Computer Generated Math Paper makes it into Peer Reviewed Journal | Hacker News Reader

Computer Generated Math Paper makes it into Peer Reviewed Journal

marginalrevolution.com

41 pointsjhull41 comments

It's not accepted. The referees were very kind and gave the benefit of doubt to the authors and gave them the option of revising the paper.

I think the Journal should be lauded for being open minded but principled. I think the authors should fess up and offer the reviewers a beer for wasting their time.

PS. Original post http://thatsmathematics.com/blog/archives/102

PPS. Well, after reading the response letter it seems the Journal was kind of eager to publish. In the Neurosciences I'm used to more guarded language from editors even for papers that have basically been accepted.

The paper was accepted, under condition that these comments are addressed. Considering that the paper contains no meaningful content whatsoever, it should have been outright rejected.

It seems that the journal is one of a number of recent "low tier" (putting it mildly) journals that mainly profit of the processing fee they collect to publish the paper.

I doubt this could have gotten accepted, even conditionally, by a more established journal with a reputation to lose.

Here is the paper in full: http://blog.richmond.edu/physicsbunn/2012/10/19/math-journal...

Well, if they rewrote the abstract, and proved their results, it would have meaningful content..

I've seen it. They'd have to rewrite the abstract, state an actual result, prove it and rewrite everything in between. The "paper" is just grammatically correct but non-sensical glibberish from start to end.

If a referee accepts this paper conditionally, he might as well accept a blank page on condition that the author will put something there.

"Benefit of doubt" contradicts the principle of peer review.

No credible working research mathematician (I am one) could look at that paper for more than fifteen seconds and regard it with anything other than uproarious laughter.

This isn't a pervasive problem in the peer-reviewed math literature. It is a problem with one publisher that is widely regarded to lack legitimacy.

See https://www.google.com/search?q=%22scirp%22+scam for some examples.

Yes, the journal has been labeled a money-making scam: if you need publications (for your academic career), they'll publish your paper in exchange for large publication fees. The "peer-reviews" are basically token reviews. I don't think there's any larger meaning here about the state of mathematics.

See one of the blog post comments for details: http://www.lrb.co.uk/blog/2012/10/17/paul-taylor/stochastica...

(And as random trivia, the four-color theorem was the first major theorem to be proved using a computer. Based on the headline, I was expecting something like that.)

Neither Marcie Rathke nor the University of Southern North Dakota at Hoople is willing to pay the ‘processing charges’ levied by Advances in Pure Mathematics, so we will never know if the work would actually have made it to publication.

-- How much $$, just out of curiosity?

$500, see http://thatsmathematics.com/blog/archives/102

Since publishing is 'valuable' people will appear to offer up 'publiching' as a service. I am surprised that being published in such journals would 'count' in your academic credentials.

The tittle first mislead me because I thought somebody generated a theorem-prover smart enough to find something worth publishing (note: I know nothing about computer generated proofs of theorems).

After reading the "article" I found it obvious it was void of sense. "Let \rho = A", what is A? And who on earth would bother to write 0 as tan(\intfy^{-1}) ?? There are also proofs "left as an exercise to the reader"! Amongs many others those are obvious clues that the article is just a joke.

But thanks to that I know that the journal "Advances in Pure Mathematics" is a scam...

The title is misleading. I came in expecting a proof found by an automated prover being explained by a natural language generator. Instead, it's just another joke paper where meaningless content is generated by some random process.

I expected to see some mathematical theorem unexpectedly proved by Mathematica or something.

How long before an ignorant humanities scholar writes a book claiming that the "Rathke Affair" proves that all math is nonsense?

We have independent verification that math works, because it is used every day in millions of companies to perform a variety of critical tasks.

The targets of the Sokal Affair were on much shakier ground to begin with.

Whether Academia is full of shit is irrelevant to the validity of mathematics. That being said, the GP comment basically demonstrates a peculiar cognitive bias. A variation of a logical fallacy: "I'm not dumb, therefore I never do dumb things." Of which we see all too often grafted into politics: Viz: Science is objective, therefore my (self-interested political) viewpoints are also objective, infallible, and inevitably correct.

The university of southern north Dakota at hoople? Yeah, it doesn't exist. http://en.wikipedia.org/wiki/P._D._Q._Bach

...making the whole thing satire.

The reviewer did not even look at the references... They are certainly the best part.

Not to worry. Someone with a graduate-level background in actual math won't make it very far into the title before deciding the paper is nonsense.

