Abstract:Although group convolution operators are increasingly used in deep convolutional neural networks to improve the computational efficiency and to reduce the number of parameters, most existing methods construct their group convolution architectures by a predefined partitioning of the filters of each convolutional layer into multiple regular filter groups with an equal spatial group size and data-independence, which prevents a full exploitation of their potential. To tackle this issue, we propose a novel method of designing self-grouping convolutional neural networks, called SG-CNN, in which the filters of each convolutional layer group themselves based on the similarity of their importance vectors. Concretely, for each filter, we first evaluate the importance value of their input channels to identify the importance vectors, and then group these vectors by clustering. Using the resulting data-dependent centroids, we prune the less important connections, which implicitly minimizes the accuracy loss of the pruning, thus yielding a set of diverse group convolution filters. Subsequently, we develop two fine-tuning schemes, i.e. (1) both local and global fine-tuning and (2) global only fine-tuning, which experimentally deliver comparable results, to recover the recognition capacity of the pruned network. Comprehensive experiments carried out on the CIFAR-10/100 and ImageNet datasets demonstrate that our self-grouping convolution method adapts to various state-of-the-art CNN architectures, such as ResNet and DenseNet, and delivers superior performance in terms of compression ratio, speedup and recognition accuracy. We demonstrate the ability of SG-CNN to generalise by transfer learning, including domain adaption and object detection, showing competitive results. Our source code is available at <a href="https://github.com/QingbeiGuo/SG-CNN.git">https://github.com/QingbeiGuo/SG-CNN.git</a>.

Characterization and Design of Generalized Convolutional Neural Network

Group Action Equivariance And Generalized Convolution In Multi-Layer Neural Networks

Group Equivariant Convolutional Networks

A Dynamic Group Equivariant Convolutional Networks for Medical Image Analysis.

Analyzing and Quantifying Generalization in Convolutional Neural Networks

Deep Global Generalized Gaussian Networks

Robust Designs For Gray-Scale Global Connectivity Detection Cnn Templates

Unified GCNs: Towards Connecting GCNs with CNNs

On the VC dimension of deep group convolutional neural networks

Generalizing the Convolution Operator in Convolutional Neural Networks

Geometric Operator Convolutional Neural Network

A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects

A Unified Framework for Convolution-Based Graph Neural Networks

Dynamic Group Convolution for Accelerating Convolutional Neural Networks

Understanding the Effect of GCN Convolutions in Regression Tasks

Self-grouping convolutional neural networks

Generalizing Graph Convolutional Networks Via Heat Kernel

Learn Generalization Feature Via Convolutional Neural Network: A Fault Diagnosis Scheme Toward Unseen Operating Conditions

Brain-inspired Dual-Pathway Neural Network Architecture and Its Generalization Analysis

VarGNet: Variable Group Convolutional Neural Network for Efficient Embedded Computing