AIDRC Seminar Series - Mikhail Belyaev

Title:

Computer vision for medical imaging: research challenges & market demands

Abstract:

Medical imaging is one of the most popular applications in modern computer vision. On one side, the total number of medical images is constantly increasing and many countries, such as the UK, face a significant shortage of radiologists. On another side, Geoffrey Hinton's 2016 statement, "we should stop training radiologists," doesn't seem correct, as the medical AI market is still in its infancy. We'll briefly cover market demands and critical industry barriers, then switch to an overview of computer science problems, including a low amount of annotated data and various domain shifts. Finally, we will cover possible solutions such as self-supervised learning and domain adaptation.

Bio:

Mikhail Belyaev is an expert in Artificial Intelligence with 14 years of experience and a solid industrial and scientific track record. He got his PhD in Data Science in 2015; then, he became a research group leader at the Institute of Information Transmission Problems and later an Assistant professor at Skolkovo Institute of Science and Technology, a young technological university established in collaboration with MIT. In 2020 Mikhail founded Intelligent Radiology Assistance Laboratories Ltd, a medical AI startup, and currently serves as CEO.