Abstract:Neural architecture search (NAS) is an effective approach for automating the design of deep neural networks. Evolutionary computation (EC) is commonly used in NAS due to its global optimization capability. However, the evaluation phase of architecture candidates in EC-based NAS is compute-intensive, limiting its application for many real-world problems. To overcome this challenge, we propose a novel progressive evaluation strategy for the evaluation phase in convolutional neural network architecture search, in which the number of training epochs of network individuals is progressively increased. In addition, a sub-population preservation strategy is proposed to preserve medium-size and large-size architectures to avoid prematurely discarding networks that may not perform well in the early stages but have the potential to excel with further optimization. Our proposed algorithm reduces the computational cost of the evaluation phase and promotes population diversity and fairness by preserving promising networks based on their distribution. We evaluate the proposed progressive evaluation and sub-population preservation of neural architecture search (PEPNAS) algorithm on the CIFAR10, CIFAR100, and ImageNet benchmark datasets, and compare it with 36 state-of-the-art algorithms, including manually designed networks, reinforcement learning (RL) algorithms, gradient-based algorithms, and other EC-based ones. The experimental results demonstrate that PEPNAS effectively identifies networks with competitive accuracy while also markedly improving the efficiency of the search process. For instance, PEPNAS discovers the architecture on CIFAR10 with a low error rate of 2.38% using only 0.7 GPU days. We directly adopt the searched architecture for the image classification on the CIFAR100 and ImageNet datasets, which achieves the top 1 error rates of 16.46% and 26.25%, respectively. The code is available at https://github.com/chajiajie/PEPNAS.

TS-ENAS:Two-Stage Evolution for Cell-based Network Architecture Search

A Hierarchical Evolution of Neural Architecture Search Method Based on State Transition Algorithm

A Fast Evolutionary Knowledge Transfer Search for Multiscale Deep Neural Architecture

Evolutionary Neural Network Architecture Search

Evolutionary Neural Architecture Search Based on Evaluation Correction and Functional Units

Multiple Population Alternate Evolution Neural Architecture Search

Automated Design of CNN Architecture Based on Efficient Evolutionary Search

Evolutionary Neural Architecture Search Combining Multi-Branch ConvNet and Improved Transformer

Reinforced Evolutionary Neural Architecture Search

Efficient Evolutionary Neural Architecture Search Based on Hybrid Search Space

A Gradient-Guided Evolutionary Neural Architecture Search

Evolving Search Space for Neural Architecture Search

Evolutionary Architecture Search Via Adaptive Parameter Control and Gene Potential Contribution

Toward Evolutionary Multitask Convolutional Neural Architecture Search

Neural Architecture Search Based on A Multi-objective Evolutionary Algorithm With Probability Stack

Surrogate-assisted Evolutionary Neural Architecture Search with Network Embedding

G-EvoNAS: Evolutionary Neural Architecture Search Based on Network Growth

Neural Architecture Search with Progressive Evaluation and Sub-Population Preservation

Stage-Wise Neural Architecture Search

Enhanced Gradient for Differentiable Architecture Search

EENA: Efficient Evolution of Neural Architecture.