SoylentNews is people
SoylentNews
Submitted via IRC for SoyCow3941
Academics share machine-learning research freely. Taxpayers should not have to pay twice to read our findings
[...] In my own field of machine learning, itself an academic descendant of Gauss’s pioneering work, modern data are no longer just planetary observations but medical images, spoken language, internet documents and more. The results are medical diagnoses, recommender systems, and whether driverless cars see stop signs or not. Machine learning is the field that underpins the current revolution in artificial intelligence.
Machine learning is a young and technologically astute field. It does not have the historical traditions of other fields and its academics have seen no need for the closed-access publishing model. The community itself created, collated, and reviewed the research it carried out. We used the internet to create new journals that were freely available and made no charge to authors. The era of subscriptions and leatherbound volumes seemed to be behind us.
The public already pays taxes that fund our research. Why should people have to pay again to read the results? Colleagues in less well-funded universities also benefit. Makerere University in Kampala, Uganda, has as much access to the leading machine-learning research as Harvard or MIT. The ability to pay no longer determines the ability to play.
Machine learning has demonstrated that an academic field can not only survive, but thrive, without the involvement of commercial publishers. But this has not stopped traditional publishers from entering the market. Our success has caught their attention. Most recently, the publishing conglomerate Springer Nature announced a new journal targeted at the community called Nature Machine Intelligence. The publisher now has 53 journals that bear the Nature name.
[...] at the time of writing, more than 3,000 researchers, including many leading names in the field from both industry and academia, have signed a statement refusing to submit, review or edit for this new journal. We see no role for closed access or author-fee publication in the future of machine-learning research.
(Score: 4, Insightful) by qzm on Wednesday May 30 2018, @10:44PM (22 children)
Anything else to say?
Perhaps some polite applause?
Pay to view journals with their often bs broken anonymous peer review systems need to die a fast death.
(Score: 0) by Anonymous Coward on Wednesday May 30 2018, @11:04PM (11 children)
If only we had a global network of computers, and a myriad of software packages and protocols for exchanging data.
One day. One day.
(Score: 2) by bob_super on Wednesday May 30 2018, @11:11PM (10 children)
The added value of the Journal model was supposed to be the peer reviews, saving you the time it takes to weed out the crummy research (and fakers) and get access to the good stuff.
Of course, that requires the Journals to do that part correctly, lest their "science" end up at the level of the Twitter feed of a Republican Congresscritter.
(Score: 0) by Anonymous Coward on Wednesday May 30 2018, @11:20PM (9 children)
People can read and sign off on papers.
If there's any virtue in anonymity, then the University's can set up a consortium that will sign publicly anonymous peer-reviewer keys as vouchers for expertise.
Otherwise, anyone can sign a paper, and thereby build trust among his web of buddies.
(Score: 2) by c0lo on Wednesday May 30 2018, @11:54PM (8 children)
Mmmm... I don't think signing is a problem, especially when the signing party is A/C.
Me thinks there are two requirements that need to be fulfilled before signing: independent and review.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 3, Interesting) by takyon on Thursday May 31 2018, @12:21AM (7 children)
You can't necessarily trust that a journal has "independent review", especially if they don't publish reviewer names and comments.
With AC's proposal, you just throw your paper out there and it can be attacked by academics or anybody else. You can give less weight to the anonymous reviewers, but still collect their analyses of the research nonetheless.
Journals already publish in-depth responses/comments to journal articles, so AC's idea is partially implemented in the real world:
Plant diversity increases with the strength of negative density dependence at the global scale [sciencemag.org] (DOI: 10.1126/science.aam5678) (DX [doi.org])
Comment on “Plant diversity increases with the strength of negative density dependence at the global scale” [sciencemag.org] (open, DOI: 10.1126/science.aar2435) (DX [doi.org])
Response to Comment on “Plant diversity increases with the strength of negative density dependence at the global scale” [sciencemag.org] (open, DOI: 10.1126/science.aar3824) (DX [doi.org])
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by c0lo on Thursday May 31 2018, @03:17AM (6 children)
Signature by itself alone this doesn't guarantees independence. And neither does guarantees pertinent reviewing.
And you have a warranty that you won't be flooded with "stellar reviews" by paid-for-reviewers? I mean, look at Yelp and others.
Do you want feedback coming for people on the "my ignorance is as good as your knowledge" position?
Just to be clear in what I'm saying:
1. signing (cryptographic or in any other way) is very likely a necessary part of the "independent review" process. But it alone is not sufficient.
2. far from my minds to say that the way the journals handle it is the best solution. The only warranty they offer is a weakish one: they stake their reputation they better make good if they want to stay in business. But any other better solution will still need to answer to the independent and pertinent review. Which the comment I was replying to does not mention.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 0) by Anonymous Coward on Thursday May 31 2018, @05:23AM (5 children)
You are willfully ignoring part of the OP's point, and therefore you are engaged in a straw man argument.
