MURAD: A Large-Scale Multi-Domain Unified Reverse Arabic Dictionary Dataset

TL;DR

The paper addresses the lack of large-scale, multi-domain Arabic lexical resources linking words to formal definitions. It introduces MURAD, a large open reverse Arabic dictionary with 96,243 word-definition triplets drawn from 17 sources across Islamic, linguistic, and scientific domains. A hybrid OCR-and-NLP pipeline (including Mistral OCR and GPT-4o) is used to extract, normalize, and structure definitions according to eight lexicographic standards, ensuring high-quality, reusable data. The dataset enables reverse dictionary modeling, semantic retrieval, definition generation, and embedding evaluation, and it is released with complete processing pipelines and open-source code to support reproducible research in Arabic lexical semantics.

Abstract

Arabic is a linguistically and culturally rich language with a vast vocabulary that spans scientific, religious, and literary domains. Yet, large-scale lexical datasets linking Arabic words to precise definitions remain limited. We present MURAD (Multi-domain Unified Reverse Arabic Dictionary), an open lexical dataset with 96,243 word-definition pairs. The data come from trusted reference works and educational sources. Extraction used a hybrid pipeline integrating direct text parsing, optical character recognition, and automated reconstruction. This ensures accuracy and clarity. Each record aligns a target word with its standardized Arabic definition and metadata that identifies the source domain. The dataset covers terms from linguistics, Islamic studies, mathematics, physics, psychology, and engineering. It supports computational linguistics and lexicographic research. Applications include reverse dictionary modeling, semantic retrieval, and educational tools. By releasing this resource, we aim to advance Arabic natural language processing and promote reproducible research on Arabic lexical semantics.

Figures (3)

• Figure 1: Overview of the Arabic Reverse Dictionary data processing workflow.
• Figure 2: Distribution of definition lengths (in number of words), with statistical highlights indicating the shortest, median, average, and longest entries.
• Figure 3: Domain-level distribution of total definitions across the four major source categories.