Automatically Evolving CNN Architectures Based on Blocks

TL;DR

This paper introduces an automatic method to evolve CNN architectures using genetic algorithms based on ResNet and DenseNet blocks, eliminating the need for domain expertise or pre/post-processing, and outperforming many existing methods in accuracy and efficiency.

Contribution

It presents a fully automatic CNN architecture design algorithm that does not require domain knowledge or pre/post-processing, and demonstrates superior performance and efficiency.

Findings

Outperforms state-of-the-art CNNs on CIFAR10 and CIFAR100 in accuracy.

Consumes less time than most peer methods to find optimal architectures.

Achieves competitive accuracy with semi-automatic approaches.

Abstract

The performance of Convolutional Neural Networks (CNNs) highly relies on their architectures. In order to design a CNN with promising performance, extended expertise in both CNNs and the investigated problem is required, which is not necessarily held by every user interested in CNNs or the problem domain. In this paper, we propose to automatically evolve CNN architectures by using a genetic algorithm based on ResNet blocks and DenseNet blocks. The proposed algorithm is \textbf{completely} automatic in designing CNN architectures, particularly, neither pre-processing before it starts nor post-processing on the designed CNN is needed. Furthermore, the proposed algorithm does not require users with domain knowledge on CNNs, the investigated problem or even genetic algorithms. The proposed algorithm is evaluated on CIFAR10 and CIFAR100 against 18 state-of-the-art peer competitors. Experimental results show that it outperforms state-of-the-art CNNs hand-crafted and CNNs designed by automatic peer competitors in terms of the classification accuracy, and achieves the competitive classification accuracy against semi-automatic peer competitors. In addition, the proposed algorithm consumes much less time than most peer competitors in finding the best CNN architectures.