A bidirectional Mamba/SSM encoder pretrained from scratch for Arabic via masked language modeling, reaching near-parity with base-sized Transformer encoders on consumer-grade hardware.
Publications
Scientific articles in NLP and Text Processing.
Efficient Multilingual Neural Machine Translation via Corpus-Driven Vocabulary Pruning: An English-Arabic Case Study
A framework combining corpus-driven vocabulary pruning and targeted fine-tuning to compress multilingual NMT models by 60% without performance loss.
Implementation and evaluation of modern Neural Machine Translation models for the Arabic-English language pair.
Sarcasm Detection in Arabic Tweets: A Comparison Between Deep Learning and Pre-trained Transformer-based Models
A study on detecting sarcasm in dialectal Arabic tweets using machine learning techniques.
Evaluating Siamese Architecture Neural Models for Arabic Textual Similarity and Plagiarism Detection
Implementation of Siamese neural networks to detect textual similarity and plagiarism in Arabic documents.
Exploring Latent Semantic Analysis (LSA) for identifying plagiarism in Arabic text.
Research on sentiment classification for Arabic text using genetic algorithms to optimize performance.