Steve Gisselbrecht on Nostr: The thing that makes me crazy about machine learning in science is that if I said in ...
The thing that makes me crazy about machine learning in science is that if I said in my Methods section, "I looked at all the data and decided which ones were affected" I'd be crucified in peer review, while if I said, "I trained a recurrent neural network for an unsupervised classification task" it'd be treated as cutting edge science, and the former is the latter with a bigger, better configured network.
Published at
2025-08-04 17:49:11 UTCEvent JSON
{
"id": "790f42e27f3d9526e72908fde2628245f1470fda999b4ea1b9fdc3e147552e8b",
"pubkey": "3fa7e50bd521ecf5f9406699b47911809a7d67f3a3f98fac191dca3412164668",
"created_at": 1754329751,
"kind": 1,
"tags": [
[
"proxy",
"https://mas.to/users/stevegis_ssg/statuses/114971754582487058",
"activitypub"
],
[
"client",
"Mostr",
"31990:6be38f8c63df7dbf84db7ec4a6e6fbbd8d19dca3b980efad18585c46f04b26f9:mostr",
"wss://relay.mostr.pub"
]
],
"content": "The thing that makes me crazy about machine learning in science is that if I said in my Methods section, \"I looked at all the data and decided which ones were affected\" I'd be crucified in peer review, while if I said, \"I trained a recurrent neural network for an unsupervised classification task\" it'd be treated as cutting edge science, and the former is the latter with a bigger, better configured network.",
"sig": "a9e134a3d771654dc9aca6c3300dfed569fa0fac866b8c833b2a83dcafc44abea318e3b2982e0970ff23180e914303e585c422a3d72e4091cd843d90a2d011cc"
}
