Guia Covid-19
Imprimir Republish

Good practices

Nonsense papers

Boas PraticasDaniel buenoThe Springer publishing house has announced the withdrawal of 18 papers published in conference proceedings in the fields of computer science and engineering between 2008 and 2013. Three months ago, the publisher was alerted to the fact that a number of papers submitted to conference proceedings published in Springer publications had been generated by a software program that creates nonsense papers. The following month, it decided to retract the papers—a form of official “de-publication”—rather than simply remove them from its system, because it was the “best available mechanism for correcting the literature and ensuring its integrity.” The publisher will also more closely monitor its review process for papers accepted at conferences. In addition to the Springer-published papers, over 100 papers published by the Institute of Electrical and Electronic Engineers (IEEE), the renowned global institution of technology professionals, were also identified as fake. Among them was a 2013 article about new methodologies for building an e-commerce website. The authors write in the abstract that they “concentrate our efforts on disproving that spreadsheets can be made knowledge-based, empathic and compact,”—a statement that is clearly meaningless. Most of the conferences that accepted fake papers were held in China, and the majority of authors have Chinese affiliation.

The software capable of producing such aberrations is SCIgen, which combines random strings of words to generate fake papers in the field of computer science. The program was invented in 2005 by researchers at the Massachusetts Institute of Technology (MIT) for the purpose of proving that many conferences held in computer science accept papers for publication without proper review. Since SCIgen is available on the Internet, anyone can use it. “I wasn’t aware of the scale of the problem, but I knew it definitely happens,” Jeremy Stribling, who cowrote the software, told the journal Nature in an interview.

The falsifications were reported by Cyril Labbé, a computer scientist and researcher at Joseph Fourier University in France. In 2012, he began the work of cataloging computer-generated papers that had been published in more than 30 conference proceedings over the past five years. Labbé developed a technique for detecting papers created by SCIgen, described in an article published in the journal Scientometrics in 2012. The method involves searching for the characteristic vocabulary produced by the program. Springer has announced a partnership with Labbé to work on developing mechanisms for detecting fake papers generated by SCIgen or similar programs that might emerge in the future.

Republish