
Federated Learning, Healthcare AI | Ittai Dayan, cofounder of Rhino Health
Infinite Curiosity Pod with Prateek Joshi
00:00
Algorithmic Bias in Health Care
A lot of early development of AI has been driven by large medical centers who are in very specific states or countries and tackle very specific patient populations. And so these biases are risky, especially since often you validate an algorithm on a fairly small test set,. But then the plan is to deploy the algorithms in thousands of institutions or at least hundreds if you want to create a successful company from that. So I think it's been really brought those two ways of tackling that one: let's train on much more data but also adapting algorithms over time.
Transcript
Play full episode