Prove It Without Revealing It: Exploring the Power of Zero-Knowledge Proofs in Data Science (Ep. 218)
Data Science at Home
00:00
What Is a Zero Knowledge Proof?
Federated learning is an approach that allows different teams for example or different owners to train their part of the model. The data will stay siloed in each of the participants infrastructure and each participant let's say promises to train to keep training the model, then pass the model to the next one. Of course also this approach has limitations due to the fact that one can run away with the trained model so far or not share the model at all after its turn. Even worse they can contaminate the model and train it with fake data or inaccurate data for which the model would start degrading. So what do we do? Well I think that one other concept that reserves some discussion goes under the
Transcript
Play full episode
