1803
He came with receipts
(mander.xyz)
A place for majestic STEMLORD peacocking, as well as memes about the realities of working in a lab.
Rules
Does 80 technical papers in 2.5 years seem kind of off to anyone else? That's more than a paper every 2 weeks. Is there really time for meaningful research if you're publishing that often? Is he advising a lot of students? If that's the case, is he providing the attention generally needed for each one? Is his field just super different than mine?
In acamedia you usually get your name on most papers where you help a bit. And if you're the boss, you get your name on papers without even helping but perhaps supplying space, material, budget.
I've been in academia. My field required a "significant intellectual contribution" to the research and the writing, so no putting your name on papers if you just supplied space/material/budget. You can get an acknowledgement for that, not an authorship credit.
And which reviewer or publishers verifies how "significant" a contribution is beyond seeing some initials matched with tags like "visualization" or "experimental design"? That's right, nobody. It's not even remotely traceable who did what if you're a reviewer.
Academia is full of fraud and people trying to secure their share of credit because in academia it's all about names, as the twitter exchange above illustrates so profoundly. And the other driver for the sad state of academia is of course having the quantity of published papers as the most important criterion for academic success. The more papers, the more citations, the bigger your name will become. It determines your chances of getting funding and therefore your career. If you want to make a career in science you have little options but to comply with this system.
That's kind of the point I was making.
Sorry, my irony detector must be malfunctioning.
Everything, everywhere is corrupt.
Everything Everywhere All Corrupt
This definitely varies by field, lab, university.
Yes, thought the same but have a quick look at this: https://scholar.google.com/citations?hl=en&user=WLN3QrAAAAAJ&view_op=list_works&sortby=pubdate Seems about right? But yeah, must be advising lots of students or something. He is rarely the first, second or even third author on the papers.
Also many arxiv preprints
Ah. I hadn't really considered preprints or workshops. If I just count the ones that seem to be published in journals or conferences, it's 28. Still prolific. But reasonable in a 10-15 person lab.
Importance of order changes by field. In my field, at least for in lab work: first is the main lab person that worked on the project. Last is the PI, everyone that helped goes in the sandwich. I'm unsure about collaborations between labs and at that point too afraid to ask.
There are some people in this world who are smarter and more motivated than we are.
And then there are people who get a head start when their rich daddy gives 'em a bunch of money and they get lucky with how they invest that money but pretend to be a genius anyway.
This is a fair question. But also, we're talking about one of the most influential minds in deep learning. If anything he's selling himself short. He's definitely not first author on most of them, but I would give all my limbs to work in his lab.
I'm not questioning his contributions to the field. Just being on that many papers. It just seemed like such a crazy amount of publishing.
Though deep learning has been on fire the last couple years. And the list posted included a lot of preprints and workshops, which I hadn't really considered.
He sounds like a hack.
Yeah, even if he is advising or contributing, the way he put it sounds very disingenuous like he's trying to inflate the number for his argument. Which MIGHT mean there likely was not many with immediately recognizable significance in that time (don't yell at me, I have not taken the time to verify this).
Either way, the way he responded comes across as very "I'm published, you're not, neener neener!" which is not a good look for anyone with a doctorates.
Also, genuine question, how significant was the contribution of LeNet-5 to the field of deep learning vs Neocognitron?
He could've just said "I have a turing award, you don't" if he wanted to show off.
He is also called one of the godfathers of deep learning, so I'd say his contributions are very significant.
He is not first named on all of them, which means he likely advised masters and PhD/post docs on their work. It’s not uncommon.
This many papers is uncommon, but how it happened is not out of the norm.
Right, I didn't mean to imply that the practice was uncommon, just that using it as a defense of ego so readily was eyebrow-raising. I'm no academic, but I feel like I'd lose respect for my advisor had they used the paper I worked hard on as a way to boost numbers used as personal defense in some petty squabble in a public forum.
Stephen King claims he writes 2000 words a day.
R. L. Stein supposedly wrote a new (admittedly short) novel every two weeks.
This Spanish romance novelist apparently wrote over 4000 novels in her lifetime.
https://en.wikipedia.org/wiki/Cor%C3%ADn_Tellado
So sure, why not 80 technical papers in 2.5 years?
It's easier to write that much if you are just making stuff up...
Successful writers generally don't just make stuff up. They do plenty of research.
You can't just compare creative writing to writing a paper.
True, but you can compare writing 4000 novels a year with being able to write 80 papers a few pages long in 2.5 and say that both are possible.
He didn't write all those papers. He put his name on them. He also finds it worth his time to publicly argue with a pig in shit, so there's that.
The writing of the paper is generally a trivial part of the work. Each technical paper is supposed to be a succinct summary of months or years of technical work.