Detecting AI-Generated Bulgarian Text: A Two-Step Multi-Class Classification Approach
Computational Linguistics in Bulgaria · 2026-08 · 2 (1) · 4-21
DOI: 10.47810/jclib.2.2026.01 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computational Linguistics in Bulgaria, 2026-08. Verfasst von Boyan Bogdanov und 4 weiteren Autor:innen.
Abstract (Originalsprache)
This paper introduces a novel two-step multi-class classification system to identify varying degrees of machine involvement in Bulgarian text. As Large Language Models (LLMs) proliferate, distinguishing original human writing from machine-assisted or machine-generated content is crucial to prevent misinformation and preserve educational integrity. We developed a comprehensive dataset in Bulgarian encompassing purely human-written texts and four distinct mixed-content categories, such as machine-continued and machine-polished text. Our proposed pipeline utilises a binary ensemble classifier combining stylometric features, XLM-RoBERTa, and an adapted Binoculars model to first filter out human-written text. Subsequently, a specialised multi-class Support Vector Machine categorises the remaining machine-involved texts. Our findings indicate that this hierarchical approach improves the reliable identification of human authorship and enhances the overall accuracy of multi-class text detection compared to single-step methods.
Zitieren
APA
Bogdanov, B., Georgiev, D., Dimitrov, D., Koychev, I., & Nakov, P. (2026). Detecting AI-Generated Bulgarian Text: A Two-Step Multi-Class Classification Approach. Computational Linguistics in Bulgaria, 2(1), 4-21. https://doi.org/10.47810/jclib.2.2026.01
Vancouver
Bogdanov B, Georgiev D, Dimitrov D, Koychev I, Nakov P. Detecting AI-Generated Bulgarian Text: A Two-Step Multi-Class Classification Approach. Computational Linguistics in Bulgaria. 2026;2(1):4-21. doi:10.47810/jclib.2.2026.01
BibTeX
@article{Bogdanov2026,
author = {Bogdanov, Boyan and Georgiev, Daniel and Dimitrov, Dimitar and Koychev, Ivan and Nakov, Preslav},
title = {Detecting AI-Generated Bulgarian Text: A Two-Step Multi-Class Classification Approach},
journal = {Computational Linguistics in Bulgaria},
year = {2026},
volume = {2},
number = {1},
pages = {4--21},
doi = {10.47810/jclib.2.2026.01}
}
EUVIMED