--- title: "A Severity-Based Curriculum Learning — Alansary 2026" description: "The paper tackles Arabic medical answer generation." authors: ["Ahmed Alansary", "Molham Mohamed", "Ali Hamdi"] year: 2026 venue: "arXiv:2604.06365" source: "https://arxiv.org/abs/2604.06365" doi: "10.48550/arxiv.2604.06365" arxiv: "2604.06365" licence: "CC BY-SA 4.0 (arXiv)" licence_url: "http://creativecommons.org/licenses/by-sa/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/alansary-2026-severity-based-curriculum-learning-strategy/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/alansary-2026-severity-based-curriculum-learning-strategy/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/) / A Severity-Based Curriculum Learning Strategy for Arabic Medical Text Generation # A Severity-Based Curriculum Learning Strategy for Arabic Medical Text Generation **Authors:** [Ahmed Alansary](https://md.research.elamur.ai/free/en/authors/ahmed-alansary/), [Molham Mohamed](https://md.research.elamur.ai/free/en/authors/molham-mohamed/), [Ali Hamdi](https://md.research.elamur.ai/free/en/authors/ali-hamdi/) · **Institutions:** [October University of Modern Sciences and Arts](https://md.research.elamur.ai/free/en/universities/october-university-of-modern-sciences-and-arts/) · **Year:** 2026 · **Venue:** arXiv:2604.06365 · **Licence:** [CC BY-SA 4.0 (arXiv)](http://creativecommons.org/licenses/by-sa/4.0/) **Links:** [Publisher's page](https://arxiv.org/abs/2604.06365) · [PDF](https://arxiv.org/pdf/2604.06365) · [DOI 10.48550/arxiv.2604.06365](https://doi.org/10.48550/arxiv.2604.06365) · [arXiv 2604.06365](https://arxiv.org/abs/2604.06365) · [OpenAlex](https://openalex.org/W7152578752) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary The paper tackles Arabic medical answer generation. In this setting, ordinary fine-tuning treats every training pair as equally important, even though the cases range from trivial complaints to emergencies. The authors take a 32K-pair subset of the MAQA (Medical Arabic Question Answering) dataset and give each question one of three urgency labels: Mild, Moderate or Critical. They assign these labels with a rule-based tagger that normalizes the Arabic text and then searches it for symptom keywords. Training then follows a curriculum: the model is fine-tuned in successive stages, starting with mild cases and moving to moderate and then critical ones, with cross-entropy loss at each stage. The authors compare ten small Arabic-capable language models under three conditions: no tuning, standard fine-tuning and the curriculum. Every model scores higher with the curriculum, and the best result, 68.00%, comes from LFM2-700M. The authors argue that the gain comes from how the data is ordered rather than from any particular architecture. This makes the approach a cheap, model-agnostic way to improve medical generation in a low-resource language. ## Key points - Data: a 32K question–answer subset of MAQA. Each question is labeled Mild, Moderate or Critical by keyword rules. - Labeling rules: - Preprocessing removes diacritics and punctuation and normalizes letter variants. - Emergency terms such as chest pain, fainting, heavy bleeding or breathing trouble mark a question as Critical. - Fever, vomiting or persistent pain mark it as Moderate. - When several cues fire, the most severe one wins. - Questions with no matching keyword default to Mild. - Models tested: AraGPT2-Base, AraGPT2-Medium, Bloomz-560M, Al-Atlas-0.5B, GPT2-Small-Arabic, LFM2-350M, LFM2-700M, Dialect-ar-gpt-2021, Qwen2.5-0.5B and Qwen3-0.6B. - Reported gains: the curriculum beats the baseline by about +4 to +7 points and standard fine-tuning by about +3 to +6, according to the authors. The table's own per-model differences are less tidy: - Bloomz-560M gains only about +2 over baseline. - AraGPT2-Base gains about +9 over baseline. - Standard fine-tuning alone adds about +5.7 for AraGPT2-Base, more than the stated ~2% ceiling. - Smaller models benefit most from the curriculum. Larger ones gain less but reach the top scores, with LFM2-700M at 68.00%. ## Interesting ideas - A crude rule-based difficulty label (a keyword lexicon, highest-severity-wins, a safe default) is enough to impose a curriculum without any human annotation. — For Mova (learn): order