MoSEs : Uncertainty-Aware AI-Generated Text Detection via Mixture of Stylistics Experts with Conditional Thresholds
Junxi Wu, Jinpeng Wang, Zheng Liu, Bin Chen, Dongjian Hu, Hao Wu, Shu-Tao Xia · 2025
Read MoSEs : Uncertainty-Aware AI-Generated Text Detection via Mixture of Stylistics Experts with Conditional Thresholds by Junxi Wu, Jinpeng Wang, Zheng Liu, Bin Chen, Dongjian Hu, Hao Wu, Shu-Tao Xia (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-02499.txt.
Description
Junxi Wu and six co-authors present a 2025 text analysis paper on spotting AI-generated writing as language models improve. Many detectors ignore writing style and use one fixed threshold, which caps their accuracy.
Their MoSEs framework mixes several stylistics experts, models tuned to style cues, and estimates its threshold for each input. A reference repository supplies matching style examples. A router picks the right references, and a threshold estimator combines word statistics with semantic features to set the cut-off and report confidence. Average detection scores rose clearly over baselines. Gains were much larger when training data was scarce, a common real-world case. The authors release their code for reuse.
Catalog record
| CREATOR | Junxi Wu, Jinpeng Wang, Zheng Liu, Bin Chen, Dongjian Hu, Hao Wu, Shu-Tao Xia |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-02499