• International Journal of Fluid Machinery and Systems
• /
• v.8 no.4
• /
• pp.264-273
• /
• 2015

This presentation describes an experimental approach for the detection of cavitation in hydraulic machines by use of ultrasonic signal analysis. Instead of using the high frequency pulses (typically 1MHz) only for transit time measurement different other signal characteristics are extracted from the individual signals and its correlation function with reference signals in order to gain knowledge of the water conditions. As the pulse repetition rate is high (typically 100Hz), statistical parameters can be extracted of the signals. The idea is to find patterns in the parameters by a classifier that can distinguish between the different water states. This classification scheme has been applied to different cavitation sections: a sphere in a water flow in circular tube at the HSLU in Lucerne, a NACA profile in a cavitation tunnel and two Francis model test turbines all at LMH in Lausanne. From the signal raw data several statistical parameters in the time and frequency domain as well as from the correlation function with reference signals have been determined. As classifiers two methods were used: neural feed forward networks and decision trees. For both classification methods realizations with lowest complexity as possible are of special interest. It is shown that two to three signal characteristics, two from the signal itself and one from the correlation function are in many cases sufficient for the detection capability. The final goal is to combine these results with operating point, vibration, acoustic emission and dynamic pressure information such that a distinction between dangerous and not dangerous cavitation is possible.

• Journal of Advanced Marine Engineering and Technology
• /
• v.36 no.5
• /
• pp.645-661
• /
• 2012

Recently, pirates are infesting on the sea and they have been hijacking the several vessels for example Samho Dream and Samho Jewelry of Korea. One of the items to reduce the risk is to adopt the invader detection system. If the pirates break in to the ship, the detection system can monitor the pirates and then call the security alarm. The crew can gain time to hide to the safe room and the report can be automatically sent to the control room to cope with the situation. For the invader detection, an unmanned observation system was proposed using the image detection algorithm that extracts the invader image from the recording image. To detect the motion area, the difference value was calculated between the current image and the prior image of the invader, and the 'AND' operator was used in calculated image and edge line. The image noise was reduced based on the morphology operation and then the image was transformed into morphological information. Finally, a neural network model was applied to recognize the invader. In the experimental results, it was confirmed that the proposed approach can improve the performance of the recognition in the invader monitoring system.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.26 no.12
• /
• pp.9-20
• /
• 2012

This paper proposes the efficiency optimization control of interior permanent magnet synchronous motor(IPMSM) drive using series connected-fuzzy neural network PI(SC-FNPI) controller. The PI controller is generally used to control IPMSM drive in industrial field. However, the PI controller has problem which is falling control performance about parameter variation such as command speed, load torque and inertia due to fixed gain of PI controller. Therefore, to improve performance of PI controller, this paper proposes SC-FNPI controller adjusted input of PI controller by FNN controller according to operating conditions. Also, this paper proposes efficiency optimization control which is improving efficiency with minimize loss. The SC-FNPI controller proposed in this paper is compared control performance with conventional FNN and PI controller about command speed, load torque and inertia variation. And the efficiency optimization control is compared with $i_d=0$ control about loss and efficiency. The SC-FNPI controller proposed in this paper shows more excellent control performance for rising time, overshoot and steady-state error. Also efficiency optimization control is increased efficiency by reducing loss.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2004.11a
• /
• pp.244-249
• /
• 2004

There have been many attempts to predict the secondary structure content of a protein from its primary sequence, which serves as the first step in a series of bioinformatics processes to gain knowledge of the structure and function of a protein. Most of them assumed that prediction relying on the information of the amino acid composition of a protein can be successful. Several approaches expanded the amount of information by including the pair amino acid composition of two adjacent residues. Recent methods achieved a remarkable improvement in prediction accuracy by using this expanded composition information. The overall average errors of two successful methods were 6.1% and 3.4%. This work was motivated by the observation that evolutionarily related proteins share the similar structure. After manipulating the values of the frequency matrix obtained by running PSI-BLAST, inputs of an artificial neural network were constructed by taking the ratio of the amino acid composition of the evolutionarily related proteins with a query protein to the background probability. Although we did not utilize the expanded composition information of amino acid pairs, we obtained the comparable accuracy, with the overall average error being 3.6%.

• Journal of Institute of Control, Robotics and Systems
• /
• v.5 no.6
• /
• pp.744-752
• /
• 1999

In this paper, we studied the response characteristics of $\alpha$, $\beta$ separated type, combined type, PI typed, and feedforward type in 2DOF-PID controller through the simulation and the experiments designed with the multivariable flow control system. The parameters $\alpha$ and $\beta$ give an affect to characteristics of controller in separated type but $\gamma$ does not give an affect to the characteristics of 2-DOF PID. The more $\beta$ increases, the more overshoot decreases and especially, in case of PI type represent clearly. The $\alpha$, $\beta$ separated type has a very small overshoot and its magnitudes in 2-DOF PID onctroller increases in order of $\alpha$, $\beta$ combined type, PI type, feedforward type, conventional type. The response characteristics of simulation are similar to that of experiments but the experimental characteristics in the multivariable flow control system has the delayed response. The time delay of response in experiments depends on 2-DOF parameter $\alpha$, $\beta$, $\gamma$ and the overshoot increase as the $\alpha$, $\beta$, $\gamma$ increase. So, we can have a satisfactory response by tuning D gain.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.8 no.3
• /
• pp.9-15
• /
• 1998

This paper presents an adaptive learning control method for unknown nonlinear systems by combining neuro control and iterative learning control techniques. In the present control system, an iterative learning controller (ILC) is used for a process of short term memory involved in a temporary adaptive and learning manipulation and a short term storage of a specific temporary action. The learning gain of the iterative learning law is estimated by using a neural network for an unknown system except relative degrees. The control informations obtained by ILC are transferred to a long term memory-based feedforward neuro controller (FNC) and accumulated in it in addition to the previously stored infonnations. This scheme is applied to a two link robot manipulator through simulations.

