Stephen Brooks 🦆 on Nostr: I tried adding a neural network to fit the figure of merit across parameter space, ...
I tried adding a neural network to fit the figure of merit across parameter space, then select an imagined optimum. It was crap and didn't help at all. I think I'd have to customise the network to get a better fit, although this whole approach could end up inefficient.
Published at
2023-12-14 02:19:12Event JSON
{
"id": "723f3231ce8a930de22a9b001b72b4957f2287669b32fcb9cafc71a2e39e95a7",
"pubkey": "e0baa8ebcaeed55330a87a40682c68d13c3a914775b416071eddee59f74b962c",
"created_at": 1702520352,
"kind": 1,
"tags": [
[
"e",
"d460baa2f7dc54b711b37549c0e1f328e501a00032adc520561f381c046e70cf",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://mstdn.io/users/sjb/statuses/111576373821982577",
"activitypub"
]
],
"content": "I tried adding a neural network to fit the figure of merit across parameter space, then select an imagined optimum. It was crap and didn't help at all. I think I'd have to customise the network to get a better fit, although this whole approach could end up inefficient.",
"sig": "1e478092232e1f8bfc9c90c2aa6b940d92d238b6594cc6f9f179308077c815ae9e01818b3fe16abc8216a88ae466bb865e066fa8d9ca4ed4d2cf9848b1645424"
}
