Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts
Purpose. The intelligent search engine Bing can be used as a method and a means of determining the author of a Ukrainian-language test. Bing helps to find information about a text fragment and its author, but the search results may be inaccurate or incomplete. The main purpose of the paper is to stu...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Dnipro National University of Railway Transport named after Academician V. Lazaryan
2023-06-01
|
| Series: | Nauka ta progres transportu |
| Subjects: | |
| Online Access: | http://stp.diit.edu.ua/article/view/288289 |
| _version_ | 1797354088825880576 |
|---|---|
| author | O. P. Ivanov V. I. Shynkarenko V. V. Skalozub A. A. Kosolapov |
| author_facet | O. P. Ivanov V. I. Shynkarenko V. V. Skalozub A. A. Kosolapov |
| author_sort | O. P. Ivanov |
| collection | DOAJ |
| description | Purpose. The intelligent search engine Bing can be used as a method and a means of determining the author of a Ukrainian-language test. Bing helps to find information about a text fragment and its author, but the search results may be inaccurate or incomplete. The main purpose of the paper is to study the effectiveness of establishing the authorship of literary texts by state-of-the-art artificial intelligence tools based on ultra-short excerpts. Methodology. Ten Ukrainian authors with a rich body of fiction reflecting various aspects of Ukrainian culture and history were selected, as well as random fragments of 3–7 words each from different works of these authors. An experiment was conducted to determine the authorship of 2,000 fragments. Findings. Using the Python programming language and the skpy package, we developed software that sends questions and receives answers from the Bing bot built into Microsoft Skype. The answers were checked for the name of the author of the phrase and the corresponding title of the work. According to the results, Ivan Franko has the highest percentage of answers where the author's name was mentioned (65%), and Oleksandr Dovzhenko has the lowest result (23%). The answers were analyzed by the length of the fragments. Of course, the longer the length of a text fragment, the greater the likelihood of accurately identifying its authorship. Features of the author's style are manifested in 20–40 % of short fragments. The remaining 60–80% may be commonly used language constructions that the author relayed from the external environment. Originality. In this work, for the first time, the method of checking the authorship of fragments of Ukrainian-language text using the Bing bot with artificial intelligence is presented. A comparative analysis was performed and experiments were given to determine the authorship of short fragments of 3–7 words. It has been established that even quite small fragments of the text have signs characteristic of the original style of the author of artistic works. Practical value. It has been determined to what extent experts in determining the authorship of natural language texts can rely on existing state-of-the-art artificial intelligence tools in combination with an extensive database of texts in the Internet space. |
| first_indexed | 2024-03-08T13:40:14Z |
| format | Article |
| id | doaj.art-61a73a92e69045b3b9b0955216d47054 |
| institution | Directory Open Access Journal |
| issn | 2307-3489 2307-6666 |
| language | English |
| last_indexed | 2024-03-08T13:40:14Z |
| publishDate | 2023-06-01 |
| publisher | Dnipro National University of Railway Transport named after Academician V. Lazaryan |
| record_format | Article |
| series | Nauka ta progres transportu |
| spelling | doaj.art-61a73a92e69045b3b9b0955216d470542024-01-16T11:34:36ZengDnipro National University of Railway Transport named after Academician V. LazaryanNauka ta progres transportu2307-34892307-66662023-06-012(102)455310.15802/stp2023/288289326560Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short ExcerptsO. P. Ivanov0https://orcid.org/0000-0003-1259-6377V. I. Shynkarenko1https://orcid.org/0000-0001-8738-7225V. V. Skalozub2https://orcid.org/0000-0002-1941-4751A. A. Kosolapov3https://orcid.org/0000-0001-8878-568XUkrainian State University of Science and TechnologiesUkrainian State University of Science and TechnologiesUkrainian State University of Science and TechnologiesUkrainian State University of Science and TechnologiesPurpose. The intelligent search engine Bing can be used as a method and a means of determining the author of a Ukrainian-language test. Bing helps to find information about a text fragment and its author, but the search results may be inaccurate or incomplete. The main purpose of the paper is to study the effectiveness of establishing the authorship of literary texts by state-of-the-art artificial intelligence tools based on ultra-short excerpts. Methodology. Ten Ukrainian authors with a rich body of fiction reflecting various aspects of Ukrainian culture and history were selected, as well as random fragments of 3–7 words each from different works of these authors. An experiment was conducted to determine the authorship of 2,000 fragments. Findings. Using the Python programming language and the skpy package, we developed software that sends questions and receives answers from the Bing bot built into Microsoft Skype. The answers were checked for the name of the author of the phrase and the corresponding title of the work. According to the results, Ivan Franko has the highest percentage of answers where the author's name was mentioned (65%), and Oleksandr Dovzhenko has the lowest result (23%). The answers were analyzed by the length of the fragments. Of course, the longer the length of a text fragment, the greater the likelihood of accurately identifying its authorship. Features of the author's style are manifested in 20–40 % of short fragments. The remaining 60–80% may be commonly used language constructions that the author relayed from the external environment. Originality. In this work, for the first time, the method of checking the authorship of fragments of Ukrainian-language text using the Bing bot with artificial intelligence is presented. A comparative analysis was performed and experiments were given to determine the authorship of short fragments of 3–7 words. It has been established that even quite small fragments of the text have signs characteristic of the original style of the author of artistic works. Practical value. It has been determined to what extent experts in determining the authorship of natural language texts can rely on existing state-of-the-art artificial intelligence tools in combination with an extensive database of texts in the Internet space.http://stp.diit.edu.ua/article/view/288289authorship detectionnatural language textartificial intelligencegenerative language modelschatgptbing botskypemicrosoftbardgoogle |
| spellingShingle | O. P. Ivanov V. I. Shynkarenko V. V. Skalozub A. A. Kosolapov Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts Nauka ta progres transportu authorship detection natural language text artificial intelligence generative language models chatgpt bing bot skype microsoft bard |
| title | Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts |
| title_full | Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts |
| title_fullStr | Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts |
| title_full_unstemmed | Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts |
| title_short | Determining the Authorship of a Ukrainian-Language Literary Text by Means of Artificial Intelligence from Ultra-Short Excerpts |
| title_sort | determining the authorship of a ukrainian language literary text by means of artificial intelligence from ultra short excerpts |
| topic | authorship detection natural language text artificial intelligence generative language models chatgpt bing bot skype microsoft bard |
| url | http://stp.diit.edu.ua/article/view/288289 |
| work_keys_str_mv | AT opivanov determiningtheauthorshipofaukrainianlanguageliterarytextbymeansofartificialintelligencefromultrashortexcerpts AT vishynkarenko determiningtheauthorshipofaukrainianlanguageliterarytextbymeansofartificialintelligencefromultrashortexcerpts AT vvskalozub determiningtheauthorshipofaukrainianlanguageliterarytextbymeansofartificialintelligencefromultrashortexcerpts AT aakosolapov determiningtheauthorshipofaukrainianlanguageliterarytextbymeansofartificialintelligencefromultrashortexcerpts |