GTC 2020: Making Radiology AI Models More Robust: Federated Learning and Other Approaches
Daniel Rubin, Stanford University
Learn about the key types of clinical use cases for AI methods in medical imaging and the critical challenges and progress in applying AI in these applications. We describe current challenges to creating robust AI models, such as insufficient quality labeled data and access to data. Next, we'll describe recent AI projects that tackle the challenges, including weak/observational learning and federated learning.