About this document

Accelerating Multi-Objective Neural Architecture Search by Random-Weight Evaluation - 2021 by Haibing Li is a document available to read on EtoBox.

This article presents a novel approach to neural architecture search (NAS) using a performance estimation metric called random-weight evaluation (RWE), which allows for efficient evaluation of CNNs by only training the last layer. The proposed method significantly reduces computational costs and achieves state-of-the-art performance on real-world datasets like CIFAR-10 and ImageNet. Additionally, it employs a multi-objective evolutionary algorithm to balance model performance and complexity, demonstrating i

Author

Haibing Li

Language

EN