Mikhail Belyaev

Mikhail Belyaev’s lab: Medical Computer Vision

Why data science for medicine

• Algorithmic challenges: heterogeneous 3D medical images require new computer vision methods.
• Social impact: it’s not about selling ads; it’s about saving lives.
• Innovations: leading tech companies such as Google and Nvidia actively work in the field; medical AI startups received more funding than any other AI sector in 2019.

Why Mikhail’s lab

• Team: we are the only group in Russia that regularly gets accepts to the top-level conferences and journals. Our multitask approach for sorting of COVID-19 patients was published in Medical Image Analysis, the top 1 journal in the field, and top-4 among all Computer Vision journals. We regularly participate at MICCAI, the leading conference in the medical image analysis field. Last year we got two papers (1, 2) at the main conference and a paper at a domain adaptation workshop; this year we got an oral talk at the main conference (~ top 4% of submitted works) and the best paper award at a workshop. The full list of Mikhail’s publications can be found at Google Scholar.
• Data & Partners: we collaborate with the leading medical centers and routinely process up to a couple of terabytes of images within our projects.
• Impact: IRA Labs LLC, a startup company, delivers the results of our research to hundreds of doctors and thousands of patients every day. IRA Labs COVID-19 solution was ranked as a leader among 15 participants of the 2020 mosmed.ai contest.

About Mikhail

Mikhail received his Master’s degree in Applied Mathematics and Physics from the Moscow Institute of Physics and Technology in 2010. After graduation, Mikhail developed and implemented several machine learning algorithms for industrial data analysis at Datadvance company as a Senior researcher. In addition to his industrial career, Mikhail also proved the developed ML method’s asymptotic optimality as a part of his Candidate of Science thesis, defended in 2015. In 2016, Mikhail created a research group to advance machine learning approaches for neuroscience data. After a couple of pivots, the lab has finally switched to Medical Computer Vision.