@inproceedings{dao-sy-etal-2026-hcmus-thefangs,
    title = "{HCMUS}{\_}{T}he{F}angs at {A}bjad{G}en{E}val Shared Task: Weighted Layer Pooling with Attention Fusion for {A}rabic {AI}-Generated Text Detection",
    author = "Dao Sy, Duy Minh  and
      Tran, Nguyen Chi  and
      Huynh, Trung Kiet  and
      Quy, Nguyen Lam Phu  and
      Phu Hoa, Pham  and
      Duong, Nguyen Dinh Ha",
    booktitle = "Proceedings of the 2nd Workshop on {NLP} for Languages Using {A}rabic Script",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/manual-author-scripts/2026.abjadnlp-1.51/",
    pages = "433--437",
    abstract = "The rapid advancement of large language mod-els poses significant challenges for content au-thenticity, particularly in under-resourced lan-guages where detection tools remain scarce.We present our winning system for the Abjad-GenEval shared task on Arabic AI-generatedtext detection. Our key insight is that AI-generated text exhibits distinctive patternsacross multiple linguistic levels-from local syn-tax to global semantics-that can be captured bylearning to fuse representations from differenttransformer layers. We introduce aWeightedLayer Poolingmechanism that learns optimallayer combinations, combined withAttentionPoolingfor sequence-level context aggregation.Through systematic experimentation with 15+ approaches, we make a surprising discovery:model architecture selection dominates over so-phisticated training techniques, with DeBERTa-v3 providing +27{\%} relative improvement overAraBERT regardless of training strategy. Oursystem achieves 0.93 F1-score, securing 1st placeamong all participants and outperform-ing the runner-up by 3 absolute points"
}