nope.. they are irritated by those people who have limited understanding of the subject/dataset and they randomly try models that 80% of them are destined to fail and only a true expert can derive some useful results out of them.
Experts are irritated by how machine learning ignores their knowledge, but it’s often that ignorance that allows it to discover things they didn’t know.
-
-
-
If 20% succeed, that’s plenty.
- Näytä vastaukset
Uusi keskustelu -
-
-
I think (domain) experts are irritated because despite grandiose claims, in practice, ML fails on most "out-of-distribution" tasks (which is where new knowledge usually lies) and doesn't have simple common sense.
-
Most of the time (any exception?) it finds common denominator for extant knowledge at best, doesn't discover new one....
Keskustelun loppu
Uusi keskustelu -
-
-
“Problems that remain persistently insoluble should always be suspected as questions asked in the wrong way.”- Alan watts Our lifespan is too short to observe the patterns and ask right questions, answers might be in the pic you last took from your ph camera.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
This is often true. The point is how to incorporate expert knowledge effectively into the ML system. This is often a key to success.
-
Expert knowledge is incorporated in feature engineering. FE is the key to build ML systems which work. DL may discover features, but as it is a cheap technology now, everybody can use it in a competitive environment such as stock markets and any competitive advantage is lost.
Keskustelun loppu
Uusi keskustelu -
-
-
It's also the same ignorance that leads to wild expectations when the algorithm games the results. Ignorance like naivety is a two-sided blade.
-
The progress we make in deep learning is a consequence of our overall ignorance about general intelligence. There are many alternative ideas on cognition developed by other fields. But these were done without the benefit of computational models.
- Näytä vastaukset
Uusi keskustelu -
-
-
Marconi invented the trasmitter, but never got a university degree. Funny story is that Marconi was wrong also in his own framework. It was mostly luck. Well ignoring previous knowledge could lead to gain a naive insight, lucky discoveries, but it will no push boundaries further.
-
On the other hand, I am an expert on being an expert in nothing. Consequently, I am ok with machine learning:)
Keskustelun loppu
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.