Slide
Centre Interdisciplinaire
de Recherche et d’Innovation
en Cybersécurité et Société
de Recherche et d’Innovation
en Cybersécurité et Société
1.
Ngouanfouo, C.; Davoust, A.
Detecting Machine-Generated Text using Grammatical Features Proceedings Article
In: Proc. Int. Conf. Tools Artif. Intell. ICTAI, pp. 843–848, IEEE Computer Society, 2025, ISBN: 10823409 (ISSN); 979-833154919-0 (ISBN).
Abstract | Links | BibTeX | Tags: AI Text Detection, CNN, Computational grammars, Detection methods, Language model, Machine-generated texts, Natural language generation, Natural language processing systems, Neural encoding, Neural modelling, Part Of Speech, Part-of Speech, Speech communication, Text detection, Written texts
@inproceedings{ngouanfouo_detecting_2025,
title = {Detecting Machine-Generated Text using Grammatical Features},
author = {C. Ngouanfouo and A. Davoust},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-105031903675&doi=10.1109%2FICTAI66417.2025.00123&partnerID=40&md5=5783b8797a3425f9dfa737343ee757d2},
doi = {10.1109/ICTAI66417.2025.00123},
isbn = {10823409 (ISSN); 979-833154919-0 (ISBN)},
year = {2025},
date = {2025-01-01},
booktitle = {Proc. Int. Conf. Tools Artif. Intell. ICTAI},
pages = {843–848},
publisher = {IEEE Computer Society},
abstract = {Large Language Models (LLMs) have advanced natural language generation but pose ethical and practical challenges, making it crucial to detect machine-generated texts. Traditional detection methods rely on complex, hard-to-interpret neural encodings and model-specific features like perplexity. This study explores whether grammatical patterns-specifically sequences of parts of speech (POS), including punctuation and symbols-can distinguish machine-written texts from human ones. Using a CNN classifier on POS sequences, the approach achieves nearly 90 % accuracy on a benchmark dataset. Combining POS-based features with neural embeddings improves performance, and the model shows robustness against adversarial attacks, though it is less effective on short texts. © 2025 IEEE.},
keywords = {AI Text Detection, CNN, Computational grammars, Detection methods, Language model, Machine-generated texts, Natural language generation, Natural language processing systems, Neural encoding, Neural modelling, Part Of Speech, Part-of Speech, Speech communication, Text detection, Written texts},
pubstate = {published},
tppubtype = {inproceedings}
}
Large Language Models (LLMs) have advanced natural language generation but pose ethical and practical challenges, making it crucial to detect machine-generated texts. Traditional detection methods rely on complex, hard-to-interpret neural encodings and model-specific features like perplexity. This study explores whether grammatical patterns-specifically sequences of parts of speech (POS), including punctuation and symbols-can distinguish machine-written texts from human ones. Using a CNN classifier on POS sequences, the approach achieves nearly 90 % accuracy on a benchmark dataset. Combining POS-based features with neural embeddings improves performance, and the model shows robustness against adversarial attacks, though it is less effective on short texts. © 2025 IEEE.
2.
Remacle, A.; Bouchard, S.; Morsomme, D.
In: Computers and Education, vol. 200, 2023, ISSN: 03601315 (ISSN), (Publisher: Elsevier Ltd).
Abstract | Links | BibTeX | Tags: Background noise, Computer aided instruction, Control groups, E-learning, Experimental groups, Oral communication, Oral communication skills, Personnel training, randomized controlled trial, Self Efficacy, Speech communication, Teacher training, Teachers', Virtual Classroom, virtual reality, Voice
@article{remacle_can_2023,
title = {Can teaching simulations in a virtual classroom help trainee teachers to develop oral communication skills and self-efficacy? A randomized controlled trial.},
author = {A. Remacle and S. Bouchard and D. Morsomme},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85152896047&doi=10.1016%2fj.compedu.2023.104808&partnerID=40&md5=990f5fff277e3edf28e78b1fab80022c},
doi = {10.1016/j.compedu.2023.104808},
issn = {03601315 (ISSN)},
year = {2023},
date = {2023-01-01},
journal = {Computers and Education},
volume = {200},
publisher = {Elsevier Ltd},
abstract = {Effective oral communication skills are essential to ensure optimal teaching while preserving the teacher's vocal health. Training these skills in representative settings is expected to promote their generalization. Since the implementation of such training in actual school situations is challenging, virtual reality (VR) may represent a solution.This study evaluated the effects of VR simulations on trainee teachers’ oral communication skills. Based on our Theoretical Framework for Teachers’ Vocal Behavior, we developed and empirically assessed a voice-related prevention program including noisy communicative situations in a virtual classroom. In a randomized controlled trial, the participants were assigned to one of two conditions: (1) individual voice training including simulations in the virtual classroom and a group information session (experimental group},
note = {Publisher: Elsevier Ltd},
keywords = {Background noise, Computer aided instruction, Control groups, E-learning, Experimental groups, Oral communication, Oral communication skills, Personnel training, randomized controlled trial, Self Efficacy, Speech communication, Teacher training, Teachers', Virtual Classroom, virtual reality, Voice},
pubstate = {published},
tppubtype = {article}
}
Effective oral communication skills are essential to ensure optimal teaching while preserving the teacher's vocal health. Training these skills in representative settings is expected to promote their generalization. Since the implementation of such training in actual school situations is challenging, virtual reality (VR) may represent a solution.This study evaluated the effects of VR simulations on trainee teachers’ oral communication skills. Based on our Theoretical Framework for Teachers’ Vocal Behavior, we developed and empirically assessed a voice-related prevention program including noisy communicative situations in a virtual classroom. In a randomized controlled trial, the participants were assigned to one of two conditions: (1) individual voice training including simulations in the virtual classroom and a group information session (experimental group