Używamy cookies i podobnych technologii m.in. w celu świadczenia usług i w celach statystycznych. Możesz określić warunki przechowywania lub dostępu do plików cookies w Twojej przeglądarce, w jej ustawieniach. Jeżeli wyrażasz zgodę na zapisywanie informacji zawartej w cookies, kliknij „Zamknij”. Jeżeli nie wyrażasz zgody – zmień ustawienia swojej przeglądarki.Więcej informacji znajdziesz w naszej Polityce cookies.

Zamknij informację o cookies
.

National Information Processing Institute

JSA has already checked 40 thousand theses this year

The Single Anti-Plagiarism System (JSA) checks the similarity of theses to be submitted in the defence process to texts, essays and other written works developed earlier. It does not determine whether a particular thesis is a plagiarism – it serves as a tool supporting the verification work of academic supervisors.

Although the system’s overload peak, i.e. summer exam session, is still ahead of us, since the beginning of the year, 40 thousand theses have already been checked by the Single Anti-Plagiarism System. 29 thousand people created accounts in the system which means – adding individuals using the JSA via university systems – that there are already over 35 thousand users logged in the system. 27.5 thousand of them are thesis supervisors.

On average, one user spent 12 minutes in the system, divided into three 4-minute sessions. One checking of thesis equals 1.058 attempts which means that in case of majority of theses users deemed one check to be sufficient. The average waiting time for the report on the thesis check was 7,5 minutes. The JSA is being constantly updated – in line with remarks submitted by the system’s users. The current version can process up to 6 thousand checks per 24 hours – the waiting time for the report is up to 10 minutes.

Current JSA users represent 216 universities (59 universities use JSA via own university systems, the others use the user interface offered by JSA). The most of theses for plagiarism-check were downloaded in the JSA by users from the Poznań University of Technology, the Ignacy Łukasiewicz Rzeszów University of Technology, the Warsaw University of Technology, the Tadeusz Kościuszko Cracow University of Technology, and the Gdańsk University of Technology (from 1,000 up to 2,500 theses per university).  

When it comes to the results of thesis checking in the JSA, for one thesis the system found on average 300 documents which could be the source of plagiarism. Combined, it gives 9 million pairs of documents which were cross-checked by computational modules using machine learning or natural language processing methods.  

The JSA system was designed and developed by the National Information Processing Institute upon instructions from the Ministry of Science and Higher Education, and financed from the funds of Knowledge Education Development Programme (POWER). Based on the amendments to the Higher Education Act every written thesis has to be checked for plagiarism in the Single Anti-Plagiarism System before a diploma exam. The system is accessible via the Internet for thesis supervisors and academic administration employees (https://jsa.opi.org.pl). The theses are checked against the National Repository of Theses (ORPPD), Polish semantic search engine Nekst (information on all Polish texts on the Internet), and other sources (6 Wikipedias, the ISAP – online system of legal acts, rulings).

Currently, the system works in a cluster of a hundred servers managing over 30 terabytes of data which makes it the biggest system of the Ministry of Science and Higher Education, both in terms of number of end users, and number of servers used by a single system, as well as in terms the size of data being processed.