Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Toronto

$46,000.00

Developing a Universal Model of Word Comprehension and Reading

FY 2024-25

Summary
The long-term aim of my research is to develop a "universal" neural network model of language processing that will explain a range of abilities in diverse languages. My short-term objectives advance this aim and form important new theoretical contributions in their own right. The main method I use is building computer simulations of how networks of neurons in the brain represent different aspects of language knowledge (e.g., representations of letters, speech sounds) and interact with one another. My modeling work is complemented by coordinated behavioural experiments and recording electrical activity while the brain processes words. Objective 1 develops a unified theory of learning, representation, and generalization in neural network models. This goal is advanced by studying how neural networks learn, represent, and generalize spelling-sound patterns from newly-learned words (e.g., a new word like gint probably rhymes with the regular pronunciation in mint and lint, not pint). Objective 2 develops a neural network model of how words are processed in different languages. It examines how neural networks adapt to different statistical regularities present in different languages to explain cross-linguistic differences and provide a more unified and inclusive account of diverse language abilities. Objective 3 develops an integrated model of how the lexical system and a response system work together to produce different patterns of effects in different experimental tasks. It also aims to settle a long-standing debate regarding how ambiguous words (e.g., chicken denotes an animal or its meat) are processed. Impact. The proposed work will shed light on core theoretical issues in the language sciences related to new word learning in first and second languages, how language processes differ as a function of which language an individual knows, and how the meanings of ambiguous words, which are ubiquitous in language, are resolved. These insights have clear potential applications such as improving language instruction. Given our multicultural society, effective language training is critical for communicating with one another, and improved language instruction would enhance first and second language learning outcomes and optimize the use of the $4.5 billion invested by provincial education ministries in language education. Similarly, it can inform our understanding of language impairments (e.g., dyslexia, aphasia), which are expected to grow more prevalent in our aging population, as well as related remediations (e.g., Brain-Computer Interfaces). Models that are validated in several languages will allow for more inclusive application their insights in our linguistically-diverse society. Understanding ambiguous word processing will provide insights into how humans make decisions and resolve ambiguity that can inform the development of AI models---one of the fastest growing and influential areas of the technology sector.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2024-05588
Name nom
Armstrong, Blair
Department département
Department of Psychology
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
46000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1502
Committee name en
Biological Systems and Functions
Committee name fr
Systèmes et fonctions biologiques
Area of application code
1210
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Psychology
Area of application fr
Psychologie
Research subject code
6306
Research subject group en
Psychology
Research subject group fr
Psychologie
Research subject en
Cognitive science -- language
Research subject fr
Science cognitive -- langage
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Developing a Universal Model of Word Comprehension and Reading
Keywords
neural network models of cognition; ambiguous word comprehension; new word learning and generalization; cross-linguistic differences; distributional semantics; corpus analyses; statistical learning; event-related potentials (erp); word recognition; lexical processing
Application summary
The long-term aim of my research is to develop a "universal" neural network model of language processing that will explain a range of abilities in diverse languages. My short-term objectives advance this aim and form important new theoretical contributions in their own right. The main method I use is building computer simulations of how networks of neurons in the brain represent different aspects of language knowledge (e.g., representations of letters, speech sounds) and interact with one another. My modeling work is complemented by coordinated behavioural experiments and recording electrical activity while the brain processes words. Objective 1 develops a unified theory of learning, representation, and generalization in neural network models. This goal is advanced by studying how neural networks learn, represent, and generalize spelling-sound patterns from newly-learned words (e.g., a new word like gint probably rhymes with the regular pronunciation in mint and lint, not pint). Objective 2 develops a neural network model of how words are processed in different languages. It examines how neural networks adapt to different statistical regularities present in different languages to explain cross-linguistic differences and provide a more unified and inclusive account of diverse language abilities. Objective 3 develops an integrated model of how the lexical system and a response system work together to produce different patterns of effects in different experimental tasks. It also aims to settle a long-standing debate regarding how ambiguous words (e.g., chicken denotes an animal or its meat) are processed. Impact. The proposed work will shed light on core theoretical issues in the language sciences related to new word learning in first and second languages, how language processes differ as a function of which language an individual knows, and how the meanings of ambiguous words, which are ubiquitous in language, are resolved. These insights have clear potential applications such as improving language instruction. Given our multicultural society, effective language training is critical for communicating with one another, and improved language instruction would enhance first and second language learning outcomes and optimize the use of the $4.5 billion invested by provincial education ministries in language education. Similarly, it can inform our understanding of language impairments (e.g., dyslexia, aphasia), which are expected to grow more prevalent in our aging population, as well as related remediations (e.g., Brain-Computer Interfaces). Models that are validated in several languages will allow for more inclusive application their insights in our linguistically-diverse society. Understanding ambiguous word processing will provide insights into how humans make decisions and resolve ambiguity that can inform the development of AI models---one of the fastest growing and influential areas of the technology sector.