Grant: Canada Research Chairs
The University of British Columbia
$70,000.00
Responsible Artificial Intelligence
EF 2024-25
Résumé
My research program at the CRC in Responsible Artificial Intelligence is pioneering an original approach to Federated Learning (FL), aimed at addressing the inherent challenges posed by data and device heterogeneity. Traditional AI models trained on homogeneous datasets fail to generalize, leading to robustness and reliability issues upon deployment across diverse and siloed datasets. This heterogeneity often undermines the performance of machine learning systems when applied to real-world applications. In addressing the pressing challenges of data heterogeneity in FL, my CRC program sets forth multi-faceted orthogonal solutions, built upon my promising preliminary results. Specifically, my team and I aim to innovate FL through model-centric strategies that enhance robustness and reliability, develop privacy-preserving synthetic data methods to secure FL training, and advance client-specific prompts for contextual learning, that emphasize model-centric, data-centric, and client-centric strategies respectively, bridging the gap between FL research and application. To this end, my methods include tailored FL model interpolation and adaptive subnetwork selection to balance model generalization and personalization. I integrate privacy-preserving synthetic data to safeguard FL training and employ in-context learning with client-specific prompts and foundational models, aligning FL deployment with the latest deep learning innovations. Training 4 PhDs, 4 MScs, and 3 UGs in machine learning, data analysis, and interdisciplinary research is also a key objective for this CRC program. This CRC proposal positions UBC at the forefront of FL research, poised to make significant contributions to the field and establish a new benchmark for responsible machine learning development. The program's comprehensive approach promises to lead to groundbreaking advancements in FL, ensuring models that are robust, fair, and reflective of the diversity inherent in real-world data.
Programme
Canada Research Chairs
Bénéficiaire
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
CRC-2023-00342
Name nom
Li, Xiaoxiao
Department département
Electrical and Computer Engineering - The University of British Columbia
Organization id
2
Institution établissement
The University of British Columbia
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
70000
Program id
CRC
Program name en
Canada Research Chairs
Program name fr
Chaires de recherche du Canada
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
9998
Committee name en
Canada Research Chair Selection Committee
Committee name fr
Comité de sélection des Chaires de recherche du Canada
Area of application code
9999
Area of application group en
Not available
Area of application group fr
Non disponible
Area of application en
Not available
Area of application fr
Non disponible
Research subject code
9999
Research subject group en
Not available
Research subject group fr
Non disponible
Research subject en
Not available
Research subject fr
Non disponible
Field of research list codes
RDF1020104
Field of research list names en
Machine learning
Field of research list names fr
Apprentissage machine
Socioeconomic objective list codes
RDS10609; RDS10610
Socioeconomic objective list names en
Computer, electronic, optical products and communication equipment;Information systems, technologies and services
Socioeconomic objective list names fr
Produits informatiques, électroniques et optiques et équipement de communication;Systèmes, technologies et services d'information
Application title
Responsible Artificial Intelligence
Keywords
Computer Vision; Deep Learning; Trustworthy AI; Machine Learning; Federated Learning; Medical Image Analysis; Healthcare
Application summary
My research program at the CRC in Responsible Artificial Intelligence is pioneering an original approach to Federated Learning (FL), aimed at addressing the inherent challenges posed by data and device heterogeneity. Traditional AI models trained on homogeneous datasets fail to generalize, leading to robustness and reliability issues upon deployment across diverse and siloed datasets. This heterogeneity often undermines the performance of machine learning systems when applied to real-world applications. In addressing the pressing challenges of data heterogeneity in FL, my CRC program sets forth multi-faceted orthogonal solutions, built upon my promising preliminary results. Specifically, my team and I aim to innovate FL through model-centric strategies that enhance robustness and reliability, develop privacy-preserving synthetic data methods to secure FL training, and advance client-specific prompts for contextual learning, that emphasize model-centric, data-centric, and client-centric strategies respectively, bridging the gap between FL research and application. To this end, my methods include tailored FL model interpolation and adaptive subnetwork selection to balance model generalization and personalization. I integrate privacy-preserving synthetic data to safeguard FL training and employ in-context learning with client-specific prompts and foundational models, aligning FL deployment with the latest deep learning innovations. Training 4 PhDs, 4 MScs, and 3 UGs in machine learning, data analysis, and interdisciplinary research is also a key objective for this CRC program. This CRC proposal positions UBC at the forefront of FL research, poised to make significant contributions to the field and establish a new benchmark for responsible machine learning development. The program's comprehensive approach promises to lead to groundbreaking advancements in FL, ensuring models that are robust, fair, and reflective of the diversity inherent in real-world data.