Convolutional Neural Networks at Constrained Time Cost (CVPR 2015) Authors : Kaiming He, Jian Sun (MSR) Presenter : Hyunjun Ju 1.

Slides:



Advertisements
Similar presentations
A brief review of non-neural-network approaches to deep learning
Advertisements

ImageNet Classification with Deep Convolutional Neural Networks
Large-Scale Object Recognition with Weak Supervision
Learning Convolutional Feature Hierarchies for Visual Recognition
DeeperVision and DeepInsight Solutions
Spatial Pyramid Pooling in Deep Convolutional
From R-CNN to Fast R-CNN
Convolutional Neural Networks for Image Processing with Applications in Mobile Robotics By, Sruthi Moola.
Hurieh Khalajzadeh Mohammad Mansouri Mohammad Teshnehlab
Deep Convolutional Nets
Convolutional Neural Network
Deep Residual Learning for Image Recognition
Lecture 3b: CNN: Advanced Layers
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition arXiv: v4 [cs.CV(CVPR)] 23 Apr 2015 Kaiming He, Xiangyu Zhang, Shaoqing.
Compression of CNNs Mooyeol Baek Xiangyu Zhang, Jianhua Zou, Xiang Ming, Kaiming He, Jian Sun: Efficient and Accurate Approximations of Nonlinear Convolutional.
When deep learning meets object detection: Introduction to two technologies: SSD and YOLO Wenchi Ma.
Wenchi MA CV Group EECS,KU 03/20/2017
Recent developments in object detection
Deep Residual Learning for Image Recognition
Learning to Compare Image Patches via Convolutional Neural Networks
Quantum Simulation Neural Networks
Computer Science and Engineering, Seoul National University
Krishna Kumar Singh, Yong Jae Lee University of California, Davis
Many slides and slide ideas thanks to Marc'Aurelio Ranzato and Michael Nielson.
Inception and Residual Architecture in Deep Convolutional Networks
Lecture 5 Smaller Network: CNN
Neural Networks 2 CS446 Machine Learning.
Training Techniques for Deep Neural Networks
Efficient Deep Model for Monocular Road Segmentation
Convolutional Networks
Deep Belief Networks Psychology 209 February 22, 2013.
CS6890 Deep Learning Weizhen Cai
Deep Residual Learning for Image Recognition
Dynamic Routing Using Inter Capsule Routing Protocol Between Capsules
ECE 599/692 – Deep Learning Lecture 6 – CNN: The Variants
Fully Convolutional Networks for Semantic Segmentation
Layer-wise Performance Bottleneck Analysis of Deep Neural Networks
Bird-species Recognition Using Convolutional Neural Network
Computer Vision James Hays
Introduction to Neural Networks
Image Classification.
NormFace:
Learning Hierarchical Features from Generative Models
Incremental Training of Deep Convolutional Neural Networks
Deep Learning Hierarchical Representations for Image Steganalysis
Very Deep Convolutional Networks for Large-Scale Image Recognition
Smart Robots, Drones, IoT
CSC 578 Neural Networks and Deep Learning
Lecture: Deep Convolutional Neural Networks
Use 3D Convolutional Neural Network to Inspect Solder Ball Defects
Outline Background Motivation Proposed Model Experimental Results
Visualizing and Understanding Convolutional Networks
Analysis of Trained CNN (Receptive Field & Weights of Network)
RCNN, Fast-RCNN, Faster-RCNN
Designing Neural Network Architectures Using Reinforcement Learning
Deep Learning Some slides are from Prof. Andrew Ng of Stanford.
Mihir Patel and Nikhil Sardana
ImageNet Classification with Deep Convolutional Neural Networks
Inception-v4, Inception-ResNet and the Impact of
Heterogeneous convolutional neural networks for visual recognition
CSC 578 Neural Networks and Deep Learning
Model Compression Joseph E. Gonzalez
Natalie Lang Tomer Malach
CS295: Modern Systems: Application Case Study Neural Network Accelerator Sang-Woo Jun Spring 2019 Many slides adapted from Hyoukjun Kwon‘s Gatech “Designing.
DRC with Deep Networks Tanmay Lagare, Arpit Jain, Luis Francisco,
VERY DEEP CONVOLUTIONAL NETWORKS FOR LARGE-SCALE IMAGE RECOGNITION
Object Detection Implementations
Example of training and deployment of deep convolutional neural networks. Example of training and deployment of deep convolutional neural networks. During.
CSC 578 Neural Networks and Deep Learning
CVPR2019 Jiahe Li SiamRPN introduces the region proposal network after the Siamese network and performs joint classification and regression.
Presentation transcript:

