Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of Waterloo

$62,000.00

Supporting Cooperation in Multiagent Systems

EF 2024-25

Résumé
Problems of cooperation, where agents seek ways to improve their welfare jointly, are ubiquitous and range from mundane daily tasks such as scheduling meetings or navigating traffic to global challenges such as pandemic preparedness or climate change negotiations. Since AI systems play an ever more significant role in our world, it is critically important to equip them with the capabilities necessary for cooperation. The emerging subfield of "Cooperative AI" encompasses a board research program "trying to help individuals, humans and machines, to find ways to improve their joint welfare".1 My long-term research goal is to advance Cooperative AI, particularly in domains with heterogeneous agents (e.g. AI agents and humans) where cooperation is hard due to complex and conflicting incentives. Progress will require research across multiple axes including improving AI systems' understanding of the world, other agents, and their incentives, and designing or learning institutions that provide the social structure required to promote cooperation. To this end, I propose a research program that will address three key objectives. First, I will investigate the role teamwork in promoting system-wide cooperation. Recent research has shown that small teams of learning agents can learn sophisticated policies, improving system-wide outcomes significantly. I aim to provide full characterizations of when teams of learning agents are beneficial or not, the role of self-interest in teamwork, and effective team-formation algorithms and processes. Second, for AI agents to cooperate effectively careful thought must be taken as to how their underlying utility functions are shaped and how rewards are shared across the system. I will develop principled methods for designing and sharing rewards across teams of agents, be they reinforcement learning agents, humans, or a combination. Finally, social infrastructure is required to support cooperation in general. I will investigate novel algorithms and methods for supporting institution design using reinforcement learning, computational game theory, and multiagent systems. Progress in this research will provide principled ways of building and reasoning about cooperative AI systems, which have the potential to impact the daily lives of people across Canada. Furthermore, this research program will train numerous students in different aspects of artificial intelligence, providing them with the skills that are, and will continue to be, in high demand across different jobs and industries. 1Dafoe, A., Bachrach, Y., Hadfield, G., Horvitz, E., Larson, K. and Graepel, T., 2021. Cooperative AI: machines must learn to find common ground. Nature, 593(7857), pp.33-36.
Programme
Discovery Grants Program - Individual
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2024-04027
Name nom
Larson, Kate
Department département
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
62000
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
2800
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Artificial intelligence (computer vision, use2603)
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
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
Supporting Cooperation in Multiagent Systems
Keywords
multiagent systems; cooperative ai; game theory; social choice
Application summary
Problems of cooperation, where agents seek ways to improve their welfare jointly, are ubiquitous and range from mundane daily tasks such as scheduling meetings or navigating traffic to global challenges such as pandemic preparedness or climate change negotiations. Since AI systems play an ever more significant role in our world, it is critically important to equip them with the capabilities necessary for cooperation. The emerging subfield of "Cooperative AI" encompasses a board research program "trying to help individuals, humans and machines, to find ways to improve their joint welfare".1 My long-term research goal is to advance Cooperative AI, particularly in domains with heterogeneous agents (e.g. AI agents and humans) where cooperation is hard due to complex and conflicting incentives. Progress will require research across multiple axes including improving AI systems' understanding of the world, other agents, and their incentives, and designing or learning institutions that provide the social structure required to promote cooperation. To this end, I propose a research program that will address three key objectives. First, I will investigate the role teamwork in promoting system-wide cooperation. Recent research has shown that small teams of learning agents can learn sophisticated policies, improving system-wide outcomes significantly. I aim to provide full characterizations of when teams of learning agents are beneficial or not, the role of self-interest in teamwork, and effective team-formation algorithms and processes. Second, for AI agents to cooperate effectively careful thought must be taken as to how their underlying utility functions are shaped and how rewards are shared across the system. I will develop principled methods for designing and sharing rewards across teams of agents, be they reinforcement learning agents, humans, or a combination. Finally, social infrastructure is required to support cooperation in general. I will investigate novel algorithms and methods for supporting institution design using reinforcement learning, computational game theory, and multiagent systems. Progress in this research will provide principled ways of building and reasoning about cooperative AI systems, which have the potential to impact the daily lives of people across Canada. Furthermore, this research program will train numerous students in different aspects of artificial intelligence, providing them with the skills that are, and will continue to be, in high demand across different jobs and industries. 1Dafoe, A., Bachrach, Y., Hadfield, G., Horvitz, E., Larson, K. and Graepel, T., 2021. Cooperative AI: machines must learn to find common ground. Nature, 593(7857), pp.33-36.