Skip to content
GCC AI Research

Search

Results for "Edge Deployment"

A compact multimodal model for real-time video understanding on edge devices

MBZUAI ·

MBZUAI researchers developed Mobile-VideoGPT, a compact and efficient multimodal model for real-time video understanding on edge devices. The system uses keyframe selection, efficient token projection, and a Qwen-2.5-0.5B language model. Testing showed that Mobile-VideoGPT is faster and performs better than other models while being significantly smaller, and the model and code are publicly available. Why it matters: This research enables on-device AI processing for video, reducing reliance on remote servers and addressing privacy concerns, which can accelerate the adoption of AI in mobile and embedded applications.

Energy-Efficient and Secure EdgeAI Systems: From Architectures to Applications

MBZUAI ·

Muhammad Shafique from NYU Abu Dhabi discusses building energy-efficient and robust EdgeAI systems. The talk covers trends, challenges, and techniques for optimizing software and hardware stacks. These optimizations aim to enable embodied AI in autonomous systems, IoT-Healthcare, Industrial-IoT, and smart environments. Why it matters: The research addresses key challenges in deploying AI on resource-constrained edge devices in the GCC region, particularly regarding energy efficiency and security.

RightNow-Arabic-0.5B-Turbo: An Open Sub-1B Arabic Language Model via Vocabulary Injection and Edge-First Deployment

arXiv ·

RightNow-Arabic-0.5B-Turbo is a new 518M-parameter Arabic-specialized decoder LLM, built on Qwen2.5-0.5B, designed to bridge the gap between small multilingual and large Arabic-specialized models. Its development pipeline included adding 27,032 Arabic tokens via vocabulary injection, continued pretraining on 504M Arabic tokens, and fine-tuning with supervised instruction and direct preference optimization. The model achieved a 35.9% mean accuracy on three Arabic benchmarks (COPA-ar, Arabic HellaSwag, ArabicMMLU), outperforming all same-class open models and recovering 67% of SILMA-9B's mean accuracy at 1/18 the parameters, with all code and weights publicly released. Why it matters: This model significantly advances efficient Arabic NLP by providing a powerful, specialized sub-1B LLM suitable for edge deployment, making advanced Arabic AI more accessible and performant on resource-constrained devices.