• Structural Engineering and Mechanics
• /
• v.85 no.4
• /
• pp.469-484
• /
• 2023

A deep recursive bidirectional Cuda Deep Neural Network Long Short Term Memory (Bi-CuDNNLSTM) layer is recruited in this paper to predict the entire force time histories, and the corresponding hysteresis and backbone curves of reinforced concrete (RC) bridge piers using experimental fast and slow cyclic tests. The proposed stacked Bi-CuDNNLSTM layers involve multiple uncertain input variables, including horizontal actuator displacements, vertical actuators axial loads, the effective height of the bridge pier, the moment of inertia, and mass. The functional application programming interface in the Keras Python library is utilized to develop a deep learning model considering all the above various input attributes. To have a robust and reliable prediction, the dataset for both the fast and slow cyclic tests is split into three mutually exclusive subsets of training, validation, and testing (unseen). The whole datasets include 17 RC bridge piers tested experimentally ten for fast and seven for slow cyclic tests. The results bring to light that the mean absolute error, as a loss function, is monotonically decreased to zero for both the training and validation datasets after 5000 epochs, and a high level of correlation is observed between the predicted and the experimentally measured values of the force time histories for all the datasets, more than 90%. It can be concluded that the maximum mean of the normalized error, obtained through Box-Whisker plot and Gaussian distribution of normalized error, associated with unseen data is about 10% and 3% for the fast and slow cyclic tests, respectively. In recapitulation, it brings to an end that the stacked Bi-CuDNNLSTM layer implemented in this study has a myriad of benefits in reducing the time and experimental costs for conducting new fast and slow cyclic tests in the future and results in a fast and accurate insight into hysteretic behavior of bridge piers.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.41 no.2
• /
• pp.123-131
• /
• 2021

The buffer is a key component of an engineered barrier system that safeguards the disposal of high-level radioactive waste. Buffers are located between disposal canisters and host rock, and they can restrain the release of radionuclides and protect canisters from the inflow of ground water. Since considerable heat is released from a disposal canister to the surrounding buffer, the thermal conductivity of the buffer is a very important parameter in the entire disposal safety. For this reason, a lot of research has been conducted on thermal conductivity prediction models that consider various factors. In this study, the thermal conductivity of a buffer is estimated using the machine learning methods of: linear regression, decision tree, support vector machine (SVM), ensemble, Gaussian process regression (GPR), neural network, deep belief network, and genetic programming. In the results, the machine learning methods such as ensemble, genetic programming, SVM with cubic parameter, and GPR showed better performance compared with the regression model, with the ensemble with XGBoost and Gaussian process regression models showing best performance.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.8
• /
• pp.446-460
• /
• 2011

The use of induction motors has been recently increasing with automation in aeronautical and automotive industries, and it playes a significant role. This has motivated that many researchers have studied on developing fault detection and classification systems of an induction motor in order to minimize economical damage caused by its fault. With this reason, this paper proposed feature vector extraction methods based on STE (short-time energy)+SVD (singular value decomposition) and DCT (discrete cosine transform)+SVD techniques to early detect and diagnose faults of induction motors, and classified faults of an induction motor into different types of them by using extracted features as inputs of BPNN (back propagation neural network) and multi-layer SVM (support vector machine). When BPNN and multi-lay SVM are used as classifiers for fault classification, there are many settings that affect classification performance: the number of input layers, the number of hidden layers and learning algorithms for BPNN, and standard deviation values of Gaussian radial basis function for multi-layer SVM. Therefore, this paper quantitatively simulated to find appropriate settings for those classifiers yielding higher classification performance than others.

• Journal of Internet Computing and Services
• /
• v.22 no.1
• /
• pp.23-30
• /
• 2021

This paper proposes a noise-tolerant image classification system using multiple autoencoders. The development of deep learning technology has dramatically improved the performance of image classifiers. However, if the images are contaminated by noise, the performance degrades rapidly. Noise added to the image is inevitably generated in the process of obtaining and transmitting the image. Therefore, in order to use the classifier in a real environment, we have to deal with the noise. On the other hand, the autoencoder is an artificial neural network model that is trained to have similar input and output values. If the input data is similar to the training data, the error between the input data and output data of the autoencoder will be small. However, if the input data is not similar to the training data, the error will be large. The proposed system uses the relationship between the input data and the output data of the autoencoder, and it has two phases to classify the images. In the first phase, the classes with the highest likelihood of classification are selected and subject to the procedure again in the second phase. For the performance analysis of the proposed system, classification accuracy was tested on a Gaussian noise-contaminated MNIST dataset. As a result of the experiment, it was confirmed that the proposed system in the noisy environment has higher accuracy than the CNN-based classification technique.

• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.3
• /
• pp.125-137
• /
• 2009

Acoustic emission-based techniques are being used for the nondestructive inspection of mechanical systems used in service. For reliable fault monitoring related to the crack growth, it is important to identify the dynamical characteristics as well as transient crack-related signals. Widely used methods which are based on physical phenomena of the three damage stages for detecting the crack growth have a problem that crack-related acoustic emission activities overlap in time, therefore it is insufficient to estimate the exact crack growth time. The proposed pattern recognition method uses the dynamical characteristics of acoustic emission as inputs for minimizing false alarms and miss alarms and performs the temporal clustering to estimate the crack growth time accurately. Experimental results show that the proposed method is effective for practical use because of its robustness to changes of acoustic emission caused by changes of pressure levels.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.9 no.5
• /
• pp.101-107
• /
• 2009

In this paper, we describe a method of detecting emergency situation using images from surveillance cameras and propose a mobile robot tracking system for detailed examination of that situation. We are able to track a few persons and recognize their actions by an analyzing image sequences acquired from a fixed camera on all sides of buildings. When emergency situation is detected, a mobile robot moves and closely examines the place where the emergency is occurred. In order to recognize actions of a few persons using a sequence of images from surveillance cameras images, we need to track and manage a list of the regions which are regarded as human appearances. Interest regions are segmented from the background using MOG(Mixture of Gaussian) model and continuously tracked using appearance model in a single image. Then we construct a MHI(Motion History Image) for a tracked person using silhouette information of region blobs and model actions. Emergency situation is finally detected by applying these information to neural network. And we also implement mobile robot tracking technology using the distance between the person and a mobile robot.

• Journal of the Korea Convergence Society
• /
• v.8 no.10
• /
• pp.1-8
• /
• 2017

Accurate forecasting enables to effectively prepare for future phenomenon. Especially, meteorological phenomenon is closely related with human life, and it can prevent from damage such as human life and property through forecasting of weather and disaster that can occur. To respond quickly and effectively to oil spill accidents, it is important to accurately predict the movement of oil spills and the weather in the surrounding waters. In this paper, we selected four representative machine learning techniques: support vector machine, Gaussian process, multilayer perceptron, and radial basis function network that have shown good performance and predictability in the previous studies related to oil spill detection and prediction in meteorology such as wind, rainfall and ozone. we suggest the applicability of oil spill prediction model based on machine learning.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.1
• /
• pp.13-22
• /
• 2013

Smartphone addiction has become more serious than internet addiction since people can download and run numerous applications with smartphones even without internet connection. However, smartphone addiction is not sufficiently dealt with in current studies. The S-scale method developed by Korea National Information Society Agency involves so many questions that respondents are likely to avoid the diagnosis itself. Moreover, since S-scale is determined by the total score of responded items without taking into account of demographic variables, it is difficult to get an accurate result. Therefore, in this paper, we have extracted important factors from all data, which affect smartphone addiction, including demographic variables. Then we classified the selected items with a neural network. The result of a comparative analysis with backpropagation learning algorithm and multiclass support vector machine shows that learning rate is slightly higher in multiclass SVM. Since multiclass SVM suggested in this paper is highly adaptable to rapid changes of data, we expect that it will lead to a more accurate self-diagnosis of smartphone addiction.

• Geomechanics and Engineering
• /
• v.25 no.1
• /
• pp.1-16
• /
• 2021

Machine learning models have been widely used for landslide susceptibility assessment (LSA) in recent years. The large number of inputs or conditioning factors for these models, however, can reduce the computation efficiency and increase the difficulty in collecting data. Feature selection is a good tool to address this problem by selecting the most important features among all factors to reduce the size of the input variables. However, two important questions need to be solved: (1) how do feature selection methods affect the performance of machine learning models? and (2) which feature selection method is the most suitable for a given machine learning model? This paper aims to address these two questions by comparing the predictive performance of 13 feature selection-based machine learning (FS-ML) models and 5 ordinary machine learning models on LSA. First, five commonly used machine learning models (i.e., logistic regression, support vector machine, artificial neural network, Gaussian process and random forest) and six typical feature selection methods in the literature are adopted to constitute the proposed models. Then, fifteen conditioning factors are chosen as input variables and 1,017 landslides are used as recorded data. Next, feature selection methods are used to obtain the importance of the conditioning factors to create feature subsets, based on which 13 FS-ML models are constructed. For each of the machine learning models, a best optimized FS-ML model is selected according to the area under curve value. Finally, five optimal FS-ML models are obtained and applied to the LSA of the studied area. The predictive abilities of the FS-ML models on LSA are verified and compared through the receive operating characteristic curve and statistical indicators such as sensitivity, specificity and accuracy. The results showed that different feature selection methods have different effects on the performance of LSA machine learning models. FS-ML models generally outperform the ordinary machine learning models. The best FS-ML model is the recursive feature elimination (RFE) optimized RF, and RFE is an optimal method for feature selection.