Since its establishment in 2019, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) has strived to expand the boundaries of AI research and innovation with a commitment to addressing real-world challenges in healthcare, sustainability, education and smart cities.
Visitors to two major technology events being held in the UAE this week will have the opportunity to see the progress that MBZUAI faculty, students, and researchers have made over the past five years in a wide range of AI-driven fields, as well as being able to preview some of the exciting new projects the university is supporting.
Visitors at IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024) in Abu Dhabi, and the annual GITEX Global technology event in Dubai, are exploring new MBZUAI projects, including:
FRAPPE (FRAming, Persuasion, and Propaganda Explorer)
FRAPPE is a system designed to analyze framing, persuasion, and propaganda techniques in media content. It provides real-time analysis of individua
l articles, focusing on their genre, framing, and persuasion strategies, allowing comparisons across media outlets, countries, and languages.
This innovative tool is aimed at researchers, media professionals, and the general public, promoting media literacy by revealing the techniques used to shape opinions and emotions.
By fostering a deeper understanding of media manipulation and encouraging informed decision-making, FRAPPE helps build trust in media and supports more transparent information dissemination globally.
BiMediX2 – AI Virtual Arabic Doctor
BiMediX2 is the first bilingual medical large language model (LLM) and large multimodal model (LMM) that can understand and generate content across multiple types of data, such as text, images, audio, and video. It can facilitate a wide range of medical interactions, including multi-turn chats and question answering, and can even understand and analyze medical images, such as x-rays, CT scans and MRIs.
The research team believes that the service will be pa
rticularly useful in remote areas where doctors use telemedicine platforms to meet with patients. They are currently expanding the functionality of BiMediX2 to enable it to locate nearby physicians and facilities and book appointments for patients who need further medical engagement.
Dog Bot and Immersive Command Center for Robot Dog Navigation
‘Dog Bot’ is a visual analytics generative pre-trained transformer (GPT) that combines image, text and audio capabilities, enabling speech and vision recognition and understanding, and audio response.
The solution could be used across a range of sectors, such as inspecting work sites for health and safety issues, assisting police work by patrolling areas, and surveying agricultural sites for crop performance. Effectively, the Dog Bot could act as the eyes and ears of important public services – even supporting healthcare by interacting with patients in remote areas and sharing diagnostic images with doctors.
GeoChat+
Teams of researchers from MBZUAI are also tackl
ing crucial issues relating to sustainability. GeoChat+, a vision-language model (VLM) for multi-modal, temporal remote sensing (RS) image analysis, aims to help monitor deforestation to help curb the loss of biodiversity, protect endangered species and limit disruption to ecosystems.
GeoChat+ is set to revolutionise geographic data analysis by integrating a wide array of satellite and imaging source. The platform utilises data from renowned Earth observation systems like Sentinel-1 and Sentinel-2, along with Landsat and high-resolution imagery from WorldView and Google Earth.
The aim is to provide comprehensive, geospatial insights for a range of applications, from environmental monitoring to disaster response and urban planning.
Source: Emirates News Agency