Automated Optimal Architecture of Deep Convolutional Neural Networks for Image Recognition

Albelwi, Saleh; Mahmood, Ausif

doi:10.1109/icmla.2016.0018

Cited by 17 publications

(13 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The training of a CNN can be varied by adjusting values of parameters such as the number of layers in a neural network, the number of filters in each convolutional layer and the order of layers. At present, there is no established theory to determine the optimal values of these parameters as well as some inherent training parameters such as the learning rate and batch size (Albelwi and Mahmood, 2016). However, these parameters can be tuned in a iterative way to find a near-optimal configuration.…”

Section: Training the Cnn-smmcmentioning

confidence: 99%

A ship movement classification based on Automatic Identification System (AIS) data using Convolutional Neural Network

et al. 2020

View full text Add to dashboard Cite

With a wide use of AIS data in maritime transportation, there is an increasing demand to develop algorithms to efficiently classify a ship's AIS data into different movements (static, normal navigation and manoeuvring). To achieve this, several studies have been proposed to use labeled features but with the drawback of not being able to effectively extract the details of ship movement information. In addition, a ship movement is in a free space, which is different to a road vehicle's movement in road grids, making it inconvenient to directly migrate the methods for GPS data classification into AIS data. To deal with these problems, a Convolutional Neural Network-Ship Movement Modes Classification (CNN-SMMC) algorithm is proposed in this paper. The underlying concept of this method is to train a neural network to learn from the labeled AIS data, and the unlabeled AIS data can be effectively classified by using this trained network. More specifically, a Ship Movement Image Generation and Labelling (SMIGL) algorithm is first designed to convert a ship's AIS trajectories into different movement images to make a full use of the CNN's classification ability. Then, a CNN-SMMC architecture is built with a series of functional layers (convolutional layer, max-pooling layer, dense layer etc.) for ship movement classification with seven experiments been designed to find the optimal parameters for the CNN-SMMC. Considering the imbalanced features of AIS data, three metrics (average accuracy, 1 score and Area Under Curve (AUC)) are selected to evaluate the performance of the CNN-SMMC. Finally, several benchmark classification algorithms (K-Nearest Neighbors (KNN), Support Vector Machine (SVM) and Decision Tree (DT)) are selected to compare with CNN-SMMC. The results demonstrate that the proposed CNN-SMMC has a better performance in the classification of AIS data.

show abstract

Section: Training the Cnn-smmcmentioning

confidence: 99%

A ship movement classification based on Automatic Identification System (AIS) data using Convolutional Neural Network

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The NMA was used in Convolutional Neural Network (CNN) optimization in [1,2], in conjunction with a relatively small optimization dataset. It works well for objective functions that are smooth, unimodal and not too noisy.…”

Section: Derivative-free Optimization: Nelder-meadmentioning

confidence: 99%

“…The progressive sampling method presented in [40] is an incremental method using progressively larger samples as long as model accuracy improves. Instance selection algorithms for CNN architectures were proposed in [1,2,28].…”

Section: Filter the Selection Criterion Uses A Selection Functionmentioning

confidence: 99%

See 1 more Smart Citation

Hyperparameter optimization in learning systems

Andonie

2019

J Membr Comput

View full text Add to dashboard Cite

While the training parameters of machine learning models are adapted during the training phase, the values of the hyperparameters (or meta-parameters) have to be specified before the learning phase. The goal is to find a set of hyperparameter values which gives us the best model for our data in a reasonable amount of time. We present an integrated view of methods used in hyperparameter optimization of learning systems, with an emphasis on computational complexity aspects. Our thesis is that we should solve a hyperparameter optimization problem using a combination of techniques for: optimization, search space and training time reduction. Case studies from real-world applications illustrate the practical aspects. We create the framework for a future separation between parameters and hyperparameters in adaptive P systems.

show abstract

“…Furthermore, large variations in image quality and appearances may occur due to differences in acquisition, reconstruction or post-processing between different vendors and/or acquisition protocols, and require algorithm retraining to achieve optimal performance. Deep learning approaches, that are rapidly becoming the state of the art for many applications, also require the manual optimization of several hyper-parameters (Greenspan et al, 2016;Albelwi and Mahmood, 2016). Self-optimization of machine learning pipelines is a common problem in computer vision, as well as in the general machine learning and data analytics community (Cerquitelli et al, 2016;Corso et al, 2017).…”

Section: Introductionmentioning

confidence: 99%

Optimization of computer aided detection systems: An evolutionary approach

Morra

Coccia

Cerquitelli

2018

Expert Systems with Applications

View full text Add to dashboard Cite

Computer Aided Diagnosis (CAD) systems are designed to aid the radiologist in interpreting medical images. They are usually based on lesion detection and segmentation algorithms whose performance depends on a large number of parameters. While time consuming and sub-optimal, manual adjustment is still widely used to adjust parameter values. Genetic or evolutionary algorithms (GA) are effective optimization methods that mimic biological evolution. Genetic algorithms have been shown to efficiently manage complex search spaces, and can be applied to all kinds of objective functions, including discontinuous, nondifferentiable, or highly nonlinear ones. In this study, we have adopted an evolutionary approach to the problem of parameter optimization. We show that the genetic algorithm is able to effectively converge to a better solution than manual optimization on a case study for digital breast tomosynthesis CAD. Parameter optimization was framed as a constrained optimization problem, where the function to be maximized was defined as weighted sum of sensitivity, false positive rate and segmentation accuracy. A modified Dice coefficient was defined to assess the segmentation quality of individual lesions. Finally, all viable solutions evaluated by the GA were studied by means of exploratory data analysis techniques, such as association rules, to gain useful insight on the strength of the influence of each parameter on overall algorithm performance. We showed that this combination was able to identify multiple ranges of viable solutions with good segmentation accuracy.

show abstract

Automated Optimal Architecture of Deep Convolutional Neural Networks for Image Recognition

Cited by 17 publications

References 15 publications

A ship movement classification based on Automatic Identification System (AIS) data using Convolutional Neural Network

A ship movement classification based on Automatic Identification System (AIS) data using Convolutional Neural Network

Hyperparameter optimization in learning systems

Optimization of computer aided detection systems: An evolutionary approach

Contact Info

Product

Resources

About