I really enjoy this series. The explanations are really clear and simple to understand.

This was concise and well put together. Thank you!

Federated learning explanation was great 👍👍👍👍👍

I prefer it when you draw things from scratch. I am even learning how to do that in preparation for my job interview.

An example of dataset and the regarding insight would have been helpful to understand why the insight is not back traceable.

nice. this gave me the intuition.

In what degree you are considering SMPC instead of other encryption for secure model aggregation? Anyway, Thanks for the comprehensive recent concepts! FTL is new!

very nicely interpreted!

Nice content , where to learn more about federated leraning? Thanks!

awesome content

Thanks man

It would be very great if, spoken about dataset distribution- iid and non iid and how FL is effected . Specifically non iid setting.

greatt!!

Any new research on Federated Learning applied to Foundation Models?

Are you mirrored or did you master the art of writing mirrored?

Identify cats? 😅

We need regulation in the industry where publicaly traded companies issue fractional stocks to those teaching AI models from web scraping for example. Lets say I post code for a tcp socket server IBM would like to use. AI scrapes it, learns it, uses it, and voila, I now am issued one thousandth of a share of IBM. But not the 30 script kiddies who copied my code and put on the Holy You Tube ... because AI would know I posted it first. Or I could say, ah, that would be cool to share, but why post it, I will just get ripped off anyway? Fractional stock issuance solves the clerical head aches of other systems I believe

You realize that sharing the updated learning gives up the individual companies competitive advantage without giving up their sensitive data? They're giving the milk for free, but keeping their cow.