Guest conversation with Prof. Schultze: machine learning on transcriptomic, radiological and healthcare data (Nature)

Join us on the 31st October at 10:15 am, for a typical Journal Club discussion, with the caveat that the senior author is joining our discussion.  We will talk about how a utilise a fully decentralised architecture, where distributed machines can automatically join, learn locally, share globally, and then update local models that have predicted who has leukaemia, tuberculosis or COVID-19. The same technology was applied to create draft annotations on X-ray images. Together, we will explore the future possibilities of this paradigm shift in how to conduct international collaboration.

 Professor Joachim Schultze leads the Systems Science research group at the University of Bonn. This integrates in vivo wet-lab data with in silico computing data (including single-cell genomic, epigenomic and transcriptomic, immunologic, metabolic, and neurobiological layers). Over the past five years professor Schultze has published twice in Nature, twice in Immunity, and five times in Cell. We are privileged that he will be available to answer our questions via Zoom.

Attendees are asked to read this paper:
Warnat-Herresthal S, et al. Swarm learning for decentralized and confidential clinical machine learning. Nature. 2021.
www.nature.com/articles/s41586-021-03583-3

You will have the opportunity to can ask any question about this paper, the process of getting there, and how this technology might be applied in the future in a relaxed and informal atmosphere.

 Journal Club discussion takes place Thursday 31th October, 10:15-12:00 (Estonian Time, and Eastern European Time)

We look forward to seeing you in person at room 3115, Delta Centre, Narva mnt 18, University of Tartu. For those unable to make it in person, online access is possible via https://ut-ee.zoom.us/j/91020203604?pwd=YnptVXc2N0NLQnA3NTQ5dndXM1JIUT09

No registration required.