Are There Only 10,000 Machine Learning Experts? LinkedIn Offers a Different Number, 651,627
January 18, 2018
I read in the dead tree edition of the New York Times (still not a tabloid sized “real” journalism delivery vehicle) that there are 10,000 machine learning experts in the world. You can find a version of this story at this link.
Just to check the validity of this magical number, which reinforces the notion of elitism, the one percent of the one percent, and the complexity of the Dark Arts of smart software, I did some research.
I turned to LinkedIn, entered the phrase “machine learning” and this is what I learned from the Microsoft professional social media search system:
I realize that the low key colors and gray type are unreadable, but contact Microsoft LinkedIn, not me.
There are more than 38,000 jobs open for experts in machine learning.
What’s the talent pool?
The number is 651,627.
Now I understand that if one is making a list of top anything, the peak of the pyramid will be, by definition, one. For music, you may have disagreements. For machine learning, it’s different.
Since machine learning and other smart software jargon is pretty vague, mostly incorrect, and generally misunderstood, the New York Times’ story missed the mark by a mere 641,627 “experts.” Keep in mind anyone can say one is an expert in anything unless the government regulates via licenses like those issued to doctors, lawyers, and beauticians. Beauticians? Yep.
Ah, you say. LinkedIn is for marketers and headhunters.
Yes, I respond.
But the point is that in jargon charged disciplines, it is tough to convince me that there are 10,000 machine learning experts in the world. My hunch is that the cream of the crop will be a handful of people, assuming that one can define what it takes to be an expert; for instance:
- Math skills that go beyond the required course in computer science with an emphasis on artificial intelligence
- Math skills which nose into the territory of Kolmogorov and his cronies (yep, my uncle, the crony)
- Database skills tuned to deal with machine learning
- Linguistics capabilities to cope with multi lingual content
- Engineering skills tuned to the peculiar demands of a real time stream of intercepted data from an outfit like WebHose
- Subject matter experts with knowledge of such exciting topics as Bayesian “drift” and how to make necessary human interventions to get the statistical ship back on course
- Operations experts who can get something useful from a ML-infused application like creating a smart home appliance which does not burn the roast chicken which must be well done for an ageing boxer.
I could go on.
Right now, anyone can claim to be an expert in machine learning. The problem is that machine learning is not one thing. Google is bundling up a bunch of stuff and making it available to LinkedIn type experts.
What could possibly go wrong? Let’s hope the New York Times knows exactly which type of expert in the components of machine learning to have a reasonable shot of reporting on the event that catches a “real” newsperson’s attention.
Stephen E Arnold, January 18, 2017
Comments
One Response to “Are There Only 10,000 Machine Learning Experts? LinkedIn Offers a Different Number, 651,627”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
This web site is really a walk-through for all the info you wanted about this and didn?t know who to ask. Glimpse here, and you?ll surely discover it.