So Chinese unicorns actually applying this interesting growth strategy in B2C space to B2B, because B2B in China is so hard, Federated learning-- you can do collaborative machine learning without centralized training data.
All of that needs to be stood up. They are federated . They have associations, nationally and globally.
But there is a lot of precedent for this. It's marginally more expensive to do federated machine learning than do everything in the center, but companies don't care, and consumers decidedly don't care.
So this is the current model, so a corporate central ownership. And their model is basically a federated cooperative model, where because the laws about healths data are so different from country to country, you could basically release your data across borders through the sort of federated cooperative data commons.
And we'll see what that means. And then I'll discuss Federated Byzantine Agreement, which is a generalization of Byzantine Agreement, to a model that could actually accommodate this kind of internet-level consensus. Then I'll talk about failure resistance and how well we can do in terms of tolerating failures.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. And the key idea in Federated Byzantine Agreement is that we're going to pick our quorums in a decentralized way.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So the first big difference between Federated Byzantine Agreement and the standard centralized Byzantine Agreement is that in the Federated model, failure
I mean it's an emergent property. It's something that isn't there when you start to do it. It emerges from the synergies among data knowledge But the idea is that we want to federate all the capabilities and have people tagging information so that they get linked up in ways that we didn't know were even possible.
The fortune that I have is a fortune you can't have. What the walkways get is like a kind of federated universe of slightly different variations on their ethics, and on their design aesthetics, and on their practices, their engineering practices, their social practices.
Aircraft have gone through a change. They used to use what we call a federated architecture. Meaning, you had one box per function for the software, which means that everything was well separated.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. And the core idea in the protocol is this technique Federated voting.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So as before, Federated voting, nodes v are issuing these vote messages.
At the last, we can use it as a marketing thing to attract this crowd. but what's being sort of described as the federated Blockchain-- the idea that an association of entities with a shared
And their model is basically a federated cooperative model, where because the laws about healths data are so different from country to country, you could basically release your data across borders through the sort of federated cooperative data commons. There is more conceptual work underway with Fairbnb, where basically groups in--
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So now I'm going to talk about this new model called Federated Byzantine Agreement, which is a generalization of standard Byzantine Agreement to a setting where you don't
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. that statement settled. So a Federated Byzantine Agreement system is basically a set of nodes V and a quorum function Q, where Q of V is the set of slices
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So effectively, the outcome of federated voting looks like this.
That sounds great, here we go. It's a lot more blended, a lot more kind of federated and inter-twingled than that.
What it meant was from foreign countries they are very interested in ramen but they don't know who to talk to about it because there isn't a big player, and there wasn't a company that federated them. When they started doing their research they stumbled upon Takumen they said that it looked like it connected a lot of ramen stores.
It maybe an exaggeration but it is very well recognized. What this means is that, for the first time, we had federated the ramen industry. Famous ramen shops, ramen shops that have long lines.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So to summarize, let's say you have a set U of well-behaved nodes in a Federated Byzantine Agreement system.
So the big difference here is that when you have a Byzantine failure, the failed node, it can actually change its vote. So now the question is, can we actually achieve this kind of optimal failure resistance in a Federated Byzantine Agreement system?
as the vehicle operations. In your slide with business models, you had different quadrants of one service provider to rule them all and then a federated system.
I mean it's an emergent property. It's something that isn't there when you start to do it. It emerges from the synergies among data knowledge So my answer so far -- until you or somebody else comes up with a different approach -- is to have two parallel systems, but have them inform each other and federates so that there's