Abdullah Ibne Hanif Arean

Presented EdgeNN Research in Cloud Computing Course Grateful to present “Breaking the Edge: Efficient Neural Network Inference on Integrated Edge Devices” in CSE 530 under Professor Dr. Md Abdur Razzaque sir at University of Dhaka. Key Takeaways: EdgeNN optimizes neural network inference on CPU-GPU integrated edge devices through semantic-aware memory management, hybrid execution, and adaptive task partitioning. Results: 3.97-8.80× speedup over edge CPUs, 24.79% latency reduction vs. cloud, and 29.14× better energy efficiency. The framework enables privacy-preserving, low-latency AIoT applications on resource-constrained hardware—highly relevant for my computer vision research! Credit: Zhang et al., IEEE Transactions on Cloud Computing, 2025 #CloudComputing #EdgeAI #MachineLearning #UniversityOfDhaka #Research