Pretrained language models have achieved great success in various natural language understanding (NLU) tasks due to their capacity to capture deep contextualized information in text using pretraining on large-scale corpora. Tokenization plays a significant role in the process of lexical analysis. Tokens become the input for other natural language processing (NLP) tasks, like semantic parsing and language modeling. However, there is a lack of research on the evaluation of the impact of tokenization on the Arabic language model. Therefore, this study aims to address this gap in the literature by evaluating the performance of various tokenizers on Arabic large language models (LLMs). In this paper, we analyze the differences between WordPiece, SentencePiece, and BBPE tokenizers by pretraining three BERT models using each tokenizer while measuring the performance of each model on seven different NLP tasks using 29 different datasets. Overall, the model pretrained with text tokenized using the SentencePiece tokenizer significantly outperforms the other two models that utilize WordPiece and BBPE tokenizers. The results of this paper will assist researchers in developing better models, making better decisions in selecting the best tokenizers, improving feature engineering, and making models more efficient, thus ultimately leading to advancements in various NLP applications.
In this paper, we introduce SaudiBERT, a monodialect Arabic language model pretrained exclusively on Saudi dialectal text. To demonstrate the model's effectiveness, we compared SaudiBERT with six different multidialect Arabic language models across 11 evaluation datasets, which are divided into two groups: sentiment analysis and text classification. SaudiBERT achieved average F1-scores of 86.15\% and 87.86\% in these groups respectively, significantly outperforming all other comparative models. Additionally, we present two novel Saudi dialectal corpora: the Saudi Tweets Mega Corpus (STMC), which contains over 141 million tweets in Saudi dialect, and the Saudi Forums Corpus (SFC), which includes 15.2 GB of text collected from five Saudi online forums. Both corpora are used in pretraining the proposed model, and they are the largest Saudi dialectal corpora ever reported in the literature. The results confirm the effectiveness of SaudiBERT in understanding and analyzing Arabic text expressed in Saudi dialect, achieving state-of-the-art results in most tasks and surpassing other language models included in the study. SaudiBERT model is publicly available on \url{https://huggingface.co/faisalq/SaudiBERT}.
4 June 2025 in Engineering, Technology & Applied Science Research
Large Language Models (LLMs) have shown outstanding performance in many Natural Language Processing (NLP) tasks for high-resource languages, especially English, primarily because most of them were trained on widely available text resources. As a result, many low-resource languages, such as Arabic and African languages and their dialects, are not well studied, raising concerns about whether LLMs can perform fairly across them. Therefore, evaluating the performance of LLMs for low-resource languages and diverse dialects is crucial. This study investigated the performance of LLMs in Moroccan Arabic, a low-resource dialect spoken by approximately 30 million people. The performance of 14 Arabic pre-trained models was evaluated on the Moroccan dialect, employing 11 datasets across various NLP tasks such as text classification, sentiment analysis, and offensive language detection. The evaluation results showed that MARBERTv2 achieved the highest overall average F1-score of 83.47, while the second-best model, DarijaBERT-mix, had an average F1-score of 83.38. These findings provide valuable insights into the effectiveness of current LLMs for low-resource languages, particularly the Moroccan dialect.
Arabic poetry, with its rich linguistic features and profound cultural significance, presents a unique challenge to the Natural Language Processing (NLP) field. The complexity of its structure and context necessitates advanced computational models for accurate analysis. In this paper, we introduce AraPoemBERT, an Arabic language model pretrained exclusively on Arabic poetry text. To demonstrate the effectiveness of the proposed model, we compared AraPoemBERT with 5 different Arabic language models on various NLP tasks related to Arabic poetry. The new model outperformed all other models and achieved state-of-the-art results in most of the downstream tasks. AraPoemBERT achieved unprecedented accuracy in two out of three novel tasks: poet's gender classification (99.34\% accuracy), and poetry sub-meter classification (97.79\% accuracy). In addition, the model achieved an accuracy score in poems' rhyme classification (97.73\% accuracy) which is almost equivalent to the best score reported in this study. Moreover, the proposed model significantly outperformed previous work and other comparative models in the tasks of poems' sentiment analysis, achieving an accuracy of 78.95\%, and poetry meter classification (99.03\% accuracy), while significantly expanding the scope of these two problems. The dataset used in this study, contains more than 2.09 million verses collected from online sources, each associated with various attributes such as meter, sub-meter, poet, rhyme, and topic. The results demonstrate the effectiveness of the proposed model in understanding and analyzing Arabic poetry, achieving state-of-the-art results in several tasks and outperforming previous works and other language models included in the study. AraPoemBERT model is publicly available on \url{https://huggingface.co/faisalq}.
This study presents EgyBERT, an Arabic language model pretrained on 10.4 GB of Egyptian dialectal texts. We evaluated EgyBERT's performance by comparing it with five other multidialect Arabic language models across 10 evaluation datasets. EgyBERT achieved the highest average F1-score of 84.25% and an accuracy of 87.33%, significantly outperforming all other comparative models, with MARBERTv2 as the second best model achieving an F1-score 83.68% and an accuracy 87.19%. Additionally, we introduce two novel Egyptian dialectal corpora: the Egyptian Tweets Corpus (ETC), containing over 34.33 million tweets (24.89 million sentences) amounting to 2.5 GB of text, and the Egyptian Forums Corpus (EFC), comprising over 44.42 million sentences (7.9 GB of text) collected from various Egyptian online forums. Both corpora are used in pretraining the new model, and they are the largest Egyptian dialectal corpora to date reported in the literature. Furthermore, this is the first study to evaluate the performance of various language models on Egyptian dialect datasets, revealing significant differences in performance that highlight the need for more dialect-specific models. The results confirm the effectiveness of EgyBERT model in processing and analyzing Arabic text expressed in Egyptian dialect, surpassing other language models included in the study. EgyBERT model is publicly available on \url{https://huggingface.co/faisalq/EgyBERT}.