Advancements and Challenges for Deep Learning in Medical Imaging

Overview

There have been rapid advances at the intersection of AI and medicine over the last few years, especially for the interpretation of medical images. In this talk, Pranav Rajpurkar, PhD, describes three key directions that present challenges and opportunities for the development of deep learning technologies for medical image interpretation. First, he discusses the development of transfer learning and self-supervised learning algorithms designed to work in low labeled medical data settings. Second, he discusses the design and curation of large, high-quality datasets and their roles in advancing algorithmic developments. Third, he discusses the real-world impact of AI technologies on clinicians’ decision making and subtleties for the promise of expert-AI collaboration. Altogether, Dr. Rajpurkar summarizes key recent contributions and insights in each of these directions with key applications across medical specialties.

About the Presenter

Pranav Rajpurkar is driven by a fundamental passion for building reliable artificial intelligence (AI) technologies for biomedical decision making. His lab approaches biomedical problems with a computational lens, developing AI algorithms, datasets and interfaces that cut across computer vision, natural language processing and structured health data. He has collaborated with clinicians across medical specialties, including radiology, cardiology and pathology, to make some of the first demonstrations of expert-level deep learning algorithms and their effects on clinician decision making. Previously, Dr. Rajpurkar received his BS, MS, and PhD. degrees, all in Computer Science from Stanford University.

His lab’s current research directions include algorithm development for limited labeled data settings, high-quality dataset curation at scale and the design of effective clinician-AI collaboration setups.