This Is Uncanny

Since we're all talking about plagiarism now, I'd like to share this video which came out last year about a paper accepted at the CVPR 2022:

For the people not in the know, the Computer Vision and Pattern Recognition conference is the biggest conference in computer science. Last year, in 2022, the paper featured in the video got accepted. A few days later, this video was posted. The first author, a PhD student, apologized and the paper was retracted and removed from the proceedings. Hilariously, the first reaction of the co-authors, including a professor at the Seoul National University, was to say that they had nothing to do with it.

My point here is that scientific papers are not rigorously checked for plagiarism, and a background in academia tells you absolutely nothing about whether or not someone will be diligent in avoiding plagiarism. The biggest difference is that there are consequences if you're caught.

I also don't want people to be too harsh on the first author of this paper, or to think the situation is equivalent to the whole Somerton debacle. For starters, you don't get paid for publishing papers, you (or more commonly your university) pay the publishers. But the phrase publish or perish exists for a reason, and everyone in the field wants to get published in the CVPR, because it's supposed to show that you're great at research. Additionally, the number of papers and the prestige of the venues they're published in criteria on which you will be evaluated as a researcher and a university employee.

The way I see it, there are basically two kinds of plagiarism that are shown in the video. The first one concerns sentences that are lifted completely unchanged from other papers. This is bad, and it is plagiarism, but I can see how this would happen. Most instances of this appear in the introduction and on background information, so if you're insecure about your mastery of English and it's not about your contribution anyway, I can understand how you would take the shortcut of copy-pasting and tell yourself that it's just so that the rest of the paper makes sense, and why waste time on phrasing things differently if others have done it already, and it's not like there are a million way to write these equations anyways.

Let me be clear. I don't approve, or condone. It's still erasing the work of the people who took the time and pain to phrase these things. It's still plagiarism. But I understand how you could get to that point.

The second kind of plagiarism is a way bigger deal in my opinion. At 0:37 , we can see that one of the contributions of the paper is also lifted from another paper. Egregiously, the passage includes "To the best of our knowledge, this is the first [...]" , which is a hell of a thing to copy-paste. So this is not only lazily passing other people's words as your own, it's also pretending that you're making a contribution you damn well know other people have already done. I also wasn't able to find a version of the plagiarized article that had been published in a peer-reviewed venue, which might mean that the authors submitted it, got rejected, and published it on arXiv (an website on which authors can put their papers so that they're accessible to the public, but doesn't "count" as a publication because it's not peer-reviewed. You can also put papers that are under review or have been published on there as long as you're careful with the copyrights and double-blind process). And then parts of it were published in the CVPR under someone else's name.

I think there's also a third kind of plagiarism going on here, one that is incredibly common in academia, but that is not shown in the video. That's the FIVE other authors, including a professor, who were apparently happy to add their name to the paper but obviously didn't do anything meaningful since they didn't notice how much plagiarism was going on.

John: Arthur it’s absolutely necessary you’re naked for the next part … Uhm because… danger.

Yes.

Danger…

👀

@oloreandil et moi étions sur la même longueur d'onde, @sometimes-gloriousstudent a confirmé, j'étais donc obligée :

I've started Master and Commander, and 2 hours in I have the following to report:

They're doing the fastest Enemies to Friends to Shipmate Speedrun.

Each of them infodumping in turn, then talking about music.

They have the following DELIGHTFUL exchange, about Castillan and Catalan (as best as I can remember it):"But they're quite similar? A putain, as they say in French? - Oh non, they're different languages. And a patois, if you please. - Oh ? I'm sure the other word exists, I think I've heard it."

A guy who Jack is talking to for the first time snitches on another by saying that he's gay. Then asks what Jack thinks of this "buggery business".

For the record, Jack is against it, but doesn't like to see a man hang for it. It turns out that Lt Snitch doesn't really have anything concrete, it's more of a vibe-based accusation.

I'm letting all nautical talk gently flow through me without making any impression. I think Maturin is doing the same.

I have some extra audible credits. Would you recommend using them on the Aubrey-Maturin serie?

always

And for the record, it does go perfectly.

Listening to these people talking themselves into doing the one thing they've been warned not to do is HILARIOUS. Well, maybe we SHOULD go down Copland Road...

Is this anything?

there was something at about the lighting and stainless steel while i was cleaning up at work that made me go WAIT HOLD UP GOTTA GRAB ME CAMERA

fellow goth transmacs do you ever read monstrous regiment, 31st novel in the discworld series by the late sir terry pratchett, and become consumed by thoughts of renaming yourself "maladict" every waking moment since putting the book down