[ANPPOM-Lista] 120 computer-generated gobblygook research papers got published in scholarly journals

[Carlos Palombini]

2/27/2014 8:00am by Dr. Mark Thoma, MD<http://americablog.com/author/mark-thoma>

A researcher in France, Cyril Labbe, found a slight problem with some 120
papers that were published either by Springer or the Institute of
Electrical and Electronic Engineers (IEEE).

The thing is, they were all computer-generated fakes.

The papers were made with software that creates random official-looking
research papers on various topics in the computer sciences. The software,
called SCIgen <http://pdos.csail.mit.edu/scigen/>, produces nonsense
results that are presented in correct format, and often have accompanying
tables and graphs. References for the "paper" are also included.

This software was developed at the Massachusetts Institute of Technology
(of course).

And not only did the papers make their ways into various research journals, but
30 of the research
papers<http://www.nature.com/news/publishers-withdraw-more-than-120-gibberish-papers-1.14763>made
their way into various conferences as well.

Springer and IEEE are now in the process of withdrawing those papers. Fraud
in scientific papers is not
new<http://americablog.com/2013/09/trust-scientific-research.html>,
but seems to be gradually increasing. This, however, is a huge number to
have been released at the same time. We will have to read papers more
skeptically in the future.

Just to show you how these "papers" can be created, I generated several
"written" by some very famous authors. Here are the intros of each article,
then at the bottom of the page I include an entire "research paper" that
the computer generated, so you can see what the entire thing looks like.
It's a pretty amazing fake.

http://americablog.com/2014/02/120-computer-generated-gobblygook-research-papers-published-scholarly-journals.html
-- 
carlos palombini
professor de musicologia ufmg
proibidao.org
ufmg.academia.edu/CarlosPalombini
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://www.listas.unicamp.br/pipermail/anppom-l/attachments/20140227/8a0bf61f/attachment.html>