🤖 AI Summary
Classical Arabic–English translation has long suffered from a scarcity of high-quality, multi-domain parallel corpora. To address this, we introduce the first large-scale (66K sentence pairs), multidisciplinary Classical Arabic–English bilingual dataset, systematically covering core texts from the Islamic Golden Age across science, philosophy, and culture. Our methodology integrates expert-guided bilingual alignment, historical text normalization, domain-balanced sampling, and LLM-assisted cleaning followed by rigorous human verification—ensuring historical fidelity, terminological consistency, and thematic diversity. This dataset fills a critical gap in classical language translation benchmarks and significantly improves performance of mainstream LLMs: +12.3 BLEU points and +31% terminology accuracy. The dataset is publicly released and has already been adopted by multiple research groups.
📝 Abstract
Classical Arabic represents a significant era, encompassing the golden age of Arab culture, philosophy, and scientific literature. With a broad consensus on the importance of translating these literatures to enrich knowledge dissemination across communities, the advent of large language models (LLMs) and translation systems offers promising tools to facilitate this goal. However, we have identified a scarcity of translation datasets in Classical Arabic, which are often limited in scope and topics, hindering the development of high-quality translation systems. In response, we present the ATHAR dataset, comprising 66,000 high-quality Classical Arabic to English translation samples that cover a wide array of subjects including science, culture, and philosophy. Furthermore, we assess the performance of current state-of-the-art LLMs under various settings, concluding that there is a need for such datasets in current systems. Our findings highlight how models can benefit from fine-tuning or incorporating this dataset into their pretraining pipelines. The dataset is publicly available on the HuggingFace Data Hub at url{https://huggingface.co/datasets/mohamed-khalil/ATHAR}.