straycat on Nostr: I think I said that wrong; the activation function of neural networks doesn’t ...
I think I said that wrong; the activation function of neural networks doesn’t multiply the weighted sum by the sigmoid function; it uses the sigmoid function to map the weighted sum onto an interval.
But the main point remains, which is that it is nonlinear.
Published at
2024-09-25 02:51:32Event JSON
{
"id": "ef5795ceb040372a56d6113c28346b70aee36e66e17689938f0b71c9fd9af665",
"pubkey": "e5272de914bd301755c439b88e6959a43c9d2664831f093c51e9c799a16a102f",
"created_at": 1727232692,
"kind": 1,
"tags": [
[
"e",
"bd5b3e31acbec7204bbaa6853dd57c780dc3b19b19cc0a92afcb1c2d0b5e4b87",
"",
"root"
]
],
"content": "I think I said that wrong; the activation function of neural networks doesn’t multiply the weighted sum by the sigmoid function; it uses the sigmoid function to map the weighted sum onto an interval. \n\nBut the main point remains, which is that it is nonlinear.",
"sig": "a2afb5b4ebef0f93f1ee7c1c47793036914381f9ab889e42a4d6a78e2e97568f7a5aa0ac05caa6379bbfd756609b9dedc3c792d939e5226e7ff35411ee95047c"
}
