Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Toronto

$50,000.00

Nuances of meaning, paraphrase, and argument identification in applications of natural language processing

FY 2012-13

Summary
Our research is in natural language processing and its applications. Two themes that run through the work are, first, the problems that are created by paraphrase and nuances of meaning in everyday language, and, second, computational notions of semantic distance between words, between larger linguistic expressions, and between documents. In the new research proposed here, we will work on the following problems: (1) Near-synonymy within and across languages: In our previous work, we developed representations of the nuances of meaning that distinguish near-synonyms, methods for automatically creating a lexical knowledge base from a dictionary of near-synonyms, and mechanisms for choosing the most appropriate near-synonym in the automatic generation of a sentence. Our next steps in this work will be to enlarge the knowledge base that we derived from near-synonym dictionaries (whose vocabulary is not sufficiently comprehensive) and to develop cross-lingual lexical knowledge bases for use in machine translation. (2) Opinion identification, extraction, and analysis: Our research will develop methods for the identification and analysis of opinions, positions, and the arguments and ideologies that underlie them, for the analysis of political, legal, and other position-based texts. The long-term goal of this work is to be able to take a text, classify its position on a topic or topics, extract the structure of the argument that it makes in support of its position, and then be able to use this in document retrieval and summarization. (3) Detecting the influence of one publication on another: It is often useful to trace the flow and influence of ideas and arguments from one person to another and from one publication to another. One particular application of opinion identification and analysis is to find documents that present the same or very similar opinions, although possibly in different words and possibly from a different viewpoint. (An instance of this is not necessarily plagiarism, though it might be.) We will develop new methods of paraphrase detection at a level sufficiently abstract that we can find such similarities.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
201-2009
Name nom
Hirst, Graeme
Department département
Computer and Mathematical Sciences (Scarborough)
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2009
Award amount
50000
Program id
RGPIN
Program name 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
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
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
4
Nb partie
5
Application title
Nuances of meaning, paraphrase, and argument identification in applications of natural language processing
Keyword
NSERC
Application summary
Our research is in natural language processing and its applications. Two themes that run through the work are, first, the problems that are created by paraphrase and nuances of meaning in everyday language, and, second, computational notions of semantic distance between words, between larger linguistic expressions, and between documents. In the new research proposed here, we will work on the following problems: (1) Near-synonymy within and across languages: In our previous work, we developed representations of the nuances of meaning that distinguish near-synonyms, methods for automatically creating a lexical knowledge base from a dictionary of near-synonyms, and mechanisms for choosing the most appropriate near-synonym in the automatic generation of a sentence. Our next steps in this work will be to enlarge the knowledge base that we derived from near-synonym dictionaries (whose vocabulary is not sufficiently comprehensive) and to develop cross-lingual lexical knowledge bases for use in machine translation. (2) Opinion identification, extraction, and analysis: Our research will develop methods for the identification and analysis of opinions, positions, and the arguments and ideologies that underlie them, for the analysis of political, legal, and other position-based texts. The long-term goal of this work is to be able to take a text, classify its position on a topic or topics, extract the structure of the argument that it makes in support of its position, and then be able to use this in document retrieval and summarization. (3) Detecting the influence of one publication on another: It is often useful to trace the flow and influence of ideas and arguments from one person to another and from one publication to another. One particular application of opinion identification and analysis is to find documents that present the same or very similar opinions, although possibly in different words and possibly from a different viewpoint. (An instance of this is not necessarily plagiarism, though it might be.) We will develop new methods of paraphrase detection at a level sufficiently abstract that we can find such similarities.