silver treebank sentences for the perceptron parser and gparse by a transparent difficulty proxy, such as length, non-projective arcs or witness-parser disagreement, and train in stages from easy to hard. - The labeling resolves conflicts with an explicit priority order, so every label can be traced to the cue that triggered it. — For Mova (frames): in crate prag, speech-act or irony cue lexicons could use the same ranked "strongest cue wins" rule, so each tagged frame carries an explainable reason. - The evaluation includes both an untuned baseline and a standard fine-tune on the same data, which isolates the effect of ordering alone. — For Mova (learn): add a negative control in which the same silver data is shuffled instead of staged, to show whether any curriculum gain is real. - The observation that low-capacity models profit most from staged exposure suggests curricula matter most where model capacity is the bottleneck. — For Mova (tree): test the curriculum first on the small perceptron parser, where the payoff may be larger than for the GPU transformer gparse. - The paper's stated gain ranges do not fully match its own table, which shows why deltas should be recomputed from the raw scores. — For Mova (learn): have evaluation reports compute per-model differences automatically from stored scores instead of quoting summary ranges by hand. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/) Topics: [Text generation](https://md.research.elamur.ai/free/en/topics/text-generation/) · [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @misc{alansary2026severity, title = {{A Severity-Based Curriculum Learning Strategy for Arabic Medical Text Generation}}, author = {Alansary, Ahmed and Mohamed, Molham and Hamdi, Ali}, year = {2026}, doi = {10.48550/arxiv.2604.06365}, eprint = {2604.06365}, archivePrefix = {arXiv}, url = {https://doi.org/10.48550/arxiv.2604.06365} } ``` - [← previous A Scalable Pipeline for Novelty Detection in Skill Extraction Using Large Language Models](https://md.research.elamur.ai/free/en/papers/seifert-2026-scalable-pipeline-novelty-detection-skill/) - [next → A Single Model Ensemble Framework for Neural Machine Translation Using Pivot Translation](https://md.research.elamur.ai/free/en/papers/oh-2026-single-model-ensemble-framework-neural/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/) ## References 10 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. 1. Abdelhay, M., Mohammed, A., Hefny, H.A.: Deep learning for arabic healthcare: Medicalbot. Social Network Analysis and Mining 13(1), 71 (2023) 2. Ahmed, S., Allam, A., Hamdi, A., Mohammed, A.: Arabic symptom classification and diagnosis using transformer models and llm-based augmentation. [find](https://scholar.google.com/scholar?q=1%2E+Abdelhay%2C+M%2E%2C+Mohammed%2C+A%2E%2C+Hefny%2C+H%2EA%2E%3A+Deep+learning+for+arabic+healthcare%3A+Medicalbot%2E+Social+Network+Analysis+and+Mining+13%281%29%2C+71+%282023%29+2%2E+Ahmed%2C+S%2E%2C+Allam%2C+A%2E%2C+Hamdi%2C+A%2E%2C+Mohammed%2C+A%2E%3A+Arabi) 2. IEEE, Egypt (July 2025) 3. Alkhurayyif, Y.: Enhancing medical knowledge access: Development and assessment of an open-domain medical question answering system with deep learning. In: 2025 4th International Conference on Computing and Information Technology (ICCIT). pp. 231–237. [find](https://scholar.google.com/scholar?q=IEEE%2C+Egypt+%28July+2025%29+3%2E+Alkhurayyif%2C+Y%2E%3A+Enhancing+medical+knowledge+access%3A+Development+and+assessment+of+an+open%2Ddomain+medical+question+answering+system+with+deep+learning%2E+In%3A+2025+4th+Internat) 3. IEEE (2025) 4. Allam, A., Ahmed, S., Hamdi, A., Mohammed, A.: Arabic large language models for medical text generation. [find](https://scholar.google.com/scholar?q=IEEE+%282025%29+4%2E+Allam%2C+A%2E%2C+Ahmed%2C+S%2E%2C+Hamdi%2C+A%2E%2C+Mohammed%2C+A%2E%3A+Arabic+large+language+models+for+medical+text+generation%2E) [our card](https://md.research.elamur.ai/free/en/papers/allam-2025-arabic-large-language-models-medical/) 4. IEEE, Malaysia (February 2025) 5. Chang, E., Yeh, H.S., Demberg, V.: Does the order of training samples matter? improving neural data-to-text generation with curriculum learning. Association for Computational Linguistics (Apr 2021) 6. Garg, R., Gupta, A.: A systematic review of nlp applications in clinical healthcare: Advancement and challenges. In: International Conference on Advances in Data-driven Computing and Intelligent Systems. pp. 31–44. [find](https://scholar.google.com/scholar?q=IEEE%2C+Malaysia+%28February+2025%29+5%2E+Chang%2C+E%2E%2C+Yeh%2C+H%2ES%2E%2C+Demberg%2C+V%2E%3A+Does+the+order+of+training+samples+matter%3F+improving+neural+data%2Dto%2Dtext+generation+with+curriculum+learning%2E+Association+for+Compu) 5. Springer (2023) 7. Guo, J., Tan, X., Xu, L., Qin, T., Chen, E., Liu, T.Y.: Fine-tuning by curriculum learning for non-autoregressive neural machine translation. