Slashdot Mirror

Why not create and manage your own open journal covering machine intelligence? (obligatory: with blackjack, and hookers)

They did. Here it is: Journal of Machine Learning Research. Re-read the open letter. The editors of the previous paywalled journal quit to create it. It's been going on for almost 2 decades now.

In just one journal (Journal of Machine Learning Research), there is hundreds of papers. 2 anecdotes is insignificant. This guy have no shame.

In just one journal (Journal of Machine Learning Research), there is hundreds of papers. 2 anecdotes is insignificant. This guy have no shame.

Yeah, the standard solution in the humanities right now is basically a person who will laboriously reformat your Word document into the journal's layout, usually using something like Adobe InDesign.

This often produces nice-looking typesetting done by a professional, but it's expensive and one barrier to making things go open-access. If you're in a field where you can expect people to do their own typesetting in LaTeX, then you can run a no-expenses journal.

The Bayesian analysis in spam filters only works on text. Spammers realized that they could get around it by filling the text portion of the message with some random passage from a Project Gutenberg file, thus making it seem innocuous, and then putting the real advertisement in a GIF or PNG file that would be displayed by HTML-capable mail readers. Bayesian analysis can still work, but only in combination with OCR software.

Bayesian filters (and other statistical filters colloqually known as Bayesian) can work on any features at all; not necessarily text. In particular they can use the markup in the header of the message, the message encoding, and so on. Some of the best-performing filters don't use 'text' at all and simply treat the entire message, images and all, as a bit string; for example, compression-based filters. Another well performing filter, OSBF-Lua, uses orthogonal sparse binomial bigrams rather than individual tokens.

Recent standardized testing shows that these methods work just fine on image spam, without any OCR component.

IANAS, but I imagine that such review might take some $ to accomplish.

IAAS, and peer review takes zero $ to accomplish. The action editor (who works for love) emails the article (in PDF) to the reviewers (who work for love), who email their reviews back, whereupon the action editor makes the call - publish, revise, or reject. The publishers do not put any money into that system, and have indeed been scamming the public for years.

Establishing a solid reputation, quality control, and peer review process are challenges for any new journal, whether online or not. It can be done though - for instance, the free online Journal of Machine Learning Research has within the few years of its existence rocketed to the top of the journal citation reports in its field.

Fortunately there is a middle way between traditional academic publishers and author self-publication on the web: online academic journals run by the scientists themselves. A good example in my field is the Journal of Machine Learning Research which was formed when the entire editorial board resigned from the overpriced Machine Learning Journal. Online access is free, while a low-price is published to satisfy the legacy requirements of libraries, copyright law, tenure review committees, and the like. Speaking of copyright: in contrast to traditional publishers, authors do not have to sign away their copyright when they publish in JMLR.

The result? Three years after its inception, JMLR is the highest-impact journal in Artificial Intelligence. This is by no means an isolated case, but part of a sea change in academic publishing. More and more such journals are being setup, often in direct competition to overpriced conventional publications, and with support from academic libraries.

The "author pays" model is a last-ditch effort by traditional academic publishers to wring profits from scientific communication, an activity that in essence has always been free (as in -dom). Apparently they haven't noticed yet that all the scarcities that their business model depended on - from trees to typesetting to transport - have simply been removed by technology. Given the free volunteer labor that scientists routinely provide, and the existing host infrastructure at the institutions where they work, the cost of running an online scientific journal is, for all practical purposes, zero.

This is already happening in Artificial Intelligence. The Journal of AI Research (JAIR), and The Journal of Machine Learning Research (JMLR) are peer-reviewed journals published on the web for free.

I'm not sure what the $20 million is for, since (at least in AI) peer-review is done for free anyway, as a service to the community. The big journals charge money while getting editing, review, and often even typsetting for free from their editorial boards or authors.

Since peer-review is the main service provided by the big journals, it was only a matter of time before the reviewers organized themselves. The tenure issue is a bit of a problem, since untenured faculty will want to publish in the best established journals. However, that should work itself out over time, as tenured researchers choose to publish in the new free journals. Eventually the new journals will be well enough established for young researchers to feel comfortable publishing in them.

That is why the people in the article have quit JML and started JMLR. There, you can view every paper for free, the turn around time from submittal to publication is potentially less than that of a conference since all papers are posted on the web site immediately after reviewers give the OK and you dont need to get any copy rights from publishers(just the authors of the paper) since the author retains the copyrights.

I think that you and the original poster are misinterpreting what is going on. What the letter suggests is not to remove the peer review process, but to remove the process of printing the journal and mailing it to people. For instance, would you prefer to pay money for the IEEE journals you now get in hard copy form, or pay nothing and simply download all the articles off of a web site and print them yourself? The same articles, published by the same people and reviewed by the same reviewers. Nothing about the process changes except the distribution medium and who holds the copyright.

In fact, with the new journal the reviewers are moving to(JMLR) you can still pay money for a hard copy if you prefer. The idea is that you dont have to pay any money if you dont want to, and authors dont lose the copyrights if they dont want to.

The peer reviewers in this article who "quit" haven't said they want to stop reviewing or do away with peer review...

In fact, if you compare the editorial board of JMLR with the people that resigned from JML, you will see that many of them are now just reviewing for the new free online journal.

Press Release on the Journal of Machine Learning Research site gives a partial answer to the question how they will finance this. In addition of getting money from printed version and paid electronic edition (with additional features) on the CatchWord, SPARC helps them. SPARC is an alliance of universities and research libraries that supports increased competition in scientific journal
publishing.