Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

McMaster University

$28,000.00

HUman partnership with bioMedical Artificial iNtelligence (HUMAN)

EF 2023-24

Résumé
Human partnership with Artificial Intelligence is fundamentally flawed because current systems are not designed with a team context or structure. In the biomedical systems domain where decision support can be critical, potentially life and death, it is vital to better balance this partnership. In simple terms, a partnership is a team that consists of two or more entities working together to complete a common goal or task. Implicit in the definition are mutual trust, defined roles, and the necessary competence/agency. For the success of the Human-AI partnership, AI cannot remain a passive rudimentary artifact; AI must become better integrated, more active partner, and support continuous learning. The primary purpose of this research program is to improve Human-AI partnership in the biomedical systems domain for the development of enhanced monitoring and diagnostic technologies related to decision support. Initial work will formulate the improved partnership framework, conceptualize implementation, and demonstrate the feasibility in application. The long term goals are to expand the research scope to broader areas of human health and performance monitoring and diagnostics, for example, in support of long duration space exploration. To improve the partnership balance we must consider the requirements of each entity. While current research examines elements of how humans can better understand artificial intelligence (explainability, interpretability, transparency, fairness, etc.) there is relatively little examination of improving artificial intelligence's perception or understanding of humans for medical advisory and decision support systems. There are initially three technical areas of examination for this program: 1. Pervasiveness - Embedding of hardware and software to support the improved integration and interfacing. This area will include data acquisition modalities, data fusion, optimized hierarchical computation, distributed computing. 2. Agency and Agility - Defining the role and expectations for partnership to improve overall performance and outcomes. This area focusses on improving autonomy, active support, and performance of medical advisory systems under various Human-AI partnership structures. 3. Active learning - Realtime and continuous feature identification and quasi-dynamic model performance enhancement. Most artificial intelligence assumes static ideal training data to produce fixed models. Medical data is often imbalanced, incomplete, and difficult to acquire, resulting in lower performing models that are expensive to retrain with the current approach. Additionally, the rate and volume of new data has become unmanageable for human ground truth labelling. Active learning will allow the artificial intelligence to identify new features and relationships in the data with more efficient model updates.
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-2022-05438
Name nom
Doyle, Thomas
Department département
Electrical and Computer Engineering
Organization id
27
Institution établissement
McMaster University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2022
Award amount
28000
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
1510
Committee name en
Electrical and Computer Engineering
Committee name fr
Génie électrique et informatique
Area of application code
518
Area of application group en
Manufacturing processes and products
Area of application group fr
Procédés et produits industriels
Area of application en
Medical equipment and apparatus
Area of application fr
Appareils et instruments médicaux
Research subject code
1901
Research subject group en
Biomedical engineering
Research subject group fr
Biomedical technology
Research subject en
Génie biomédical
Research subject fr
Technologie biomédicale
Application title
HUman partnership with bioMedical Artificial iNtelligence (HUMAN)
Keyword
active learning; artificial intelligence; autonomous medical advisory systems; biomedical signal processing; diagnostic instrumentation; human in the loop; intelligent and integrated systems; joint cognitive system; machine learning; pervasive computing
Application summary
Human partnership with Artificial Intelligence is fundamentally flawed because current systems are not designed with a team context or structure. In the biomedical systems domain where decision support can be critical, potentially life and death, it is vital to better balance this partnership. In simple terms, a partnership is a team that consists of two or more entities working together to complete a common goal or task. Implicit in the definition are mutual trust, defined roles, and the necessary competence/agency. For the success of the Human-AI partnership, AI cannot remain a passive rudimentary artifact; AI must become better integrated, more active partner, and support continuous learning. The primary purpose of this research program is to improve Human-AI partnership in the biomedical systems domain for the development of enhanced monitoring and diagnostic technologies related to decision support. Initial work will formulate the improved partnership framework, conceptualize implementation, and demonstrate the feasibility in application. The long term goals are to expand the research scope to broader areas of human health and performance monitoring and diagnostics, for example, in support of long duration space exploration. To improve the partnership balance we must consider the requirements of each entity. While current research examines elements of how humans can better understand artificial intelligence (explainability, interpretability, transparency, fairness, etc.) there is relatively little examination of improving artificial intelligence's perception or understanding of humans for medical advisory and decision support systems. There are initially three technical areas of examination for this program: 1. Pervasiveness - Embedding of hardware and software to support the improved integration and interfacing. This area will include data acquisition modalities, data fusion, optimized hierarchical computation, distributed computing. 2. Agency and Agility - Defining the role and expectations for partnership to improve overall performance and outcomes. This area focusses on improving autonomy, active support, and performance of medical advisory systems under various Human-AI partnership structures. 3. Active learning - Realtime and continuous feature identification and quasi-dynamic model performance enhancement. Most artificial intelligence assumes static ideal training data to produce fixed models. Medical data is often imbalanced, incomplete, and difficult to acquire, resulting in lower performing models that are expensive to retrain with the current approach. Additionally, the rate and volume of new data has become unmanageable for human ground truth labelling. Active learning will allow the artificial intelligence to identify new features and relationships in the data with more efficient model updates.