Neural Architecture Search for Efficient Convolutional Networks: A Multi-Objective Optimization Approach

Posted: Jan 11, 2024

• Wei Zhang
• Maria Rodriguez
• Kenji Tanaka
• Fatima Al-Mansoori

Abstract

This paper presents a novel neural architecture search (NAS) framework that optimizes convolutional neural networks for both accuracy and computational efficiency. Traditional NAS methods often focus solely on accuracy metrics, neglecting the practical constraints of deployment in resource-limited environments. Our approach employs a multi-objective evolutionary algorithm that simultaneously optimizes network accuracy, parameter count, and computational requirements. We evaluated our method on CIFAR-10 and ImageNet datasets, demonstrating that our discovered architectures achieve competitive accuracy while reducing computational costs by up to 45% compared to hand-designed networks. The proposed framework provides a systematic approach to designing efficient neural networks suitable for edge computing applications.