Suggestions
Asmaa Ibrahim
AI & Cloud Solution Manager - Mentor
Asmaa Ibrahim is an accomplished professional with a diverse educational background and extensive experience in the technology industry.
She holds a bachelor's degree in Computer Engineering & Systems from Zagazig University and has furthered her education by studying AI Product Management at Udacity, Pre-Master in Management of Technology at Nile University, and a scholarship in Open Source Technologies at ITI.
Asmaa has held various key roles at Vodafone, starting as a Technology Consultant at ITIDA and transitioning to roles such as Sr. Cloud Engineer at ITI, Cognitive Business Architect, Sr. Solution Architect, Cognitive Solutions Architect, AI/ML Solution Owner, AI Product Owner, and currently serving as an AI&Cloud Solution Manager.
Her experience spans across different domains including Artificial Intelligence, Cloud Solutions, Machine Learning, and Cognitive Solutions, showcasing her expertise in these areas.
Aside from her corporate experience, Asmaa has also sharpened her skills through trainee positions at Information Technology Institute (ITI) and in roles like Linux System Engineer at ValueSYS.
Asmaa Ibrahim's career trajectory exemplifies a strong commitment to continuous learning and growth within the technology sector.
Highlights
》Building on my earlier post about ways to make Large Language Models (LLMs) better, 1-bit approach is method that's been around since 2016 aimed at reducing the high processing demands of neural networks.
☆ Below are several studies that explore this technique and its promising uses.
■ Known as: Binarized Neural Networks (BNN) ■ Problem: High computational and energy costs of traditional LLMs. ■ Solution: Use ternary model parameters (-1, 0, 1) for efficiency in latency, memory, throughput, and energy. ■ Example of Applications: ● Mobile and Telecom: Enhancing on-device AI ● Healthcare: On-device patient monitoring. ● Automotive: Improved navigation and voice-activated controls. ● Edge Computing: Facilitating real-time data processing at the edge of the network, reducing the need for data to travel to centralized cloud servers,
☆ Related Papers: ● arXiv:2402.10193v2 ● arXiv:2402.17764v1 ●https://lnkd.in/ePjrV82p ●https://lnkd.in/e-FU5-Cb
