Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition

碩士 === 國立中央大學 === 數學系 === 107 === In recent years, deep learning has flourished and people have begun to use deep learning to solve problems. Deep neural networks can be used for speech recognition, image recognition, object detection, face recognition, or driverless. The most basic neural network i...

Full description

Bibliographic Details
Main Authors: Jing-Wun Chen, 陳靖玟
Other Authors: Feng-Nan Hwang
Format: Others
Language:en_US
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/fpx347
id ndltd-TW-107NCU05479012
record_format oai_dc
spelling ndltd-TW-107NCU054790122019-10-22T05:28:09Z http://ndltd.ncl.edu.tw/handle/fpx347 Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition Jing-Wun Chen 陳靖玟 碩士 國立中央大學 數學系 107 In recent years, deep learning has flourished and people have begun to use deep learning to solve problems. Deep neural networks can be used for speech recognition, image recognition, object detection, face recognition, or driverless. The most basic neural network is the Multilayer Perceptron (MLP), which consists of multiple node layers, each layer is fully connected to each other, and one of the drawbacks of MLP is that it ignores the shape of the data which is important for image data. Compare to traditional neural networks, the convolutional neural network (CNN) has additional convolution and pooling layers which are used for preserving and capturing image features. The accuracy rate for prediction using neural network depends on many factors, such as the architecture of neural networks, the cost functions, and the selection of an optimizer. The goal of this work is to investigate the effects of optimizer selection and their hyperparameter tuning on the performance of deep neural networks for image recognition problems. We use three data sets including MNIST, CIFAR-10 and train route scenarios as test problems and test six optimizers (Gradient descent, Momentum, Adaptive gradient algorithm, Adadelta, Root Mean Square Propagation, and Adam). Our numerical results show that Adam is a good choice because of its efficiency and robustness. Feng-Nan Hwang 黃楓南 2019 學位論文 ; thesis 56 en_US
collection NDLTD
language en_US
format Others
sources NDLTD
description 碩士 === 國立中央大學 === 數學系 === 107 === In recent years, deep learning has flourished and people have begun to use deep learning to solve problems. Deep neural networks can be used for speech recognition, image recognition, object detection, face recognition, or driverless. The most basic neural network is the Multilayer Perceptron (MLP), which consists of multiple node layers, each layer is fully connected to each other, and one of the drawbacks of MLP is that it ignores the shape of the data which is important for image data. Compare to traditional neural networks, the convolutional neural network (CNN) has additional convolution and pooling layers which are used for preserving and capturing image features. The accuracy rate for prediction using neural network depends on many factors, such as the architecture of neural networks, the cost functions, and the selection of an optimizer. The goal of this work is to investigate the effects of optimizer selection and their hyperparameter tuning on the performance of deep neural networks for image recognition problems. We use three data sets including MNIST, CIFAR-10 and train route scenarios as test problems and test six optimizers (Gradient descent, Momentum, Adaptive gradient algorithm, Adadelta, Root Mean Square Propagation, and Adam). Our numerical results show that Adam is a good choice because of its efficiency and robustness.
author2 Feng-Nan Hwang
author_facet Feng-Nan Hwang
Jing-Wun Chen
陳靖玟
author Jing-Wun Chen
陳靖玟
spellingShingle Jing-Wun Chen
陳靖玟
Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
author_sort Jing-Wun Chen
title Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
title_short Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
title_full Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
title_fullStr Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
title_full_unstemmed Exploring Effects of Optimizer Selection and Their Hyperparameter Tuning on Performance of Deep Neural Networks for Image Recognition
title_sort exploring effects of optimizer selection and their hyperparameter tuning on performance of deep neural networks for image recognition
publishDate 2019
url http://ndltd.ncl.edu.tw/handle/fpx347
work_keys_str_mv AT jingwunchen exploringeffectsofoptimizerselectionandtheirhyperparametertuningonperformanceofdeepneuralnetworksforimagerecognition
AT chénjìngwén exploringeffectsofoptimizerselectionandtheirhyperparametertuningonperformanceofdeepneuralnetworksforimagerecognition
_version_ 1719273955864346624