Convolutional Neural Networks at Constrained Time Cost (CVPR 2015) Authors : Kaiming He, Jian Sun (MSR) Presenter : Hyunjun Ju 1

Motivation Most of the recent advanced CNNs are time consuming. They take a high-end GPU or multiple GPUs one week or several weeks to train, which can sometimes be too demanding for the rapidly changing industry. This paper investigates the accuracy of CNN architecture at constrained time cost. Factors : depth, width(the number of filter), filter size, stride. 2 Goal Find the efficient and relatively accurate CNN model.

Time Complexity of Convolutions 3 The time cost of fc layers and pooling layers is not involved in the above formulation. These layers often take 5-10% computational time.

Baseline Model 4 Full Full Full8 Softmax 1000 SPP (spatial pyramid pooling)

3-stage Design 5 Stage Layers between two nearby pooling layers. convolution pooling convolution pooling Stage 2 Stage 1 conv pooling Stage 3

Model Designs by Layer Replacement 6 Model A is baseline. The others are variation of the model A. In each model, a few layers are replaced with some other layers that preserve time costs. There are trade offs by replacing the layers.

Trade-offs between Depth and Filter Sizes 7

8 The depth is more important than the filter sizes. When the time complexity is roughly the same, the deeper networks with smaller filters show better results than the shallower networks with large filters.

Trade-offs between Depth and Width 9

Increasing the depth leads to considerable gains, even the width needs to be properly reduced. 10 But, G is only better than F marginally.

Trade-offs between Width and Filter Size 11

Trade-offs between Width and Filter Size Unlike the depth that has a high priority, the width and filter sizes do not show apparent priorities to each other. 12

But, Is Deeper Always better? In experiments, they find the accuracy is stagnant or even reduced in some of their very deep attempts. Two possible explanations 1.The width/filter sizes are reduced overly and may harm the accuracy. 2.Overly increasing the depth will degrade the accuracy even if the other factors are not traded. 13

But, Is Deeper Always better? To understand the main reason Do not constrain the time complexity (just add conv layers) 14 Overly increasing depth can harm the accuracy, even if the width/filter sizes are unchanged. The errors not only get saturated at some point, but get worse if going deeper. The degradation is not due to over-fitting. (training errors are also worse)

Adding Pooling Layer(Feature map size and width) 15 The model J(low feature map size, high width) results in new error rates which is better than that of model E.

Delayed Subsampling of Pooling Layers Max pooling Layer has two different role 1.Lateral suppression that increase the invariance to small local translation 2.Reducing the spatial size of feature maps by subsampling Usually, max pooling layer plays the two roles simultaneously with stride > 1. We can separate these two roles using two different layers. Pooling layer : by setting the stride = 1 A convolutional layer : by setting the stride > 1 (original stride of the pooling layer) this operation doesn’t change the complexity of all convolutional layer. 16 Delayed model has lower Top-5 error rates than original model’s.

Comparisons (Fast models) 17 (J’) Model J’ has the best performance and complexity is low.

Comparisons (Accurate models) VGG-16 and GoogLeNet are trained by more additional data augmentation than others. 18 Model J’ has low complex and relatively good performance.

Conclusion Constrained time cost is practical issue in industrial and commercial requirement. They proposed models that are fast for practical applications yet are more accurate than existing fast models Not the best performance model or the fastest model. 19

20