Tags:
A recent research article led by Muhammad Zawish, PhD Student at Walton Institute at South East Technological University (SETU), along with supervisors Dr. Lizy Abraham and Dr. Steven Davy, has been published in the prestigious IEEE Transactions on Artificial Intelligence. The study focuses on making complex AI (deep learning) models more efficient for resource-constrained and energy-efficient execution on mobile edge devices.
In simple terms, the research addresses the challenges of deploying advanced deep learning models like VGG-16 and ResNets for computer vision on devices with limited resources. These AI models have too many parameters and floating-point operations, making them impractical for low-power devices. To overcome this, Zawish proposes a method called network pruning, a form of model compression that accelerates CNNs on low-power devices.
The article, titled “Complexity-Driven Model Compression for Resource-constrained Deep Learning on Edge” can be found HERE.
Zawish says, “I believe this research contributes to the ongoing efforts to make AI models more accessible and efficient for a variety of applications, particularly in scenarios with limited resources.”
This research was carried out under the SFI VistaMilk Research Centre supported by Science Foundation Ireland and the Department of Agriculture, Food and Marine on behalf of the Government of Ireland under grant 16/RC/3835.