Употреба великих језичких модела у тестирању софтвера
Zbornik radova Fakulteta tehničkih nauka u Novom Sadu · 15.07.2026 · 41 (07) · 816-820
DOI: 10.24867/35oi03fejes ↗Kurz gefasst
Fachartikel in der Zeitschrift Zbornik radova Fakulteta tehničkih nauka u Novom Sadu, online erschienen am 15.07.2026. Verfasst von David Feješ.
Abstract (Originalsprache)
The advancement of artificial intelligence has enabled the broad adaptation of large language models across various domains of software development. These models are increasingly used to support automation, analysis, and evaluation of tests, thereby improving the overall efficiency of software testing. This paper explores the role and potential of large language models in modern software testing, with particular emphasis on their advantages, limitations, and factors that affect reliability. Special attention is given to the issues of consistency and stability, which represent major challenges in real-world problems. This finding suggests that large language models currently provide the greatest value as intelligent support tools that complement existing methods, while their full autonomy in testing remains a subject of ongoing research.
Zitieren
APA
Feješ, D. (2026). Употреба великих језичких модела у тестирању софтвера. Zbornik radova Fakulteta tehničkih nauka u Novom Sadu, 41(07), 816-820. https://doi.org/10.24867/35oi03fejes
Vancouver
Feješ D. Употреба великих језичких модела у тестирању софтвера. Zbornik radova Fakulteta tehničkih nauka u Novom Sadu. 2026;41(07):816-820. doi:10.24867/35oi03fejes
BibTeX
@article{Feje2026,
author = {Feješ, David},
title = {Употреба великих језичких модела у тестирању софтвера},
journal = {Zbornik radova Fakulteta tehničkih nauka u Novom Sadu},
year = {2026},
volume = {41},
number = {07},
pages = {816--820},
doi = {10.24867/35oi03fejes}
}
EUVIMED