“nuclear corrosive” to “nucleic acid”. “False consciousness” instead of “artificial intelligence” … these are the “tortured” expressions that Guillaume Cabanac traces in scientific publications to identify those that are unreliable. The program he co-created for this purpose earned him a place in Nature’s Top 10 Vintage 2021.
Scientific Journal temper nature, One of the oldest and most famous nicknames in the world disappointment. Literally: “to put an end to the renunciation of deception.” His talents for exposing unreliable publications earned Guillaume Cabanac, 39, a teacher and researcher at the Computer Science Research Institute in Toulouse, a place on a list of Ten people who had the greatest impact on science in 2021 According to the review.
Conflicting combinations of words, symptoms of misconduct and fraud going through the filter of some magazine…
In collaboration with Grenoble computer scientist Cyril Labe and Russian mathematician Alexander Magazinov, the program was designed and developed Problem Paper Checker (“The Radar Paper Problem”). Every night, this tool’s torture detector cycles through 120 million publications indexed by the bibliographic database to flush out tormented expressions.
This designation refers to inconsistent combinations of words, symptoms of misconduct and fraud which, unfortunately, pass the filter of some defective or even fraudulent journal, and plague a small part of the scientific literature.
Team of Invisible Detectives
To understand why, you have to put yourself in the shoes of scammers … For them, it is no longer possible to copy simple paste in order to restore paragraphs of serious articles, and initial bricks to create a false one, due to plagiarism detectors.
Paste the result without proofreading and you will see that kidney disappointments thrive instead of kidney failure. Breast risk rather than breast cancer.
Unless … paraphrase the text with synonyms using custom software. Even if it means getting famous torture phrases. Paste the result without proofreading and you will see that kidney disappointments thrive instead of kidney failure. Breast risk rather than breast cancer. Corrosive nuclear material replaces DNA. And fake awareness replaces artificial intelligence!
The trio chose a participatory approach to post-publication re-evaluation. Apparently an invisible college of about fifty “investigators”, researchers and amateurs, defines new tormented expressions that Guillaume Cabanac injects into the program. The articles it contains, which have been flagged as problematic, are then reported on PubPeer.org, a platform run by two CNRS researchers, where scholars from around the world and from all disciplines contribute their expertise by commenting on it for re-evaluation.
Of the 6 million articles in 2021, more than 1,500 contained at least two torturous expressions, a criterion guaranteeing that at least 98% of them contained genuine plagiarism.
Currently, the program tracks 550 tormented expressions in a new Daily Dimensions index of 16,000. Of the nearly 6 million articles published in 2021, it cited more than 1,500 with at least two tormented expressions, a standard that ensures 98% of them are on least contain, after reassessment, real plagiarism. And the phenomenon is increasing, as by the same standards, the numbers drop to 1135 in 2020, 638 in 2019, and 416 in 2018.
In addition, another of their radar detectors has problematic papers, developed in collaboration with Cyril Labbé and Australian researchers, enabling them to point out more than 700 articles on tumors, published between 2014 and 2018, had errors in genetic sequencing To date, it has been cited more than 20,000 times in other scientific articles.
His colleague from Toulouse, CyrilLabbé . often entertains Discover some ‘surreal’ prints. But for Guillaume, discontent often ends in victory … He says. Less than three “pollutions” per 10,000 articles, may not seem like much. corn Do we accept that three out of 10,000 daily flights end in crash?, Guillaume Cabanac insists. The flag is built like a brick house wall. The use of unreliable or even rotten items – and their reuse by the scientific community – could, in the long run, threaten the building! »
Gas scrubbers vs paper mills
This enthusiasm for “cleaning” may have its origins in the researcher’s family and moral values. A grandfather representative of the guild, spurred by the lights of lifelong learning, a Toulouse citizen of humble origins has always been drawn to and believes in culture.The spirit of scienceCloseEthical rules determine the activity of members of the scientific community according to the American sociologist of science Robert K. Merton.especially indifference and organized suspicion.
Due to the multitude of publications, the “paper mills” were trampled by organized indifference and suspicion.
However, due to the large number of publications (which nearly doubled in ten years), “These two guarantees laid down by the sociologist Robert K. Merton are today trampled upon by ‘paper mills’, Kilometer makers steal reliable publications and reformulate them. All in a highly competitive world where researchers are often under the influence of celebrities publish or perish (“Spread or die”) …
At DUT in Management Computing, research has caught the attention of Guillaume Cabanac in earnest. His internship as a programmer with a PhD student at Paul Sabatier University in Toulouse turned out to be a challenging, motivating and grounding experience. He earned his master’s degree at the university and secured funding for his thesis from the ministry on the collective annotation activity of digital texts, in the same laboratory that received him as a lecturer in 2009, nine months after his defense.
When does science fail to correct itself?
Motivated by interdisciplinarity, social bonds, and human relationships, he approaches researchers in the sociology of science and, in 2016, supports a license to conduct direct research entitled “Scientific Text Questioning”. His research, which requires very few resources, reflects its social origins to him: “All I need is a pencil, a computer and an internet connection!” says to himself full stack (“Versatile” in IT parlance): A researcher, he has ideas. The engineer solves problems. Technician, doing programming and getting his hands dirty digitally. As a teacher, he believes that his courses in front of his students fuel his research.
The three investigators were called by dozens of publishing houses to join the Radar paper problem at the start of the publishing chain.
Today, Guillaume Cabanac is an appointed member of the Scientific Council of the CNRS’ Institute of Humanities and Social Sciences. Acts on the protective component to decontaminate inside Franco-Dutch project Nanobubbles ERC For the crucial issue: How, when and why science fails to correct itself? » And the three investigators are called by dozens of publishing houses such as ACM, Elsevier, IEEE, Institute of Physics, SAGE and Springer Nature to incorporate the Radar paper problem into the beginning of the publishing chain.
I saw the current the informationClose“Information” and “epidemic” have shrunk to refer to a deluge of information, many of which turn out to be false or even malicious., The approach of the Toulouse investigator finds more and more echo. Even the French Office of Scientific Integrity (Ofis) is now judging it Post-publish comment activities. located within “The normal activities of the researcher, in the same way as the traditional evaluation activities before peers”, remembers the computer scientist. Excited, he invites everyone to join him to clean up the scientific literature together. ♦
“Music guru. Incurable web practitioner. Thinker. Lifelong zombie junkie. Tv buff. Typical organizer. Evil beer scholar.”