3 ms·
I always thought this kind of stuff would be the killer app for machine learning. Products slowly drift out of tolerance, and the ml applies corrections to get
by jfoutz 8y ago
I always thought this kind of stuff would be the killer app for machine learning. Products slowly drift out of tolerance, and the ml applies corrections to get back into tolerance. Apparently there’s just not enough data to respond well.
- harperlee 8y agoYou don’t even need to get that “modern”; feed back information and use some classic control theory and you should be able to have a system that compensates for change in tolerances.