@inproceedings{alharthi-etal-2026-abjad,
    title = "Abjad {AI} at {KSAA}-2026 Shared Task 2: Grouped Speech Conditioning for {A}rabic Diacritization",
    author = "Alharthi, Naif Saad  and
      Ghannam, Ahmad  and
      Alasmary, Faris  and
      Al Tabash, Kholood  and
      Sadah, Shouq  and
      Ghouti, Lahouari",
    editor = "Al-Khalifa, Hend  and
      El-Haj, Mo  and
      Ezzini, Saad",
    booktitle = "The 7th Workshop on Open-Source {A}rabic Corpora and Processing Tools ({OSACT}7) with 5 Shared Tasks",
    month = may,
    year = "2026",
    address = "Palma, Mallorca (Spain)",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/paragraph-normalization/2026.osact-1.33/",
    doi = "10.63317/4t229uhjkkot",
    pages = "252--255",
    abstract = "We describe Abjad AI{'}s submission to KSAA-2026 Shared Task 2 on automatic diacritization of Arabic speech dictation. The task requires generating fully diacritized text given speech audio and an undiacritized transcript. Because text-only diacritization cannot resolve ambiguities that are recoverable from the acoustic signal, we propose conditioning a character-level encoder-only Transformer (CATT) (Alasmary et al., 2024) on speech representations. We introduce grouped speech conditioning, which downsamples speech encoder features into a small set of pooled tokens concatenated to the text input, enabling efficient fusion without architectural changes to CATT. We train with a two-phase schedule that first freezes the text encoder, then fine-tunes the full model. Our best system, using Whisper-small (Rad-ford et al., 2022) features with five grouped tokens, achieves a Diacritization Error Rate (DER) of 6.60 and a Word Error Rate (WER) of 18.66 (without case endings, including no-diacritic) on the official test set. Notably, we find that Whisper-small consistently outperforms Whisper-large-v3, suggesting that compact speech representations better suit this fusion setting. This is an extended and revised version of our previous work (Ghannam et al., 2025)."
}