A Study of Reuse and Plagiarism in LREC papers

A Study of Reuse and Plagiarism in LREC papers Gil Francopoulo 1, Joseph Mariani 2, Patrick Paroubek 2 1 LIMSI, CNRS, Université Paris-Saclay + Tagmatica (France) 2 LIMSI, CNRS, Université Paris-Saclay (France) gil.francopoulo@wanadoo.fr, joseph.mariani@limsi.fr, pap@limsi.fr Abstract The aim of this experiment is to present an easy way to compare fragments of texts in order to detect (supposed) results of copy & paste operations between articles in the domain of Natural Language Processing (NLP). The search space of the comparisons is a corpus labeled as NLP4NLP gathering a large part of the NLP field. The study is centered on LREC papers in both directions, first with an LREC paper borrowing a fragment of text from the collection, and secondly in the reverse direction with fragments of LREC documents borrowed and inserted in the collection. Keywords: Plagiarism Detection, Natural Language Processing 1. Introduction Everything starts with a copy & paste and, of course the flood of documents that we see today could not exist without the practical ease of copy & paste. This is not new but what is new is that the availability of archives allows us to study a vast amount of papers in our domain (i.e. Natural Language Processing, NLP, both for written and spoken materials) and to figure out the level of reuse and plagiarism. 2. Context Our work comes after the various studies initiated in the Workshop entitled: “Rediscovering 50 Years of Discoveries in Natural Language Processing” on the occasion of ACL’s 50th anniversary in 2012 [Radev et al 2013] where a group of researchers studied the content of the corpus recorded in the ACL Anthology [Bird et al 2008]. Among these studies, one was devoted to reuse and it is worth quoting Gupta and Rosso [Gupta et al 2012]: “It becomes essential to check the authenticity and the novelty of the submitted text before the acceptance. It becomes nearly impossible for a human judge (reviewer) to discover the source of the submitted work, if any, unless the source is already known. Automatic plagiarism detection applications identify such potential sources for the submitted work and based on it a human judge can easily take the decision”. Let’s add that this subject is a specific and active domain ruled yearly by the PAN international plagiarism detection competition1. 3. Objectives Our aim is not to present the state-of-art or to compare the various metrics and algorithms for reuse and plagiarism detection, see [Hoad et al 2003] [HaCohen-Kerner et al 2010] for instance. In order to avoid any misunderstanding, we position our work as an extrinsic detection, the aim of which is to find near-matches between texts, as opposed to intrinsic detection whose aim is to show that different parts of a presumably single-author text could not have been written by the same 1 http://pan.webis.de author [Stamatatos et al 2011a], [Stein et al 2011], [Bensalem et al 2014]. In contrast, our main objective is to deal with the entry level of the detection. The main question is: Is there a …

Content truncated to 500 words