9 February 2024 · Okinawa, Japan (virtual)

Invited Talk at the Okinawa Institute of Science and Technology Graduate University, Japan

Machine Learning Insights and Innovations (MLII) online seminar series, OIST ↗

Invited Talk at the Okinawa Institute of Science and Technology Graduate University, Japan
Photo: Albarqouni Lab
Online seminar “Unlocking the Potential of Federated Learning in Medical Imaging” in the MLII series of the Okinawa Institute of Science and Technology.

On 9 February 2024 I gave the online seminar “Unlocking the Potential of Federated Learning in Medical Imaging” in the Machine Learning Insights and Innovations (MLII) series of the Okinawa Institute of Science and Technology (OIST), Japan. Deep learning needs large datasets, but privacy rules make it hard to pool medical data. Federated learning lets hospitals train models together by sharing model parameters instead of data. The talk covered the clinical workflow and its AI challenges, such as heterogeneous and scarce data, domain shift and multi-modal imaging, then the promise, pitfalls and solutions of federated learning, with a case study on brain MRI from our Nature Machine Intelligence paper. Source: OIST event page.

Where