Join the Machine Learning in Medical Imaging Consortium (MaLMIC) for an opportunity to network on machine learning
Forum on Federated Learning for Medical Imaging
Friday, February 18, 2022
3:30 to 5:30 p.m. EST
Kam Kafi, Director of Clinical Strategy and Oncology at Imagia will outline the case for federated learning and work underway between Imagia and collaborators. Spyridon Bakas, Assistant Professor at UPenn, will talk about an open platform he has developed to support federated learning for radiology. Aline Talhouk, Assistant Professor at UBC, will talk about privacy issues and introduce the LEAP platform. The presentations will be followed by group discussion on lessons learned, opportunities to collaborate, and sharing of resources.
Interested in joining? Please contact us.
Forum on Federated Learning for Medical Imaging
Friday, February 18, 2022
3:30 to 5:30 p.m. EST
Kam Kafi, Director of Clinical Strategy and Oncology at Imagia will outline the case for federated learning and work underway between Imagia and collaborators. Spyridon Bakas, Assistant Professor at UPenn, will talk about an open platform he has developed to support federated learning for radiology. Aline Talhouk, Assistant Professor at UBC, will talk about privacy issues and introduce the LEAP platform. The presentations will be followed by group discussion on lessons learned, opportunities to collaborate, and sharing of resources.
Interested in joining? Please contact us.
The promise of federated learning : Challenges and progress towards a workable solution
Kam Kafi, MD
Director of Clinical Strategy and Oncology, Imagia
Talk summary: Kam will provide an overview of federated learning, data preparation and patient privacy considerations, and show some of the work they have been doing internally at Imagia.
The Federated Tumor Segmentation (FeTS) Initiative
Spyridon Bakas, PhD
Assistant Professor, University of Pennsylvania
Talk summary: In this talk, the Federated Tumor Segmentation (FeTS) initiative will be presented, which represents a collaboration between UPenn and Intel that successfully completed what-seems-to-be the largest to-date real-world federated learning study involving 59 geographically-distinct healthcare sites, and the development of a suite of tools (including the OpenFL framework) to facilitate it. Results will be presented from an initial feasibility simulated study (the first of its kind), the real-world endeavor, as well as future considerations.
Barriers to the implementation of federated learning in healthcare settings
Aline Talhouk, PhD
Assistant Professor, University of British Columbia
Talk summary: Aline will address real and perceived barriers to implementation of federated systems and federated learning in health settings. These are based on preliminary data from qualitative interviews with ethics and privacy officers and data stewards as part of a multi-national study.