Vitalik Buterin
@vitalik.eth
One of the fascinating things about AI and cryptography is how similar the math sometimes is. Like, the way that the poseidon hash works (matrix mult -> per-unit nonlinear layer -> matrix mult -> per-unit nonlinear layer) is *exactly* what deep neural networks do.
16 replies
4 recasts
19 reactions
tina
@howdai
Can you expand on what this means to you in terms of the interplay between crypto x AI?
2 replies
0 recast
0 reaction
Cassie Heart
@cassie
There’s a lot of potential for overlap — secure MPC enables multiple dataset owners to combine their sets for private training of a neural network without revealing their inputs to the other parties, yet getting an end output of a network trained on all participants’ data.
3 replies
0 recast
1 reaction
chandresh 🪴
@chandresh.eth
something smart going on in this conversation, no idea what not i am sure it's useful 🤞
1 reply
0 recast
0 reaction
Cassie Heart
@cassie
Decentralization offers a place where people can come together to collectively agree on the state of data. MPC is a process where two or more parties can leverage cryptography to securely compute a resulting output without necessarily revealing their individual inputs. MPC can be applied to any computable function,
2 replies
0 recast
0 reaction
Cassie Heart
@cassie
at the expense of how much overhead MPC may bring to do so. Efficient techniques in MPC for particular mathematical operations (matrix multiplication, for example) exist. A decentralized platform for MPC could combine the benefits of crypto (self-ownership of data, monetization of data) and ML such that you have a
1 reply
0 recast
0 reaction
chandresh 🪴
@chandresh.eth
this is so much more clear, thank you cassie 🙌🙏
0 reply
0 recast
0 reaction
