Multi-Loss Fusion: Angular and Contrastive Integration for Machine-Generated Text Detection
Iqra Zahid, Yue Chang, Tharindu Madusanka, Youcheng Sun, Riza Batista-Navarro
Abstract
Modern natural language generation (NLG) systems have led to the development of synthetic human-like open-ended texts, posing concerns as to who the original author of a text is. To address such concerns, we introduce DeB-Ang: the utilisation of a custom DeBERTa model with angular loss and contrastive loss functions for effective class separation in neural text classification tasks. We expand the application of this model on binary machine-generated text detection and multi-class neural authorship attribution. We demonstrate improved performance on many benchmark datasets whereby the accuracy for machine-generated text detection was increased by as much as 38.04% across all datasets.
BibTeX
@inproceedings{zahid-etal-2024-multi,
title = "Multi-Loss Fusion: Angular and Contrastive Integration for Machine-Generated Text Detection",
author = "Zahid, Iqra and
Chang, Yue and
Madusanka, Tharindu and
Sun, Youcheng and
Batista-Navarro, Riza",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-emnlp.421/",
doi = "10.18653/v1/2024.findings-emnlp.421",
pages = "7189--7202"
}