Hacker News Reader

Comment by contemporary343 | Hacker News Reader

Her research work is well-known and has had an impact, whether you like it or not: https://scholar.google.com/scholar?q=Timnit+gebru

Edit: And to be clear, her most cited paper (> 1000 citations) had an impact on what some companies did: https://www.theverge.com/2020/6/8/21284683/ibm-no-longer-gen...

What is disturbing here, from a researcher's point of view (even within a company!), is the lack of transparency on why a paper is being blocked from publication. The implicit understanding of being a researcher in a company like Google is that, while papers are subject to review prior to submission/publication, that the company will not wholesale reject it and instead will work with you to make modifications and changes. Clearly something different happened here that should be of concern to people considering working at Google Brain.

View on HN

Citations by like minded publications hardly constitutes impact. Contrary to what most academics believe, impact isn't measured between the ears of individuals in academia.

Uh, what? She's a researcher and this is a Google research division. A 1000 citations is a large number for any single paper, even in ML. And yes, even inside Google research, this aspect of impact plays a big role in evaluation of researchers.

1000 citations is A LOT. I think the person you're responding to is that it's a self-fulfilling prophecy. Popular people get quoted regardless if their paper is actually impactful and that it's, unfortunately, often a popularity contest. People are lazy and only read/cite popular papers

I work in research, and citation count, especially of a single paper, is such a dumb metric for evaluating performance. There are about a million reasons that one paper gets cited over another that don’t have to do with quality.

Does the same paper get 1000+ citations if it were published in 2010? Or even 2015? Probably not.

It's an original research paper (not review) and it was published in 2018. So yes, 1000 citations at the very least conveys that it was noticed and had an impact.

Also, anyone who works in the field of AI fairness and ethics research knows about it. So yeah, it's an important study.

Andrew Wakefield's paper which started the anti vax movement got much more citations than that.

What does that tell you?

Attention != quality. Especially not the type of quality that keeps one from getting fired after publicly badmouthing ones boss.

As far as I can see, the only way she would keep her job, is if her research was so fundamental to the company that they literally could not function without her. Citations don’t tell you that. My boss is a star researcher in his field with multiple 1000+ citation papers. He still has to act respectfully of his boss and peers.

There are multiple types of impact. The question is whether it brings value to Google as a corporation, and thus leverage to her as an employee.

FAANG hires top diverse ethics researchers for good publicity and they demand meaningful changes that aren't in FAANG's interest and get silences/fired resulting in bad publicity

FAANG hires top diverse ethics researchers for good publicity and they demand clout/unreasonable changes that aren't in FAANG's interest and they get kicked resulting in bad publicity.

Unclear which is happening here. I agree it isn't in FAANG interest to hire these people but that misses the point.

It seems plausible that a key result of her research (likely including the unnamed paper which she was asked to retract) is driving change in policy and social opinion that is apparently considered harmful to Google, so the research does not add value to Google, it benefits others (e.g. particular underprivileged groups) but just imposes extra restrictions and costs to her employer.

Industry research is not academia, where you'd expect to be free to choose the avenues of your research, and you should expect that publication is restricted. The technical improvements you make may be kept under wraps to gain a competitive advantage; and policy agenda work is essentally part of PR strategy, it's accepted if and only if that agenda benefits the strategic interests of the company. Tenure in academia offers some freedom of research (though not even then abolute), but when working for private companies, you don't have that freedom unless you have a specific contract asserting that, and even then you'd be wary of biting the hand that feeds you.

Like, if you want to expose climate change risks, it's not going to work out while employed in an oil company, and if you want to advocate for regulation of ML research while working in a company that would be bound by this regulation, this is certainly something that the management would be expected to care about.

I generally agree with your comment - industrial research is not academia. But (relative) freedom to publish is a very important consideration when attracting top-flight researchers. Especially in ML! I mean, even Apple's ML team publishes.

I'm not sure Google Brain's recruiting pitch includes a description of them blocking publication of a paper you'd worked a long time on (and even given them a heads up about, as Gebru apparently had). Their pitch is almost certainly about being able to do research similar to how you would in academia (except with more money and resources, etc.).

Sure, companies may have interests in some areas and not others, but it is counterproductive and even irritating if the company does not clearly communicate what the areas of interest are and which are not. An order to retract without explanation is rude and counterproductive.

Yes, sure, the current situation a sign of a misalignment in goals (and likely values) that has gone unaddressed far too long; and that's the responsibility of managers to ensure that this communication happens early. However, it somehow seems that if this situation was managed properly, the result would simply be that she would quit or get fired some time earlier.

It is not exactly a wholesale reject. She submits the paper to the conference first and asks for approvals later, which might be common if you are rushing for a deadline. This has the side effect of approvers forcing you to retract the paper (because the conference deadline has passed and you can't make modifications on it).

Reply on news.ycombinator.com