EUVIMEDEUVIMED entdecken
Crossref

Multi-level Style Preference Optimization: An Adaptive Detection Framework for Human-Machine Hybrid Text

Zehao Wang, Lianwei Wu, Wenbo An, Hang Zhang, Yaxiong Wang

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (40) · 33746-33754

DOI: 10.1609/aaai.v40i40.40665 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Zehao Wang und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language model (LLM) generated texts now rival human quality, creating four text categories: purely machine-generated, machine-rewritten, machine-polished, and human-written content. Traditional detection methods face significant challenges in human-machine hybrid scenarios where LLMs perform rewriting or polishing, as existing approaches focus on single-level features and fail to capture subtle, multi-layered machine traces. To address this, we propose the Multi-level Style Preference Optimization (MSPO) framework, capturing machine style features at multiple granularities: sequence-level (overall consistency), phrase-level (distinctive n-gram patterns), and lexical-level (word selection distributions). We further incorporate four text complexity indicators (Type-Token Ratio, Average Sentence Length, Average Word Length, and Punctuation Ratio) to dynamically adjust optimization parameters based on human-machine text complexity differences, enhancing adaptability across diverse text types. Additionally, we construct a comprehensive detection dataset spanning three representative domains (scientific writing, news articles, and creative writing) across four text types (human-written, purely machine-generated, machine-rewritten, and machine-polished), generated using state-of-the-art LLMs for robust evaluation. Experimental results demonstrate that MSPO significantly outperforms existing methods across all text types. On the challenging rewritten texts, MSPO achieves up to 82.14% AUROC, representing an improvement of 11.15 percentage points over the strongest baseline ImBD, while maintaining robust cross-domain generalizability across scientific, news, and creative writing domains.

Zitieren

APA

Wang, Z., Wu, L., An, W., Zhang, H., & Wang, Y. (2026). Multi-level Style Preference Optimization: An Adaptive Detection Framework for Human-Machine Hybrid Text. Proceedings of the AAAI Conference on Artificial Intelligence, 40(40), 33746-33754. https://doi.org/10.1609/aaai.v40i40.40665

Vancouver

Wang Z, Wu L, An W, Zhang H, Wang Y. Multi-level Style Preference Optimization: An Adaptive Detection Framework for Human-Machine Hybrid Text. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(40):33746-33754. doi:10.1609/aaai.v40i40.40665

BibTeX
@article{Wang2026,
  author = {Wang, Zehao and Wu, Lianwei and An, Wenbo and Zhang, Hang and Wang, Yaxiong},
  title = {Multi-level Style Preference Optimization: An Adaptive Detection Framework for Human-Machine Hybrid Text},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {40},
  pages = {33746--33754},
  doi = {10.1609/aaai.v40i40.40665}
}
Kostenlos starten →