Authorship Verification in software engineering education
ASCILITE Publications · 28.11.2023 · 195-204
DOI: 10.14742/apubs.2023.559 ↗Kurz gefasst
Fachartikel in der Zeitschrift ASCILITE Publications, online erschienen am 28.11.2023. Verfasst von Shannon Rios und 2 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The prevalence of academic misconduct, specifically contract cheating, is a rising concern in higher education institutions globally. Among the recent advancements, Generative Artificial Intelligence (genAI) has exacerbated the situation by offering authentically generated writings, making detection through traditional plagiarism tools ineffective. This paper explores the development and application of students' academic writing profiles, using a combination of word embedding (Word2Vec) and stylistic feature extraction techniques. By leveraging a Siamese neural network, our method focuses on recognising distinctive writing styles, a concept rooted in Authorship Verification (AV). Our approach's efficacy evaluates favourably against other AV methods and is tested against AI-generated texts deliberately designed to mimic student writing. The study emphasises the importance of understanding individual academic writing styles to identify outsourcing or AI-generated work effectively.
Zitieren
APA
Rios, S., Zhang, Y., & Oliveira, E. (2023). Authorship Verification in software engineering education. ASCILITE Publications, 195-204. https://doi.org/10.14742/apubs.2023.559
Vancouver
Rios S, Zhang Y, Oliveira E. Authorship Verification in software engineering education. ASCILITE Publications. 2023:195-204. doi:10.14742/apubs.2023.559
BibTeX
@article{Rios2023,
author = {Rios, Shannon and Zhang, Yu and Oliveira, Eduardo},
title = {Authorship Verification in software engineering education},
journal = {ASCILITE Publications},
year = {2023},
pages = {195--204},
doi = {10.14742/apubs.2023.559}
}
EUVIMED