It’s incredible to me that they’re getting worse- although it seems they are. Especially if you compare the Euro from 2-3 years ago.
Ive had opportunity to see some of the advancements in what’s called : deep learning technology and I can’t believe it’s yet to be implemented into this field?!?!
Deep learning is going to revolutionize data analysis. Up until recently, most Computer data was comprised of a majority of noise. The value portion, at least in Radiology, was 20%. Deep Learning wipes away that 80%, which is Useless noise, and leaves you with a superior product.
I wonder when this will happen in the models? Maybe it has?
I’m in the weeds here but it’s area that fascinates me (in a layman’s way).
Take a single data point, say temperature, at KROC, on Dec 23, at 1200, up until now, most computers would assess this temperature on an open scale, it says, “jeez, this temp could be anywhere from 0kelvin to ~ kelvin. Deep Learning says, “nah, remove all that nonsense, we know this temp will be between -20f and 80f.”
It doesn’t seem like much. But in the end you’re opening up tremendous opportunities and narrowing noise and error and speeding things up.