ML is useful for many things, but not for predicting scientific replicability

Re:ML is useful for many things, but not for predicting scie... by mirekrusin on Saturday August 12, 2023 @11:27PM (#37106404)

Reads like release notes for some early bayesian spam filter.

[ Parent | on HN ]

Re:ML is useful for many things, but not for predicting scie... by mdp2021 on Sunday August 13, 2023 @03:17AM (#37107460)

I saw the same: I was writing my comment (then I noticed this parent) as «Without the implementation of Understanding, a ML system would see, if not a bag of words, an ordered set of words (anything more?)... May or may not get more input Information than a Bayesian filter».

[ Parent | on HN ]