Deep Roots: Improving CNN Efficiency with Hierarchical Filter Groups

05/20/2016
by   Yani Ioannou, et al.
0

We propose a new method for creating computationally efficient and compact convolutional neural networks (CNNs) using a novel sparse connection structure that resembles a tree root. This allows a significant reduction in computational cost and number of parameters compared to state-of-the-art deep CNNs, without compromising accuracy, by exploiting the sparsity of inter-layer filter dependencies. We validate our approach by using it to train more efficient variants of state-of-the-art CNN architectures, evaluated on the CIFAR10 and ILSVRC datasets. Our results show similar or higher accuracy than the baseline architectures with much less computation, as measured by CPU and GPU timings. For example, for ResNet 50, our model has 40 45 For the deeper ResNet 200 our model has 25 44 GoogLeNet, our model has 7 (GPU).

READ FULL TEXT

page 5

page 11

page 12

page 13

page 14

page 15

research
06/09/2019

HGC: Hierarchical Group Convolution for Highly Efficient Neural Network

Group convolution works well with many deep convolutional neural network...
research
12/15/2014

MatConvNet - Convolutional Neural Networks for MATLAB

MatConvNet is an implementation of Convolutional Neural Networks (CNNs) ...
research
03/07/2018

HENet:A Highly Efficient Convolutional Neural Networks Optimized for Accuracy, Speed and Storage

In order to enhance the real-time performance of convolutional neural ne...
research
08/22/2016

Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures

Major winning Convolutional Neural Networks (CNNs), such as AlexNet, VGG...
research
02/17/2018

Towards Principled Design of Deep Convolutional Networks: Introducing SimpNet

Major winning Convolutional Neural Networks (CNNs), such as VGGNet, ResN...
research
03/03/2016

Decision Forests, Convolutional Networks and the Models in-Between

This paper investigates the connections between two state of the art cla...
research
08/07/2018

Efficient Fusion of Sparse and Complementary Convolutions for Object Recognition and Detection

We propose a new method for exploiting sparsity in convolutional kernels...

Please sign up or login with your details

Forgot password? Click here to reset