Skip to content

Comment on TimesFM: Time Series Foundation Model for time-series forecastingparent

Comments

Every time i've actually tried something like this it has not outperformed statistical process control.

There are clear counterexamples to your experience, most notably in maintaining plasma stability in tokamak reactors: https://www.nature.com/articles/s41586-021-04301-9

Interesting. Could you point me to where it is compared against SPC? I didn't find it from a cursory read.

task specific model

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.