; Ministry of Finance;

finances; reform; tax reform

populationbudget, estimate;

finances; economy; document

government; state power; Minister of

Finance

State Duma; state power;

state

Thematic representation of a text:Thematic Node i||+ == Thematic Node j

Thematic node in the text

University Information System RUSSIA(http://www.cir.ru, http://uisrussia.msu.ru )

- Database of Fulltext Documents (1,5 mln): Legal Acts, Newspaper articles, Scientific Reports

- Database “Statistics of Russian Federation” (Socio-economic Statistics, Demographic Statistics, Agrarian Statistics, Urban Statistics)

- Database “Budget system of Russian Federation”) (www.budgetrf.ru)

Visualisation of Data in Dynamic Tables and Maps

Convertors Processing Interfaces Services

Unified Technology Platform (Constructor) Russian University Social Sciences

I nformation and Analytical consortiumwww.cir.ru

www.echr- base.ru

БД Статистика Россииwww.budgetrf.ru

Cross-Language Information

Retrieval

Applications of technology

• Concept-based information retrieval (monolingual, bilingual)

• Information-Retrieval systems combining word-based and concept-based serach

• Concept-based automatic text categorization

• Automatic Question-Answering

• Automatic Text Summarization

Main Projects

State Duma of RF (1999 - …) Central Election Commission of RF (1997 - …) Legal Company “Garant” (2002 – …) Ministry of Education (2005-2006) Accounting Chamber of RF (2003 – …) Central Bank of RF (2006 – …) Grants:

– McArthur Foundation (1994, 1995, 2004 - …)

– Ford Foundation (2002, 2003)

– Russian Foundation for Basic Research (9)

– Russian Foundation for Humanitarian (5)

– Eurasia Foundation (2002, 2003)

Participance in International Forums

• Participance in Text REtrieval Conference TREC

organized by NIST DARPA (TREC-6, TREC-8)

• Participance in Summarizarion Conference SUMMAC

organized by NIST DARPA (1st place)

• Cross-Language Evaluation Forum CLEF

(DELOS program)

– paricipance in Steering Committee

– provision of Russian collections for evaluation purposes

– information retrieval of domain-specific information

retrieval

• Organizers of Russian Information Retrieval Evaluation

Seminar ROMIP (www.romip.ru/en/ )