A Severity-Based Curriculum Learning Strategy for Arabic Medical Text Generation

TL;DR

This paper proposes a severity-based curriculum learning approach for Arabic medical text generation, improving model performance by gradually training on more severe cases.

Contribution

It introduces a novel curriculum learning strategy that orders training data by severity levels, enhancing Arabic medical text generation models.

Findings

Performance improved by 4-7% over baseline models.

Severity-aware training benefits all tested models.

Annotated severity levels enable structured curriculum learning.

Abstract

Arabic medical text generation is increasingly needed to help users interpret symptoms and access general health guidance in their native language. Nevertheless, many existing methods assume uniform importance across training samples, overlooking differences in clinical severity. This simplification can hinder the model's ability to properly capture complex or high-risk cases. To overcome this issue, this work introduces a Severity-based Curriculum Learning Strategy for Arabic Medical Text Generation, where the training process is structured to move gradually from less severe to more critical medical conditions. The approach divides the dataset into ordered stages based on severity and incrementally exposes the model to more challenging cases during fine-tuning, allowing it to first learn basic medical patterns before addressing more complex scenarios. The proposed method is evaluated on a subset of the Medical Arabic Question Answering (MAQA) dataset, which includes Arabic medical questions describing symptoms alongside corresponding responses. In addition, the dataset is annotated with three severity levels (Mild, Moderate, and Critical) using a rule-based method developed in this study. The results demonstrate that incorporating severity-aware curriculum learning leads to consistent performance improvements across all tested models, with gains of around +4% to +7% over baseline models and +3% to +6% compared with conventional fine-tuning approaches.