Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

Simon Fraser University

$27,000.00

Semi-supervised learning for natural language processing

EF 2008-09

Résumé
Natural language is the main repository of human knowledge, and with the advent of large amounts of readily available machine-readable natural language text such as the web and newswire feeds, there has been a steady increase in the number of computational applications that attempt to extract information from these sources. The field of natural language processing now has a well-developed suite of approaches that use supervised machine learning. However, adapting to a new domain or a new language, especially a language with insufficient data, is still a major bottleneck when building software applications that process natural language. This proposal aims to address this crucial issue by exploring the use of semi-supervised machine learning methods for natural language, which can learn even when only a small amount of training data is available. Here are two examples of benefits from this research: First, by using domain-general information about English and then adapting to a specialized language, such as bio-medical research papers, natural language applications can be used to extract information in this new domain. Second, machine translation systems are trained on sentences previously translated by human translators, but such translations are only available in large quantities in a domain like parliament proceedings, e.g. the French-English Canadian Hansards. But this domain is not well suited to build a machine translation system for another domain, e.g. political weblogs. Our research enables translation systems to adapt to this new domain. Our research also enables machine translation in a domain where only small amounts of training data are available, e.g. the Nunavut Hansards.
Programme
Discovery Grants Program - Individual
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
262313-2008
Name nom
Sarkar, Anoop
Department département
Computing Science
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2008
Competition year année de concours
2008
Award amount
27000
Program id
RGPIN
Program nane en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
331
Committee name en
Computing and Information Sciences - B
Committee name fr
Informatique - B
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2801
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Natural language and speech understanding
Research subject fr
Langage naturel et reconnaissance de la parole
Partie
1
Nb partie
5
Application title
Semi-supervised learning for natural language processing
Keyword
NA
Application summary
Natural language is the main repository of human knowledge, and with the advent of large amounts of readily available machine-readable natural language text such as the web and newswire feeds, there has been a steady increase in the number of computational applications that attempt to extract information from these sources. The field of natural language processing now has a well-developed suite of approaches that use supervised machine learning. However, adapting to a new domain or a new language, especially a language with insufficient data, is still a major bottleneck when building software applications that process natural language. This proposal aims to address this crucial issue by exploring the use of semi-supervised machine learning methods for natural language, which can learn even when only a small amount of training data is available. Here are two examples of benefits from this research: First, by using domain-general information about English and then adapting to a specialized language, such as bio-medical research papers, natural language applications can be used to extract information in this new domain. Second, machine translation systems are trained on sentences previously translated by human translators, but such translations are only available in large quantities in a domain like parliament proceedings, e.g. the French-English Canadian Hansards. But this domain is not well suited to build a machine translation system for another domain, e.g. political weblogs. Our research enables translation systems to adapt to this new domain. Our research also enables machine translation in a domain where only small amounts of training data are available, e.g. the Nunavut Hansards.