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.

A New Deep Neural Architecture Search Pipeline for Face Recognition.

Neural Architecture Search for Deep Face Recognition

Advances in neural architecture search

Neural Architecture Search with Progressive Evaluation and Sub-Population Preservation

Automatic improvement of deep learning-based cell segmentation in time-lapse microscopy by neural architecture search

PV-NAS: Practical Neural Architecture Search for Video Recognition

Robust 3D Face Alignment with Multi-Path Neural Architecture Search

NAS-FCOS: Fast neural architecture search for object detection

Neural Architecture Search Using Genetic Algorithm for Facial Expression Recognition

HZS-NAS: Neural Architecture Search with Hybrid Zero-Shot Proxy for Facial Expression Recognition

Neural Architecture Search for Microscopy Cell Segmentation

Searching for the Fakes: Efficient Neural Architecture Search for General Face Forgery Detection

Video Action Recognition with Neural Architecture Search.

Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation

Network Architecture Search for Domain Adaptation

XC-NAS: A New Cellular Encoding Approach for Neural Architecture Search of Multi-path Convolutional Neural Networks

FBNetV5: Neural Architecture Search for Multiple Tasks in One Run

RelativeNAS: Relative Neural Architecture Search Via Slow-Fast Learning

EAT-NAS: Elastic Architecture Transfer for Accelerating Large-scale Neural Architecture Search

Neural Architecture Search Survey: A Computer Vision Perspective