The comment to which you were replying mentioned webs of trust.
Whether the reviewing was independent, or whether there was reviewing at all is a probability to be calculated by each interested party; such a calculation is based on the interested party's web of trust.
(Score: 2) by c0lo on Thursday May 31 2018, @08:44AM (4 children)
Ah, ignorance! What a blissful life you are living, A/C.
Web-of-trust [wikipedia.org] guarantees that a signing key really corresponds to a a certain person identity. So, it answers to the identity problem and that's about how far it goes.
It does not guarantee out-of-band communication (so that the reviews may be "arranged" instead of fair), it doesn't even cover the pertinence of the reviews (i.e. the reviewer is qualified to offer an relevant opinion about the matter), it doesn't even cover the effort of assessing the relevance of the received reviews (from a possible flood of irrelevant ones).
"For every complex problem, there is solution that is clear, simple and wrong" - in this case, digital signature alone (with or without Web-of-trust).
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 0) by Anonymous Coward on Thursday May 31 2018, @01:01PM (3 children)
Part of maintaining one's web of trust is re-calculating whether an individual should be kept in that web of trust.
In this case, that's a matter of comparing one's expectation of independence/expertise with actual results.
In other words, the reviews are also subject to reviews; it's reviews all the way down UNTIL someone decides that enough reviewing has been done and just accepts a certain web of trust. It's an ongoing, iterative process.
(Score: 2) by c0lo on Thursday May 31 2018, @01:47PM (2 children)
Seems to me you are associating a lot more meaning to Web of trust [wikipedia.org] than the rest of the world familiar with the term.
As such, I find you "willfully missing the point" accusation ill-founded: I can't wilfully miss a point nobody else but you knows about.
This looks like vetting the fitness of someone to a purpose (in particular, the reviewing purpose). Has nothing to do with the common meaning of "web of trust" - I'll be grateful if you use different name for the construct that you have in mind if you'd be inclined to explain.
If you want to go on the path of explaining/detailing your proposal, I'd suggest you to consult the Byzantine fault tolerance [wikipedia.org] in addressing the weakness I seem to detect what you propose on the line of "gaining trust until vetted, misbehaving later". Also, the role of anonymity in lending credence to the independence trait the review process requires (what to do with the case in many reviewers favour or disfavour a certain author or group of them [soylentnews.org])
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 0) by Anonymous Coward on Thursday May 31 2018, @04:28PM
You download software and check to make sure that at least Joe Dev has signed it, because you trust that he is producing high quality software.
Yet, you start to notice that a few glaring bugs have been slipping through, and so you start to have doubts about Joe Dev; maybe, he's having marital troubles, and just isn't devoting as much time to the project as he used to do. So, you decide that you don't really trust him all that much anymore, and therefore remove him from your Web of Trust.
Next time you download an update, your signature-verification software warns: Nobody in your Web of Trust has signed this software; well, now you know that you need to do a little more investigation of your own—either you need to check the changes that have been made, or you need to find someone else whom you can trust to do so.
To borrow a phrase, GET IT YET MUTHAFUCKA?!!!1111
(Score: 0) by Anonymous Coward on Thursday May 31 2018, @04:31PM
An expert registers in the flesh with a consortium of Universities, presenting all the usual Old World criteria (e.g., an expensive piece of paper); the expert generates a one-time private-key for reviewing some article, and then asks the consortium to sign his public key, indicating to the world that his public key has been verified to be that of an expert.
The rest of the world doesn't know how this expert is, but because they trust the consortium and its long history of success, they also accept reviews by that fairly anonymous expert.
Come on. COME ON!
(Score: 0) by Anonymous Coward on Wednesday May 30 2018, @11:17PM
Pay to view journals are not de wey.
Ok I'm a tard.
ICCF Holland link [iccf-holland.org] to atone for my racist sins. What have Emacs users done for Uganda lately?
(Score: 3, Interesting) by unauthorized on Thursday May 31 2018, @01:08AM (7 children)
Come on, that's not a reasonable assessment. Mainstream pay-per-view journals are pretty good at weeding out bullshit, the quality is leagues above free journals and pay2play journals. Sure, occasionally you'll get a weak article published, but I'm yet to see those journals publish troll papers (see this for example [skeptic.com]).
I'm all for transitioning to open journals, but right now traditional publications is the best thing we have. Even when we replace them with something better one day, their historic value will not be diminished - these journals contributed a lot to humanity and for that they deserve to be cherished, even as they become outdated in the march of progress.