After that, it's all fun and games trying to assign meanings to things. "Arrows of Equations and Problems in Formal Applied PDE" sounds like what happens when someone gets frustrated with their homework to me. ;)

I'm really impressed with the quality of the lorem ipsum, though. If you don't actually read what it says, just kinda skim over the formatting, it's surprisingly believable.

Funny. One would imagine that filtering stuff like this should be very easy, because stuff in math either makes sense, or not, at least to someone knowledgeable in the field.

I propose a simple experiment:

Print out a copy of this paper. Take it to the nearest university campus. Ask math grad students and profs to look it over. I suspect almost all of them will ask if it's some kind of joke, or simply tell you it doesn't make any sense.

That's the filter we expect to apply. But the journal this was submitted to doesn't use a math filter; it uses a "did you pay me $500" filter, which isn't really useful for identifying legitimate mathematics.

I performed a variant of my suggested experiment. I went on facebook and found a friend of mine, who has a B.S. in mathematics and did some coursework toward a Masters.

It took him about 5 minutes to identify it as a hoax. And he was distracted (with facebook, football, and his 3 year old.)

He's how the exchange went:

12:11pm ME: tell me what you think of http://thatsmathematics.com/blog/wp-content/uploads/2012/09/...

12:11pm HIM: How in-depth a review do you want?

12:12pm ME: just give me your first impressions

12:14pm HIM: My first thought: "Holy crap, there's a ton here I don't even remember..."

12:14pm ME: second thought?

12:15pm HIM: Second thought: "Is this legit? It's all over the place..."

12:15pm ME: third thought?

12:16pm HIM: "This is starting to look like something thrown together to appear like a legit math paper."

12:17pm ME: Thanks. That's adequate. It actually is.

12:17pm HIM: Yeah, I was about to say I was 100% certain after one more glance.

EDIT: my wife and I both have advanced math degrees, and we spent the morning laughing at this paper. But since we already knew it was a hoax from the HN post, we weren't valid targets for this experiment. My friend didn't know anything about it before this chat, so he was a valid experimental target. I think his response is adequate to demonstrate that this paper wouldn't pass peer review by actual mathematicians working for a legitimate journal.

we spent the morning laughing at this paper

And how. I still can't breathe quite right. I still think differentiable category theory was my favorite bit, but I'll admit the recent paper by Pythagoras has appeal.

It says "provisionally accepted". What's the provision?

That the only thing left to do is to pay the "processing fees".

Oh, sounds like a very prestigious journal.

http://blog.richmond.edu/physicsbunn/2012/10/19/math-journal...

Rewrite abstract, explain your expressions, prove your results basically.

What should give you pause for thought is not so much the 'quality' of the review but the fact that there is a ready market for such journals. Else they would not exist.

What drives this market? Can it in any sense be said to be good and in public interest?

Reminds me of the story about someone submitting a fake or computer generated paper on Derrida and deconstructionist analysis to a number of literary journals and having it accepted. I can't find the link though, unfortunately.

The Sokal Affair?

http://en.wikipedia.org/wiki/Sokal_affair

Why should we privilege one literary journal over another, anyway?

Can it be that the journal reviews the papers using computer-generated referees?

Yes, a gobbledydook author (the program, not its creator) passed peer review of gobbledygook reviewers.

The insinuation that the mathematical community at large had any involvement in this self-proclaimed journal is false, and then headline is linkbait unworthy of HNm

This finally happened. Narrow specialization at its best.

No, that's not what happened: this journal is a scam, they publish anything vaguely math paper-like in exchange for $500.

This!

It ends up making open access journals look bad because people make the wrong associations.

Then why even bother giving a referee report? But I guess scammers do many strange things...

Because an accepted paper without some report is something that has never been seen and thus highly suspicious. No typos to report? All the prose is perfect and understandable? No missing references? No formatting issues?

This sounds to me like asking "Why would a scammer do something that makes the scam more plausible?". The answer is, of course, to make more people fall for it. Their target is tiny (people that want to write math papers!), so they need to give at least the appearance of being legit.

It depends if the authors are in on the scam or not -- is it scamming authors of their money, or scamming universities who are checking CVs?

You need a 'peer-reviewed' journal for a paper to be relevant in your cv in Spain. And by relevant I mean "it DOES count as a paper". If it is not 'peer-reviewed' it does not count.

(Even if they are cr*p, yes).

You may have proved anything. You may even be Thurston. If you do not have 'papers' (like most of Thurston's second era works), they are NOTHING in Spain.

I guess something similar happens in other countries...

Reply on news.ycombinator.com