Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Western Ontario

$29,000.00

Active cost-sensitive learning

FY 2009-10

Summary
Both "active learning" and "cost-sensitive learning" are fast-growing and promising research areas in machine learning. I propose to further my research in these two areas, and to combine the two to foster a new learning paradigm called "Active Cost-sensitive Learning" (or ACL for short). Traditional machine learning is "passive"; that is, the learning program passively learns from the provided training examples. Active learning, on the other hand, actively seeks for extra information (the effort) to improve learning (the gain). Most previous works of active learning treat the "effort" and "gain" in different terms and do not unify the two. On the other hand, ACL (Active Cost-sensitive Learning) unifies the two aspects of active learning (effort and gain) using cost; that is, we study active learning in the cost-sensitive framework. In the proposed research we will study active learning algorithms that are able to recognize the opportunity to "buy" extra information in order to minimize the total cost (or maximize the net benefit). The extra information can include new examples, new class labels, and values of missing attributes in the either training or test datasets. The advantage of studying ACL is that it aims to optimize the business goals of real-world applications. The ultimate goal of such applications is to maximize the net profit or to minimize the total cost. Thus, ACL can be applied to solve a wide range of real-world problems. For example, in medical diagnosis our learning program can actively select a small set of crucial medical tests to reduce the total cost of medical tests and cost of misdiagnosis of patients, thus relieve the financial strain in the medical insurance system, such as OHIP (Ontario Health Insurance Plan). An ACL system of image classification can discover the optimal number of training examples (images) to "buy" or collect to reduce the total cost. Applied to real-world applications, ACL allows Canadian businesses, companies, and society's social services to save money. The impact of the proposed research is therefore very significant.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
46392-2007
Name nom
Ling, Charles
Department département
Computer Science
Organization id
36
Institution établissement
University of Western Ontario
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2009
Competition year année de concours
2007
Award amount
29000
Program id
RGPIN
Program nane 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
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
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
Active cost-sensitive learning
Keyword
NA
Application summary
Both "active learning" and "cost-sensitive learning" are fast-growing and promising research areas in machine learning. I propose to further my research in these two areas, and to combine the two to foster a new learning paradigm called "Active Cost-sensitive Learning" (or ACL for short). Traditional machine learning is "passive"; that is, the learning program passively learns from the provided training examples. Active learning, on the other hand, actively seeks for extra information (the effort) to improve learning (the gain). Most previous works of active learning treat the "effort" and "gain" in different terms and do not unify the two. On the other hand, ACL (Active Cost-sensitive Learning) unifies the two aspects of active learning (effort and gain) using cost; that is, we study active learning in the cost-sensitive framework. In the proposed research we will study active learning algorithms that are able to recognize the opportunity to "buy" extra information in order to minimize the total cost (or maximize the net benefit). The extra information can include new examples, new class labels, and values of missing attributes in the either training or test datasets. The advantage of studying ACL is that it aims to optimize the business goals of real-world applications. The ultimate goal of such applications is to maximize the net profit or to minimize the total cost. Thus, ACL can be applied to solve a wide range of real-world problems. For example, in medical diagnosis our learning program can actively select a small set of crucial medical tests to reduce the total cost of medical tests and cost of misdiagnosis of patients, thus relieve the financial strain in the medical insurance system, such as OHIP (Ontario Health Insurance Plan). An ACL system of image classification can discover the optimal number of training examples (images) to "buy" or collect to reduce the total cost. Applied to real-world applications, ACL allows Canadian businesses, companies, and society's social services to save money. The impact of the proposed research is therefore very significant.
University of Western Ontario - $29,000 Grant (2009-1… | Canada Spends