(Score: 4, Insightful) by AthanasiusKircher on Thursday May 31 2018, @05:04AM (5 children)
They absolutely do NOT deserve to be "cherished." Many of them once served the greater good but have been co-opted by giant corporate conglomerates that seek outrageous profits for doing next to nothing.
The work is primarily done by editorial boards and peer reviewers, most of whom are academics who work for free. The logistics of copyediting, formatting, and other basic publication matters can often be done for a fraction of the cost of mainstream journals. Publishers exploit the labor of highly skilled academics and play off their past reputations of journals, but the cycle can be broken.
Rather than "cherishing" co-opted corporate monstrosities just because they share a journal name with something that used to be for the greater good, we should encourage those big name academics to take their work and reputation elsewhere, to declare independence [simmons.edu] from the corporations that leech off their work, and found open-access journals that can do the same work for significantly less cost and make research available to all.
(Score: 3, Insightful) by unauthorized on Thursday May 31 2018, @05:48AM (4 children)
If you can't value something despite it's flaws, then you can never value anything because nothing is perfect. There are many wrongs with today's journals, but many of those wrongs are endemic in mainstream western civilization, rather than being specific to the journals. Do you think we should stop valuing culture because corporations like Disney have de facto usurped it? Should we no longer consider the Internet a great thing just because Google's grubby fingers are everywhere and you can't even as much as look up a web page without having to install a mandatory buttplug.js?
Journals still work and provide value, despite being parasitized by talentless leeches. I can't see a good reason why this is not worthy of appreciation.
(Score: 2) by AthanasiusKircher on Thursday May 31 2018, @04:50PM (2 children)
I truly don't understand your point here. I am not arguing and did not argue against journals, only against profiteering journals. As pointed out in discussion below, there are plenty of claimed "open-access" journals that exist as profiteering enterprises too, if not outright scams. I don't value those "journals" either.
And I'm not at all trying to take away from the historical value of historical issues of classic journals. But if said journals have been bought out by a publisher that is now charging ridiculous subscription rates, why should good academics waste their voluntary time being on editorial boards and doing peer review when they could be using their time for journals that don't have the "middleman" skimming large profits off the top?
In a given academic discipline and definitely within an academic subdiscipline, everybody knows the "good" journals -- the ones that have good peer review processes and publish quality research. Everyone knows the relative "pecking order." How do journals get those reputations? By having good academic editors and good editorial boards that run a quality academic enterprise that sifts out the better research from the crap to publish a quality journal. Journal reputations can go up and down over time depending on standards of those who run it -- again, almost always VOLUNTEERING their time.
Those are the people who make the reputation of the journal. Those are the people who maintain it and make it a positive contribution to science. So why should they (or peer reviewers or authors who want to publish) continue to support such an enterprise if it's become infected by profiteering corporate bureaucracy that adds NOTHING to the scientific quality??
Let them do as I say and migrate to a journal with more responsible policies. Because without the editorial board, you know what a journal is? A name. That's it. It has no legitimacy beyond the academics who maintain it. The companies that are making money off of this stuff are simply leeches -- let them have the name, and let the better academics migrate to found new names and create new reputations for better journals in the long-run.
(Score: 1) by unauthorized on Friday June 01 2018, @12:13AM (1 child)
I think I see the point of disagreement clearly now, I myself tend see journals as a holistic institution (ie yhe reviewers are ARE the journal), and you only see them as the discrete publishing entity. Or to put it in other words, it's like the difference of seeing a nation as it's citizenry, rather than purely as the state itself.
(Score: 2) by AthanasiusKircher on Friday June 01 2018, @04:04AM
Okay, I understand that perspective. But note the link I gave in my first reply to you: it was about editorial boards leaving extant journals en masse and forming new journals often with better access and less unnecessary corporate bureaucracy.
That's what I'm advocating for... Effectively preserving extant scholarly communities but migrating the people to better platforms. I don't see the point in continuing to support existing corporate journal infrastructure if they are not responsive to the demands of the academics who run them.
(Score: 0) by Anonymous Coward on Thursday May 31 2018, @05:06PM
you're just another apologist whore. fuck you and your precious information prisons.
(Score: 2) by qzm on Thursday May 31 2018, @06:47AM
The are not bad at weeding out bullshit, but that is not what they claim, they claim they are providing some kind of factual validation.
Unfortunately the fact, which a number of researchers I know see regularly, is the journals are often so specific, a small around of experts at the top
of the field end up reviewing each others work almost all the time, and it becomes a mixture of horse trading and 'a quick chat on the phone, yes, of course I will!'
Hardly what the public thinks of as peer review..
(Score: 2) by Thexalon on Thursday May 31 2018, @02:57AM
Yes, there is: Get me off your fucking mailing list [vox.com].
The only thing that stops a bad guy with a compiler is a good guy with a compiler.