Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) building. Image Courtesy: Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)
ABU DHABI – Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) is rapidly expanding its influence in artificial intelligence (AI) research, achieving remarkable milestones in 2024. Between January and June, the university’s community—comprising over 80 world-class faculty members, 200 researchers, and hundreds of students—published more than 300 papers at leading AI conferences. Among these, 39 papers were presented at the prestigious International Conference on Learning Representations 2024 (ICLR) in May.
This success follows the publication of 612 papers in 2023, including notable contributions to the International Conference on Computer Vision, IEEE / CVF Computer Vision and Pattern Recognition Conference, Empirical Methods in Natural Language Processing, and the Conference on Neural Information Processing Systems.
In just five years since its inception, MBZUAI has earned a place among the world’s top 100 universities in computer science and is ranked within the top 20 globally in AI, computer vision, machine learning, natural language processing (NLP), and robotics, according to CSRankings.
Five standout research papers from MBZUAI published in the first half of 2024 include:
- Tackling Misuse of LLM-Generated Text: A collaborative effort by MBZUAI researchers and international partners produced resources to identify text generated by large language models (LLMs). This work could significantly impact fields such as journalism, academia, and education.
- Enhancing Gene-Sequencing Analysis: Professor Kun Zhang, along with his Ph.D. student Gongxu Luo and American university researchers, developed a model that improves the accuracy of gene-sequencing processes, potentially advancing our understanding of diseases like cancer and enhancing treatment outcomes.
- Improving Machine Learning Efficiency: MBZUAI researchers proposed a new algorithm to enhance training efficiency in models that use “hard-thresholding.” This approach reduces errors and accelerates results, with promising applications in financial portfolios and cybersecurity.
- Advancing Vision Language Models: MBZUAI led a global team in creating the Grounding Large Multimodal Model (GLaMM), which improves the interaction between text and images at the pixel level, enabling enhanced automated image captioning, reasoning, and object manipulation.
- Boosting Vision Transformer Efficiency: Dr. Xiaodan Liang and Professor Xiaojun Chang, in collaboration with international researchers, developed a technique to make vision transformers—key components in modern image and video analysis—more efficient.
These achievements underscore MBZUAI’s commitment to advancing AI research and solidifying its position as a global leader in the field.