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 34, pp. 7839–7846 (2020) 8. Hamdi, A., Mohamed, M., Emad, R., Shaban, K.: An ensemble classification approach in a multi-layered large language model framework for disease prediction. [find](https://scholar.google.com/scholar?q=Springer+%282023%29+7%2E+Guo%2C+J%2E%2C+Tan%2C+X%2E%2C+Xu%2C+L%2E%2C+Qin%2C+T%2E%2C+Chen%2C+E%2E%2C+Liu%2C+T%2EY%2E%3A+Fine%2Dtuning+by+curriculum+learning+for+non%2Dautoregressive+neural+machine+translation%2E+In%3A+Proceedings+of+the+AAAI+Conference+) [our card](https://md.research.elamur.ai/free/en/papers/hamdi-2025-ensemble-classification-approach-multi-layered/) 6. IEEE, Qatar (October 2025) 9. Klila, J., Souihi, S.B., Boujelben, R., Semmar, N., Belguith, L.H.: Adapting large language models to biomedical domain: A survey of techniques and approaches. [find](https://scholar.google.com/scholar?q=IEEE%2C+Qatar+%28October+2025%29+9%2E+Klila%2C+J%2E%2C+Souihi%2C+S%2EB%2E%2C+Boujelben%2C+R%2E%2C+Semmar%2C+N%2E%2C+Belguith%2C+L%2EH%2E%3A+Adapting+large+language+models+to+biomedical+domain%3A+A+survey+of+techniques+and+approaches%2E) 7. Springer Nature Switzerland, Cham (June 2024) 10. Kumichev, G., Blinov, P., Kuzkina, Y., Goncharov, V., Zubkova, G., Zenovkin, N., Goncharov, A., Savchenko, A.: Medsyn: Llm-based synthetic medical text generation framework. [find](https://scholar.google.com/scholar?q=Springer+Nature+Switzerland%2C+Cham+%28June+2024%29+10%2E+Kumichev%2C+G%2E%2C+Blinov%2C+P%2E%2C+Kuzkina%2C+Y%2E%2C+Goncharov%2C+V%2E%2C+Zubkova%2C+G%2E%2C+Zenovkin%2C+N%2E%2C+Goncharov%2C+A%2E%2C+Savchenko%2C+A%2E%3A+Medsyn%3A+Llm%2Dbased+synthetic+medical+tex) 8. Springer (2024) 11. Liu, F., Ge, S., Wu, X.: Competence-based multimodal curriculum learning for medical report generation. Association for Computational Linguistics (Aug 2021) 12. Mohamed, M., Emad, R., Hamdi, A.: Ensemble transformers for arabic medical text classification. In: 2025 Intelligent Methods, Systems, and Applications (IMSA). pp. 1–6. [find](https://scholar.google.com/scholar?q=Springer+%282024%29+11%2E+Liu%2C+F%2E%2C+Ge%2C+S%2E%2C+Wu%2C+X%2E%3A+Competence%2Dbased+multimodal+curriculum+learning+for+medical+report+generation%2E+Association+for+Computational+Linguistics+%28Aug+2021%29+12%2E+Mohamed%2C+M%2E%2C+Emad%2C+) 9. IEEE, Egypt (2025) 13. Mohamed, M., Emad, R., Hamdi, A.: A multi-layered large language model framework for disease prediction. In: International Congress on Information and Communication Technology. pp. 259–270. [find](https://scholar.google.com/scholar?q=IEEE%2C+Egypt+%282025%29+13%2E+Mohamed%2C+M%2E%2C+Emad%2C+R%2E%2C+Hamdi%2C+A%2E%3A+A+multi%2Dlayered+large+language+model+framework+for+disease+prediction%2E+In%3A+International+Congress+on+Information+and+Communication+Technology%2E+) 10. Springer Nature Singapore, Singapore (2025) 14. Nazi, Z.A., Peng, W.: Large language models in healthcare and medical domain: A review. Informatics 11(3), 57 (August 2024) 15. Ren, M., Yan, Y., Chen, H., Hu, D., Xu, J., Zeng, X.: Tacl: Threshold-adaptive curriculum learning strategy for enhancing medical text understanding. arXiv preprint arXiv:2510.15269 (2025) 16. Sarella, P.N.K., Mangam, V.T.: Ai-driven natural language processing in healthcare: transforming patient-provider communication. Indian Journal of Pharmacy Practice 17(1), 21–26 (2024) 17. Shen, L., Feng, Y.: CDL: Curriculum dual learning for emotion-controllable response generation. Association for Computational Linguistics (Jul 2020) 18. Zhou, B., Yang, G., Shi, Z., Ma, S.: Natural language processing for smart healthcare. IEEE Reviews in Biomedical Engineering 17, 4–18 (2022) [↗](https://arxiv.org/abs/2510.15269)