Grant: Discovery Grants Program - Individual
University of Toronto
$81,000.00
The AI Symbol-Signal Nexus: Next Generation Human-Compatible Sequential Decision Making
EF 2024-25
Résumé
Striking advances in machine and deep learning are transforming how we build AI systems, including many sequential decision-making systems - from conversational agents and personal assistants, to logistics planners, recommender systems, personal and industrial robots, and more generally to computer programs that automate a myriad of tasks. The synthesis of automated decision-making systems from data, in whole or in part, presents fundamental challenges to how we construct safe, predictable systems that can interact seamlessly with each other and with humans, leverage human knowhow, reflect human norms and preferences, and that are taskable and understandable to people who may not know how to program.
This research program explores the AI symbol-signal nexus in the context of sequential decision making. It will develop mathematical principles, computational models, and algorithms for plan, program, and controller synthesis with a focus on building decision-making systems that are human compatible. Importantly, it will advance our understanding of how to marry conceptual ideas and methods from symbolic AI and from machine and reinforcement learning to build effective AI sequential decision-making systems. Our algorithms will reflect the highly optimized combinatorial and long-horizon reasoning of symbolic planning and search, the synthesis from language specification and correct-by-construction properties of program synthesis, and the generalization afforded by structured knowledge representation. They will do so both in the context of established methods for sequential decision making, with and without data, as well as in the context of recent advances in reinforcement learning as supervised learning using transformers, and generative AI. Our techniques will be applied to problems ranging from personal assistive and household robots to customized personal assistants and Internet of Things.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2024-06913
Name nom
McIlraith, Sheila
Department département
Computer Science
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
81000
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
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
2803
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Knowledge representation
Research subject fr
Répresentation des connaissances
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
The AI Symbol-Signal Nexus: Next Generation Human-Compatible Sequential Decision Making
Keywords
artificial intelligence; knowledge representation and reasoning; automated planning; reinforcement learning and machine learning; human-compatible ai; formal languages and automata; abstraction
Application summary
Striking advances in machine and deep learning are transforming how we build AI systems, including many sequential decision-making systems - from conversational agents and personal assistants, to logistics planners, recommender systems, personal and industrial robots, and more generally to computer programs that automate a myriad of tasks. The synthesis of automated decision-making systems from data, in whole or in part, presents fundamental challenges to how we construct safe, predictable systems that can interact seamlessly with each other and with humans, leverage human knowhow, reflect human norms and preferences, and that are taskable and understandable to people who may not know how to program.
This research program explores the AI symbol-signal nexus in the context of sequential decision making. It will develop mathematical principles, computational models, and algorithms for plan, program, and controller synthesis with a focus on building decision-making systems that are human compatible. Importantly, it will advance our understanding of how to marry conceptual ideas and methods from symbolic AI and from machine and reinforcement learning to build effective AI sequential decision-making systems. Our algorithms will reflect the highly optimized combinatorial and long-horizon reasoning of symbolic planning and search, the synthesis from language specification and correct-by-construction properties of program synthesis, and the generalization afforded by structured knowledge representation. They will do so both in the context of established methods for sequential decision making, with and without data, as well as in the context of recent advances in reinforcement learning as supervised learning using transformers, and generative AI. Our techniques will be applied to problems ranging from personal assistive and household robots to customized personal assistants and Internet of Things.