Fast Neural Architecture Construction using EnvelopeNets

Abstract

Fast Neural Architecture Construction (NAC) is a method to construct deep network architectures by pruning and expansion of a base network. In recent years, several automated search methods for neural network architectures have been proposed using methods such as evolutionary algorithms and reinforcement learning. These methods use a single scalar objective function (usually accuracy) that is evaluated after a full training and evaluation cycle. In contrast NAC directly compares the utility of different filters using statistics derived from filter featuremaps reach a state where the utility of different filters within a network can be compared and hence can be used to construct networks. The training epochs needed for filters within a network to reach this state is much less than the training epochs needed for the accuracy of a network to stabilize. NAC exploits this finding to construct convolutional neural nets (CNNs) with close to state of the art accuracy, in < 1 GPU day, faster than most of the current neural architecture search methods. The constructed networks show close to state of the art performance on the image classification problem on well known datasets (CIFAR-10, ImageNet) and consistently show better performance than hand constructed and randomly generated networks of the same depth, operators and approximately the same number of parameters.

arXiv commentA shorter version of this paper appeared in the Workshop on MetaLearning 2018 (MetaLearning 2018 at NeurIPS 2018)

References in corpus (19)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Residual Learning for Image Recognition
• Neural Architecture Search with Reinforcement Learning
• Densely Connected Convolutional Networks
• Pruning Convolutional Neural Networks for Resource Efficient Inference
• Pruning Filters for Efficient ConvNets
• PathNet: Evolution Channels Gradient Descent in Super Neural Networks
• Efficient Neural Architecture Search via Parameter Sharing
• Neural Architecture Optimization
• Visualizing and Understanding Convolutional Networks
• SMASH: One-Shot Model Architecture Search through HyperNetworks
• Xception: Deep Learning with Depthwise Separable Convolutions
• Neural Architecture Search with Bayesian Optimisation and Optimal Transport
• Provable Bounds for Learning Some Deep Representations
• Simple And Efficient Architecture Search for Convolutional Neural Networks
• Convolutional Neural Fabrics
• DeepArchitect: Automatically Designing and Training Deep Architectures
• Highway and Residual Networks learn Unrolled Iterative Estimation
• Practical Block-wise Neural Network Architecture Generation

Cited by in corpus (6)

• ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
• Weight-Sharing Neural Architecture Search: A Battle to Shrink the Optimization Gap
• NASIB: Neural Architecture Search withIn Budget
• Direct Federated Neural Architecture Search
• EENA: Efficient Evolution of Neural Architecture
• Revisiting Neural Architecture Search