Why Nostr? What is Njump?
2024-01-09 10:15:38
in reply to

Nafnlaus 🇮🇸 🇺🇦 on Nostr: npub1vypq6…xaghx Ultimately, though, we're a collection of weightings. We run a ...

Ultimately, though, we're a collection of weightings. We run a predictive model for our senses, and the difference between the predicted senses and the actual senses forms a measure of error. This "backpropagates" throughout the system, not just adjusting the weights of the output layers, but all the way down through the deep hidden "layers" that represent complex conceptual classifications.

** I have to put "backpropagates" and "layers" in quotes because we're structured somewhat..
Author Public Key
npub1dnwv0sr3qm5vc3ea8dxqxpt98e0g49cvzkd3rkgxhhxdyx83mwnsdz2fyw