This research addresses the underexplored area of aligning Arabic Large Language Models (LLMs) with country-specific cultural behavioral preferences. The authors propose ArabPref, the first large-scale pan-Arab preference dataset designed for cultural behavior alignment conditioned on country. ArabPref includes both a culturally grounded preference dataset for training and evaluation, and a multiple-choice benchmark for assessing culturally aligned behavior across various Arab nations, with all test data validated by native speakers. Why it matters: This dataset and benchmark provide a crucial foundation for systematically studying and improving the cultural nuance and appropriateness of LLMs for the diverse Arab world.
KAUST-developed biomedical microgravity experiments, SSA-1 and SSA-2, were launched to the International Space Station via SpaceX CR35. The project, involving KAUST, the Saudi Space Agency's BioGravity Program, Axiom Space, and the National Stem Cell Foundation, aims to study brain cells in microgravity to advance neurodegenerative disease research. This environment accelerates conditions mimicking health challenges on Earth, providing insights into diseases like Alzheimer’s and multiple sclerosis. Why it matters: This initiative could significantly accelerate drug discovery for neurodegenerative diseases and develops specialized local infrastructure and human capital in biomedical microgravity research for Saudi Arabia.
Researchers developed a force-data-driven robotic harvesting system for cherry tomatoes, integrating six-axis force/torque sensing, a bio-inspired compliant gripper, and a mobile robotic platform. Manual trials informed the design of a novel Finray gripper and a hybrid grip-twist-pull strategy, applying specific forces and torques to detach the fruit gently. Validated at a commercial greenhouse in Abu Dhabi, UAE, the system achieved an 86.0% picking success rate with only 2.3% fruit damage and an 8.4-second cycle time per fruit. Why it matters: This development demonstrates a practical and efficient solution for delicate robotic harvesting in controlled agricultural environments, with direct relevance to food security initiatives in the Middle East.
Researchers have developed MARTS (Modular Agricultural Robot for Targeted Spraying), a low-cost robotic platform designed for precise disease control in greenhouse tomato production. MARTS integrates a mobile base, a UR5 manipulator, a modular spraying end-effector, and an Intel RealSense RGB-D camera, utilizing a vision-language model for disease detection. Deployed at the SILAL commercial greenhouse in Al Ain, UAE, the robot successfully treated 312 diseased leaves across 12 trials, achieving 92.4% targeting success and a 71% reduction in pesticide volume compared to broadcast spraying. The spraying subsystem costs significantly less than comparable systems, demonstrating the feasibility of accurate leaf-level targeting with commodity hardware. Why it matters: This innovation provides a precise, cost-effective, and environmentally friendly solution for agricultural disease management, reducing pesticide use and worker exposure, with successful application and cost efficiency demonstrated in the UAE.