المجلة الدولية لنشر البحوث والدراسات

International Journal of Research and Studies Publishing

المجلة الدولية لنشر البحوث والدراسات
القائمة الرئيسية

The Geometry of Arabic Poetry Rhythm (An Energy-Based Artificial Intelligence Approach to Sub-Meter Detection)

By: Zaid Abdulsattar Al-Qudsi (1), Moadh Abdulsattar Abdulhameed (2)

PhD Candidate, Department of Electrical and Computer Engineering, King Abdulaziz University, Saudi Arabia (1)
Supervisor, Operations Sector, center3, Riyadh, Saudi Arabia (2)


Abstract:

This paper proposes an Energy-Based Model (EBM) that formulates sub-meter detection as a geometric optimization problem over rhythmic structures. An energy function measures the compatibility between the latent rhythmic skeleton of an undiacritized verse and the canonical constraints of candidate sub-metrical templates, conditioned on a known primary meter. The sub-meter is identified by selecting the configuration that minimizes the energy value, representing the most rhythmically stable state. Experiments on the Rajaz meter demonstrate clear energy separation between valid and invalid rhythmic configurations, highlighting the effectiveness of energy-based geometric modeling as a robust and data-efficient approach for fine-grained rhythmic analysis of Arabic poetry. Arabic is one of the most morphologically rich and structurally complex Semitic languages and underpins a vast literary and cultural heritage. At the center of this heritage is Arabic poetry, characterized by a strictly regulated rhythmic system derived from the temporal arrangement of vocalized and non-vocalized syllables and formalized through prosodic meter. The Arabic prosodic system comprises sixteen canonical meters, each associated with a distinct rhythmic pattern and capable of generating multiple sub-metrical forms through systematic variations in verse length. The hierarchical relationship between rhythm, meter, and sub-meter poses significant challenges for computational modeling. In modern NLP, the absence of diacritics in standard Arabic text has driven the adoption of deep learning techniques for metrical analysis. Although existing models achieve reasonable performance in identifying primary meters, they remain largely ineffective in detecting fine-grained sub-metrical forms such as Majzūʾ, Mashṭūr, and Manhūk. This limitation stems from both missing phonetic cues and the scarcity of finely annotated datasets for supervised sub-meter classification, often leading to the misclassification of structurally valid rhythmic forms.


Keywords:

Arabic Poetry; Rhythm; Sub-meter; Detection; Energy-Based Model

DOI

Email

About the Authors

PhD Candidate, Department of Electrical and Computer Engineering, King Abdulaziz University, Saudi Arabia (1)
Supervisor, Operations Sector, center3, Riyadh, Saudi Arabia (2)

References

- Alḥanafī, J. (1978). Prosody Refinement and Rewriting. Baghdad: Alʻānī press
- Alkātib, M. Ṭ. (1971). Arabic Poetry Meters Using Binary Numbers (1st ed.). Basrah: Iraqi Ports Authority Press
- Allāh-Verdī, M. (1948). Bdāʼiʻ alʻurūḍ (1st ed.). Damascus: Ibn Zaydoun Printing press
- AlNagdawi, M. A. (2013). Finding Arabic Poem Meter using Context Free Grammar. Journal of Communications and Computer Engineering, 3(1), 52
- Alqudsi, Z., Bilal, M., & Wasly, S. (2025). Detecting Meter of Arabic Poetry Using Deep Learning: Testing Parameters and Model Design. In International Conference on Innovation in Artificial Intelligence and Internet of Things (AIIT). Jeddah, Saudi Arabia
- Alṣṣaḥḥāny, H. (2018). The helper in prosody and rhyme (1st ed.). Damascus: Dār alʻaṣmāʼ.
- Bhattacharyya, S., Rooshenas, A., Naskar, S., Sun, S., Iyyer, M., & McCallum, A. (2021). Energy-Based Reranking: Improving Neural Machine Translation Using Energy-Based Models. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics (ACL)
- Dawid, A., & LeCun, Y. (2024). Introduction to latent variable energy-based models: a path toward autonomous machine intelligence. Journal of Statistical Mechanics: Theory and Experiment, 2024(10), 104011. https://doi.org/10.1088/1742-5468/ad7d94
- Diyāb, M. (2021). History of Arabic Literature. Hindawi
- Mannāʻ, H. Ṣ. (2003). al-Shāfī fī al-ʻarūḍ wa al-qawāfī (4th ed.). Beirut: Dār al-Fikr al-ʻArabi
- Pang, B., Han, T., Nijkamp, E., Zhu, S. C., & Wu, Y. N. (2020). Learning Latent Space Energy-Based Prior Model. In Advances in Neural Information Processing Systems (NeurIPS), 33, 12617-12628
- Saleh, M. I., et al. (2023). 100 Questions About the Arabic Language (1st ed.).King Salman Global Academy for Arabic Language
- Sulṭānī, M. (2009). Prosody and rhythm of Arabic poetry (1st ed.). Damascus: Dār alʻaṣmāʼ.
- Tu, L. (2021). Learning Energy-Based Approximate Inference Networks for Structured Applications in NLP (Doctoral dissertation). Toyota Technological Institute at Chicago
- Unified Energy-Based Framework for Unsupervised Learning. In Proceedings of the 11th International Conference on Artificial Intelligence and Statistics (AISTATS). San Juan, Puerto Rico
- Versteegh, K. (1997). The Arabic Language. New York: Columbia University Press

Download PDF





المجلة الدولية لنشر البحوث والدراسات

المجلة الدولية لنشر البحوث والدراسات مجلة علمية محكمة متخصصة في نشر الابحاث العلمية في مختلف التخصصات، تصدر المجلة الدولية دورياً كل شهر. تصدر المجلة في المملكة الأردنية الهاشمية

تعليمات النشر للباحث

  • » تنسيقات البحث
  • » قالب البحث
  • » أخلاقيات النشر
  • » اقرار الباحث

    • سياسات المجلة

  • » سياسة التحرير
  • » سياسة الوصول المفتوح
  • » حقوق النشر والترخيص
  • » سياسة الاقتباس

    • للإقتراحات

    نحن نعمل باستمرار على تحسين مجلتنا العلمية وعملية النشر لدينا بهدف تزويدك بأفضل تجربة نشر علمية. فإننا نقدر رأيك ونرحب بأي اقتراحات عبر الإيميل التالي: info@ijrsp.com

    جميع الحقوق محفوظة © المجلة الدولية لنشر البحوث والدراسات 2019-2025م