Log-DenseNet: How to Sparsify a DenseNet

Abstract

Skip connections are increasingly utilized by deep neural networks to improve accuracy and cost-efficiency. In particular, the recent DenseNet is efficient in computation and parameters, and achieves state-of-the-art predictions by directly connecting each feature layer to all previous ones. However, DenseNet's extreme connectivity pattern may hinder its scalability to high depths, and in applications like fully convolutional networks, full DenseNet connections are prohibitively expensive. This work first experimentally shows that one key advantage of skip connections is to have short distances among feature layers during backpropagation. Specifically, using a fixed number of skip connections, the connection patterns with shorter backpropagation distance among layers have more accurate predictions. Following this insight, we propose a connection template, Log-DenseNet, which, in comparison to DenseNet, only slightly increases the backpropagation distances among layers from 1 to ($1+{\log }_{2}L$), but uses only $L{\log }_{2}L$ total connections instead of $O(L^2)$. Hence, Log-DenseNets are easier than DenseNets to implement and to scale. We demonstrate the effectiveness of our design principle by showing better performance than DenseNets on tabula rasa semantic segmentation, and competitive results on visual recognition.

References in corpus (7)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Distilling the Knowledge in a Neural Network
• Do Deep Nets Really Need to be Deep?
• Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation
• Deep Roots: Improving CNN Efficiency with Hierarchical Filter Groups
• Deep Networks with Stochastic Depth
• Geometric Context from Videos

Cited by in corpus (7)

• Gated Convolutional Networks with Hybrid Connectivity for Image Classification
• ThreshNet: An Efficient DenseNet Using Threshold Mechanism to Reduce Connections
• Convolutional Networks with Dense Connectivity
• HarDNet: A Low Memory Traffic Network
• Sparsely Aggregated Convolutional Networks
• A novel Deep Structure U-Net for Sea-Land Segmentation in Remote Sensing Images
• Comb Convolution for Efficient Convolutional Architecture