Russia Develops Unique System for Detecting Plagiarism in 100 Languages

A unique system for tracking down plagiarism of materials by fraudsters has been developed in Russia. Although there are various neural networks around the world that detect plagiarism in scientific works, literary creations, and documents, the Russian system has no analogs. This was reported by the RG newspaper.

The uniqueness of this search system lies in its ability to detect translated segments of plagiarism across any combination of 100 languages.

Yuri Chekhovich, the head of the company that created the system named "Antiplagiate," and a candidate of physical and mathematical sciences, explains that the neural network does not make decisions on its own and does not "accuse" anyone of copying and borrowing other people's texts.

"Essentially, it is a search engine. You upload the text of a thesis or article into the system, and in the end, it delivers the results: if it finds any segment from another source, it will be indicated. However, it cannot determine how correctly the author has utilized that segment, whether it is presented as a citation or as their own writing. Only an expert can do that," Chekhovich clarifies, adding that this way, the tool and the human complement each other in this energy-intensive work.

Every day, the neural network processes up to 500,000 materials, averaging around a dozen per second.

The scientist mentioned that "Antiplagiate" deals with skilled fraudsters who are seriously manipulating other people's texts in an attempt to modify them. The more "advanced" ones utilize foreign language materials, translating and inserting them into their works as if they were their own products.

"We have developed a unique system that checks more than 30 million documents annually. Among them are texts that sometimes contain hundreds or even thousands of pages," emphasized Chekhovich.

The work has been carried out within the framework of the national project "Digital Economy" under a grant program for artificial intelligence amounting to 180 million rubles.