arXiv ·
MBZUAI researchers release 'Fann or Flop', a new benchmark for evaluating Arabic poetry understanding in LLMs. The benchmark covers 12 historical eras and 14 poetic genres, assessing semantic understanding, metaphor interpretation, and cultural context. Evaluation of state-of-the-art LLMs reveals challenges in poetic understanding despite strong performance on standard Arabic benchmarks.
MBZUAI ·
Researchers from MBZUAI and other institutions have developed a new framework called STEREO to improve the safety of text-to-image diffusion models. STEREO uses a two-stage approach: STE (Search Thoroughly Enough) based on adversarial training and REO (Robustly Erase Once) for batch concept erasure. This framework aims to enhance safety without significantly impacting the model's performance on normal queries. Why it matters: The framework addresses vulnerabilities in AI image generation, reducing the creation of inappropriate images while preserving performance on harmless queries.
MBZUAI ·
MBZUAI researchers are working on digital twin technology that can replicate human beings in detail, with real-time data flow between the physical and virtual. This project aims to extend digital twins from objects to organic entities like humans, plants and animals. The technology mines data from cameras, sensors, wearables, and other sources to predict health issues before they arise. Why it matters: This research has the potential to transform healthcare by enabling the prediction and prevention of health issues.
TII ·
Dr. Abdelrahman AlMahmoud from the Secure Systems Research Center (SSRC) has been named a founding member of the Emirates Robotic and Automation Society (ERAS). He also joins ERAS's Board of Directors, alongside 20 other members from government, academia, and the private sector. This appointment will help align SSRC's robotics research goals with other entities and strengthen TII's ability to attract talent. Why it matters: This establishes a formal body to advance robotics and automation in the UAE, signaling increased focus and coordination in this sector.
KAUST ·
KAUST researchers developed a low-cost, AI-powered drone system to recognize and track camels, addressing challenges faced by local herders. The system uses commercial drones, cameras, and machine learning to monitor camel herds in real time without expensive GPS collars. The AI model revealed insights into camel migration patterns, showing coordinated grazing and sensitivity to drone sounds. Why it matters: This system offers an affordable solution to preserve Saudi Arabia's camel herding tradition while providing valuable insights into camel behavior and contributing to the local economy.
KAUST ·
The Communications and Computing Systems Lab (CCSL) at KAUST received two awards in the International Telecommunication Union AI for Good Machine Learning Challenge and tinyML Hackathon Challenge 2023: Pedestrian Detection. The KAUST team's solution achieved high accuracy in pedestrian identification using event-based cameras, while consuming less power and achieving lower latency. They also received an award for innovative use of "Edge Impulse" for building datasets and training models. Why it matters: This recognition highlights KAUST's growing influence in AI research, particularly in edge computing and computer vision applications for public safety.
KAUST ·
KAUST doctoral student Jose Filho won the 2022 InnovateFPGA Design Contest for his "Customized Medicine for Corals" project. The project uses an automatic feeder technology to deliver coral probiotics and monitor their efficacy via cloud connectivity, computer vision, and an FPGA. The system gathers data from cameras, temperature sensors, and luminosity sensors, using AI to determine the coral's bleaching stage and deploy beneficial microorganisms. Why it matters: This win highlights KAUST's innovative research in applying AI and cloud technology to address critical environmental challenges like coral bleaching, demonstrating the potential for technology to aid marine conservation efforts.
arXiv ·
Team NimbRo presented four UAVs tailored for the MBZIRC 2020 challenges, including target chasing, wall building, and fire fighting. The UAVs utilized onboard object detection, aerial manipulation, LiDAR, and thermal cameras to perform their tasks autonomously. The team's software stack, which is mostly open-source, includes tools for system configuration, monitoring, and agile trajectory generation. Why it matters: The work demonstrates advanced robotics capabilities developed in the context of a major regional competition, advancing machine vision and trajectory generation, and showcasing potential applications in various sectors.