Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

McGill University

$25,000.00

Sequential decision-making under model uncertainty

EF 2010-11

Résumé
As we begin to integrate intelligent autonomous systems in challenging real-world applications, we are confronted with a major scientific challenge: to build intelligent systems that are robust enough to handle the uncertain and dynamic nature of real-world task domains without requiring exact models of these domains. Some attention has been given to the question of uncertainty in action effect and sensor data, but most existing techniques require a well-defined parameterized representation of the domain. The long-term goal of this project is the deployment of intelligent autonomous agents in complex real-world domains, including inherently uncertain systems. In support of this, we propose to investigate new models and algorithms for addressing the problem of sequential decision-making under model uncertainty. The first objective of this research is to develop new methods for learning good state represntations directly from data, with a particular attention to domains where the input data is high-dimensional, or the state of the system depends on a long sequence of prevous inputs. The second objective is to improve current methods for joint parameter and value estimation in Partially Observable Markov Decision Processes. The third objective is to develop effective methods for dealing with domains where the objective function is unknown or complex. The research outlined in this proposal falls within a larger research context, which integrates two complementary objectives: the development of representations and algorithms for decision-making, and their application to the problem of sequential decision-making in health-care domains. An important aspect of this proposal is that any findings arising from the theoretical investigations will be validated in the context of complex medical and robotic decision-making problems.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
311949-2008
Name nom
Pineau, Joelle
Department département
Computer Science, School of
Organization id
61
Institution établissement
McGill University
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2010
Competition year année de concours
2008
Award amount
25000
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
2805
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Learning and inference theories
Research subject fr
Théories de l'apprentissage et l'inférence
Partie
3
Nb partie
5
Application title
Sequential decision-making under model uncertainty
Keyword
NSERC
Application summary
As we begin to integrate intelligent autonomous systems in challenging real-world applications, we are confronted with a major scientific challenge: to build intelligent systems that are robust enough to handle the uncertain and dynamic nature of real-world task domains without requiring exact models of these domains. Some attention has been given to the question of uncertainty in action effect and sensor data, but most existing techniques require a well-defined parameterized representation of the domain. The long-term goal of this project is the deployment of intelligent autonomous agents in complex real-world domains, including inherently uncertain systems. In support of this, we propose to investigate new models and algorithms for addressing the problem of sequential decision-making under model uncertainty. The first objective of this research is to develop new methods for learning good state represntations directly from data, with a particular attention to domains where the input data is high-dimensional, or the state of the system depends on a long sequence of prevous inputs. The second objective is to improve current methods for joint parameter and value estimation in Partially Observable Markov Decision Processes. The third objective is to develop effective methods for dealing with domains where the objective function is unknown or complex. The research outlined in this proposal falls within a larger research context, which integrates two complementary objectives: the development of representations and algorithms for decision-making, and their application to the problem of sequential decision-making in health-care domains. An important aspect of this proposal is that any findings arising from the theoretical investigations will be validated in the context of complex medical and robotic decision-making problems.