[jjcon]

Not as a rule though - so many ML systems are utilizing data that is streaming in from passive sensors or transactional streams and is not human curated at all. The human aspect isn’t an intrinsic property of ML or even these algorithms only particular applications (and I would guess a minority of applications too).

Given that, it seems to be a clear miss to apply that logic generally. I have to believe it most likely stems from a lack of basic understanding and competence on the authors part.

Edit: Here are two examples that I have personally worked on:

Global calculations of weather composite reflectivity using 20 years of historical satellite imaging data from NASA

Superior RF demodulation (under certain circumstances) using RF transmissions as received and sent.

Both of these utilize modern ML imaging models, neither require any human labeling only streaming data (which in these cases began collection long before modern ML techniques were in widespread use). The applications in the natural sciences are endless not to mention the applications more on the business intelligence side using transactional data. Only in specific cases is human labeling required but because of the high cost of that data it is by its nature dwarfed by that which is collected naturally (not to mention often error prone). It is for that reason that techniques to ingest data that is more and more natural to collect are growing in favor.

[YeGoblynQueenne]

>> I have to believe it most likely stems from a lack of basic understanding and competence on the authors part.

That is unlikely, given that one of the authors is Timint Gebru. I'm quoting below select passages from her wikipedia page indicating her background:

In 2001, Gebru was accepted at Stanford University.[2][5] There she earned her Bachelor of Science and Master of Science degrees in electrical engineering[8] and her PhD in computer vision[9] in 2017.[10] Gebru was advised during her PhD program by Fei-Fei Li.[10]

Gebru presented her doctoral research at the 2017 LDV Capital Vision Summit competition, where computer vision scientists present their work to members of industry and venture capitalists. Gebru won the competition, starting a series of collaborations with other entrepreneurs and investors.[11][12]

Gebru joined Apple as an intern while at Stanford, working in their hardware division making circuitry for audio components, and was offered a full-time position the following year. Of her work as an audio engineer, her manager told Wired she was "fearless," and well-liked by her colleagues

https://en.wikipedia.org/wiki/Timnit_Gebru

[jjcon]

I didn’t say they were uneducated but an audio hardware engineer does not imply a good working knowledge of industry trends in ML applications.

Regardless, my point still stands, they completely ignore (willingly or ignorantly) that human labeled data is not intrinsic to ML or even the algorithms themselves and in all likelihood is a small minority of datasets used by modern ML applications. To then apply that critique generally to ML shows ignorance and a misunderstanding of the ecosystem.

[YeGoblynQueenne]

Gebru is not an audio hardware engineer. I call your attention to this passage I quoted above:

Gebru presented her doctoral research at the 2017 LDV Capital Vision Summit competition, where computer vision scientists present their work to members of industry and venture capitalists. Gebru won the competition, starting a series of collaborations with other entrepreneurs and investors.[11][12]

And to the fact that she got her PhD in computer vision, i.e. the main area of AI research that the article seems to be criticising.

[jjcon]

Her work experience is as an audio engineer - but again it doesn’t matter what her credentials are, she is wrong regardless and you are ignoring my whole point. She shows her ignorance of the subject matter (again willingly or not) when she applies her critique generally at Ml and not just at these specific applications - not sure how many times I need to say that.

[YeGoblynQueenne]

>> Her work experience is as an audio engineer

Her PhD research is in computer vision and she and her co-authors are writing mainly about computer vision, but you spoke of "a lack of basic understanding and competence on the authors part". That is clearly incorrect and I don't understand what saying the same thing many times will change about that.

[jjcon]

Computer Vision is a domain and is not equivalent to machine learning. They overlap yes, but not necessary. Again though you have completely ignored my point again and again. The authors ignorantly conflate specific applications of ML with the entire industry. That plainly demonstrates a clear lack of competence in this area.