This paper introduces AraLLaMA, a new Arabic large language model (LLM) trained using a progressive vocabulary expansion method inspired by second language acquisition. The model utilizes a modified byte-pair encoding (BPE) algorithm to dynamically extend the Arabic subwords in its vocabulary during training, balancing the out-of-vocabulary (OOV) ratio. Experiments show AraLLaMA achieves performance comparable to existing Arabic LLMs on various benchmarks, and all models, data, and code will be open-sourced. Why it matters: This work addresses the need for more accessible and performant Arabic LLMs, contributing to democratization of AI in the Arab world.
MBZUAI researchers introduce Web2Code, a new large-scale dataset and evaluation framework for training and benchmarking multimodal LLMs on webpage understanding and HTML code generation. The dataset includes webpage images, HTML code, and QA pairs about webpage content. Experiments demonstrate the dataset's utility in webpage understanding, code generation, and general visual domain tasks, with code and data available on Github.
This paper explores the impact of tokenization strategies and vocabulary sizes on Arabic language model performance across NLP tasks like news classification and sentiment analysis. It compares four tokenizers, finding that Byte Pair Encoding (BPE) with Farasa performs best overall due to its morphological analysis capabilities. The study surprisingly found limited impact of vocabulary size on performance with fixed model sizes, challenging assumptions about vocabulary size and model performance. Why it matters: The findings provide insights for developing more effective and nuanced Arabic language models, particularly for handling dialectal variations and promoting responsible AI development in the region.
KAUST has upgraded its connectivity with 200 Gbps links to Amsterdam and Singapore, connecting to major research networks in Europe and Asia. This upgrade provides researchers with fast data transmission and access to global scientific resources. The increased bandwidth reduces data transfer times significantly, enabling high-performance science applications. Why it matters: This connectivity boost is unprecedented in the Middle East and empowers KAUST to enhance global research collaboration and fully utilize its advanced data processing capabilities.
Dr. Nesma Aboulkhair, Additive Manufacturing Lead Researcher at TII, has been ranked among Stanford's top 2% of scientists for single-year citations for the second consecutive year. Dr. Aboulkhair joined TII in April 2021 and also holds a visiting academic position at the University of Nottingham in the UK. Her research focuses on metal additive manufacturing processes, including powder-based and droplet-on-demand methods, and she has secured over GBP2 million in funding. Why it matters: This recognition highlights the growing expertise in advanced manufacturing within UAE research institutions and TII's ability to attract and foster high-impact researchers in critical fields.
KAUST's Technology Innovation and Entrepreneurship (TIE) students excelled at the International Business Plan Competition (IBPC) in Austin, Texas, securing positions on the winning, second-place, and third-place teams. The competition focused on transportation and logistics, with KAUST students developing a drone-based solution for organ transport. KAUST sponsored the event, highlighting its commitment to research commercialization and international partnerships. Why it matters: KAUST's strong showing in this international competition underscores its role in fostering innovation and entrepreneurship, aligning with Saudi Vision 2030 and KAUST's strategic priorities.
KAUST postdoctoral fellow Elisabetta Fiacco won the Best Poster Prize at the Spetses Summer School 2018 on Chromatin and Metabolism for her research on the role of carbohydrate-responsive element binding protein (ChRBP) during the onset of type 2 diabetes. Fiacco's research at KAUST, under Assistant Professor Antonio Adamo, uses human induced pluripotent stem cells to understand the epigenetic and transcriptional mechanisms dysregulated in type 2 diabetes. She also gave a 15-minute oral presentation on her work at the event, which gathered over 80 participants from top global universities. Why it matters: This recognition highlights KAUST's contribution to cutting-edge research in regenerative medicine and the study of metabolic disorders prevalent in the region.
MBZUAI researchers have developed "Culturally Yours," a reading assistant that highlights and explains culturally-specific items on webpages to help users understand unfamiliar terms. The tool addresses the "cold-start problem" by asking users for demographic information to personalize the identification of potentially unfamiliar cultural references. It was presented at the 31st International Conference on Computational Linguistics in Abu Dhabi. Why it matters: This tool can help bridge linguistic and cultural gaps, particularly for underrepresented languages and cultures, and aid businesses in reaching diverse audiences.