Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Waterloo

$41,000.00

Towards General-Purpose Machine Pragmatic Reasoning

FY 2024-25

Summary
Humans are remarkably efficient in reasoning about the communication intents of others and generating appropriate utterances in response. This is usually achieved through pragmatic reasoning, which allows humans to infer the underlying meanings of utterances beyond their literal meanings. For example, the utterance "I have been waiting for a million years" should almost always be interpreted as a complaint about a long wait in real life, instead of a literal statement about the actual waiting time. My long-term goal is to empower machines with the ability to communicate with people as efficiently and effectively as humans do---a good model should understand not only the literal meanings carried by words, but also the underlying meanings inferred from contexts. These models should possess the ability to understand context across various modalities, such as text, images, and videos, mirroring the multifaceted way humans perceive information. On the one hand, the main body of natural language processing (NLP) research has been focused on understanding the literal meanings of utterances, which is necessary but insufficient for my long-term goal. On the other hand, existing work on machine pragmatic reasoning has been largely limited to synthetic settings, pure-text environments or task-specific scenarios, which are not enough to capture the complexity of human communication. My research experience in multimodal NLP, syntax, and semantics (literal meanings) has prepared me well for this research. To achieve my long-term goal, I will work on addressing the following short-term objectives in the emerging field of machine pragmatic reasoning: O1.1: Constructing an English multimodal pragmatic reasoning benchmark. O1.2: Extending the benchmark to multiple languages, with awareness of underlying cultural differences. O2: Developing pragmatic listener models that interpret the communication goal behind the utterances. O3: Developing pragmatic speaker models that generate utterances given a certain context. This research will have five significant impacts: 1. Scientific and technical innovation in machine pragmatic reasoning will be disseminated in leading NLP and machine learning venues. 2. The benchmark developed will be open-sourced to facilitate future research. 3. Findings will be applied to create more human-like chatbots to enhance customer experiences in various application scenarios, from education to healthcare, benefiting individuals globally.  4. Students will gain high-quality hands-on training in data collection, analysis, model design and implementation, and communication under the guidance of the PI. 5. The emphasis on multilingual NLP will enhance the deployment of cutting-edge technologies for underrepresented languages, such as the Indigenous languages of Canada, and reduce bias in NLP models through increased cultural awareness.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2024-04395
Name nom
Shi, Haoyue
Department département
David R. Cheriton School of Computer Science
Organization id
33
Institution établissement
University of Waterloo
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
41000
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
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
1207
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Information, computer and communication technologies
Area of application fr
Technologie de l'information, des ordinateurs et des réseaux de communications
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
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
Towards General-Purpose Machine Pragmatic Reasoning
Keywords
natural language processing; machine learning; pragmatics; language grounding; reasoning; language models; multilingualism; multi-modal nlp
Application summary
Humans are remarkably efficient in reasoning about the communication intents of others and generating appropriate utterances in response. This is usually achieved through pragmatic reasoning, which allows humans to infer the underlying meanings of utterances beyond their literal meanings. For example, the utterance "I have been waiting for a million years" should almost always be interpreted as a complaint about a long wait in real life, instead of a literal statement about the actual waiting time. My long-term goal is to empower machines with the ability to communicate with people as efficiently and effectively as humans do---a good model should understand not only the literal meanings carried by words, but also the underlying meanings inferred from contexts. These models should possess the ability to understand context across various modalities, such as text, images, and videos, mirroring the multifaceted way humans perceive information. On the one hand, the main body of natural language processing (NLP) research has been focused on understanding the literal meanings of utterances, which is necessary but insufficient for my long-term goal. On the other hand, existing work on machine pragmatic reasoning has been largely limited to synthetic settings, pure-text environments or task-specific scenarios, which are not enough to capture the complexity of human communication. My research experience in multimodal NLP, syntax, and semantics (literal meanings) has prepared me well for this research. To achieve my long-term goal, I will work on addressing the following short-term objectives in the emerging field of machine pragmatic reasoning: O1.1: Constructing an English multimodal pragmatic reasoning benchmark. O1.2: Extending the benchmark to multiple languages, with awareness of underlying cultural differences. O2: Developing pragmatic listener models that interpret the communication goal behind the utterances. O3: Developing pragmatic speaker models that generate utterances given a certain context. This research will have five significant impacts: 1. Scientific and technical innovation in machine pragmatic reasoning will be disseminated in leading NLP and machine learning venues. 2. The benchmark developed will be open-sourced to facilitate future research. 3. Findings will be applied to create more human-like chatbots to enhance customer experiences in various application scenarios, from education to healthcare, benefiting individuals globally.  4. Students will gain high-quality hands-on training in data collection, analysis, model design and implementation, and communication under the guidance of the PI. 5. The emphasis on multilingual NLP will enhance the deployment of cutting-edge technologies for underrepresented languages, such as the Indigenous languages of Canada, and reduce bias in NLP models through increased cultural awareness.