|
This article is cited in 1 scientific paper (total in 1 paper)
SPECIAL ISSUE: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING TECHNOLOGIES
Artificially generated text fragments search in academic documents
G. M. Gritsaiab, A. V. Grabovoyabc, A. S. Kildyakova, Yu. V. Chekhovichac a Antiplagiat Company, Moscow, Russia
b Moscow Institute of Physics and Technology (National Research University), Moscow, Russia
c Federal Research Center "Computer Science and Control" of Russian Academy of Sciences, Moscow, Russia
Abstract:
Recent advances in text generative models make it possible to create artificial texts that look like human-written texts. A large number of methods for detecting texts obtained using large language models have already been developed. But the improvement of detection methods occurs simultaneously with the improvement of generation methods. Therefore, it is necessary to explore new generative models and modernize existing approaches to their detection. In this paper, we present a large analysis of existing detection methods, as well as a study of lexical, syntactic and stylistic features of the generated fragments. Taking into account the developments, we have tested the most qualitative, in our opinion, methods of detecting machine-generated documents for their further application in the scientific domain. Experiments were conducted for Russian and English languages on the collected datasets. The developed methods improved the detection quality to a value of 0.968 on the F1-score metric for Russian and 0.825 for English, respectively. The described techniques can be applied to detect generated fragments in scientific, research and graduate papers.
Keywords:
machine-generated text, natural language processing, multiple hypothesis testing, paraphrasing, detection of generated texts.
Citation:
G. M. Gritsai, A. V. Grabovoy, A. S. Kildyakov, Yu. V. Chekhovich, “Artificially generated text fragments search in academic documents”, Dokl. RAN. Math. Inf. Proc. Upr., 514:2 (2023), 308–317; Dokl. Math., 108:suppl. 2 (2023), S434–S442
Linking options:
https://www.mathnet.ru/eng/danma475 https://www.mathnet.ru/eng/danma/v514/i2/p308
|
Statistics & downloads: |
Abstract page: | 50 | References: | 5 |
|