• Journal of Information Processing Systems
• /
• v.15 no.4
• /
• pp.904-919
• /
• 2019

Many security systems rely solely on solutions based on Artificial Intelligence, which are weak in nature. These security solutions can be easily manipulated by malicious users who can gain unlawful access. Some security systems suggest using fingerprint-based solutions, but they can be easily deceived by copying fingerprints with clay. Image-based security is undoubtedly easy to manipulate, but it is also a solution that does not require any special training on the part of the user. In this paper, we propose a multi-factor security framework that operates in a three-step process to authenticate the user. The motivation of the research lies in utilizing commonly available and inexpensive devices such as onsite CCTV cameras and smartphone camera and providing fully secure user authentication. We have used technologies such as Argon2 for hashing image features and physically unclonable identification for secure device-server communication. We also discuss the methodological workflow of the proposed multi-factor authentication framework. In addition, we present the service scenario of the proposed model. Finally, we analyze qualitatively the proposed model and compare it with state-of-the-art methods to evaluate the usability of the model in real-world applications.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.148-151
• /
• 2010

분류 기법은 데이터 마이닝 기술 중 가장 잘 알려진 방법으로서, Decision tree, SVM(Support Vector Machine), ANN(Artificial Neural Network) 등 기법을 포함한다. 분류 기법은 이미 알려진 상호 배반적인 몇 개 그룹에 속하는 다변량 관측치로부터 각각의 그룹이 어떤 특징을 가지고 있는지 분류 모델을 만들고, 소속 그룹이 알려지지 않은 새로운 관측치가 어떤 그룹에 분류될 것인가를 결정하는 분석 방법이다. 분류기법을 수행할 때에 기본적으로 특징 공간이 잘 표현되어 있다고 가정한다. 그러나 실제 응용에서는 단일 특징으로 구성된 특징공간이 분명하지 않기 때문에 분류를 잘 수행하지 못하는 문제점이 있다. 본 논문에서는 이 문제에 대한 해결방안으로써 많은 정보를 포함하면서 빈발패턴에 대한 정보의 순실이 없는 닫힌 빈발패턴 기반 분류에 대한 연구를 진행하였다. 본 실험에서는 ${\chi}^2$(Chi-square)과 정보이득(Information Gain) 속성 선택 척도를 사용하여 의미있는 특징 선택을 수행하였다. 그 결과, 이 연구에서 제시한 척도를 사용하여 특징 선택을 수행한 경우, C4.5, SVM 과 같은 분류기법보다 더 향상된 분류 성능을 보였다.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.

• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.105-122
• /
• 2019

Dimensionality reduction is one of the methods to handle big data in text mining. For dimensionality reduction, we should consider the density of data, which has a significant influence on the performance of sentence classification. It requires lots of computations for data of higher dimensions. Eventually, it can cause lots of computational cost and overfitting in the model. Thus, the dimension reduction process is necessary to improve the performance of the model. Diverse methods have been proposed from only lessening the noise of data like misspelling or informal text to including semantic and syntactic information. On top of it, the expression and selection of the text features have impacts on the performance of the classifier for sentence classification, which is one of the fields of Natural Language Processing. The common goal of dimension reduction is to find latent space that is representative of raw data from observation space. Existing methods utilize various algorithms for dimensionality reduction, such as feature extraction and feature selection. In addition to these algorithms, word embeddings, learning low-dimensional vector space representations of words, that can capture semantic and syntactic information from data are also utilized. For improving performance, recent studies have suggested methods that the word dictionary is modified according to the positive and negative score of pre-defined words. The basic idea of this study is that similar words have similar vector representations. Once the feature selection algorithm selects the words that are not important, we thought the words that are similar to the selected words also have no impacts on sentence classification. This study proposes two ways to achieve more accurate classification that conduct selective word elimination under specific regulations and construct word embedding based on Word2Vec embedding. To select words having low importance from the text, we use information gain algorithm to measure the importance and cosine similarity to search for similar words. First, we eliminate words that have comparatively low information gain values from the raw text and form word embedding. Second, we select words additionally that are similar to the words that have a low level of information gain values and make word embedding. In the end, these filtered text and word embedding apply to the deep learning models; Convolutional Neural Network and Attention-Based Bidirectional LSTM. This study uses customer reviews on Kindle in Amazon.com, IMDB, and Yelp as datasets, and classify each data using the deep learning models. The reviews got more than five helpful votes, and the ratio of helpful votes was over 70% classified as helpful reviews. Also, Yelp only shows the number of helpful votes. We extracted 100,000 reviews which got more than five helpful votes using a random sampling method among 750,000 reviews. The minimal preprocessing was executed to each dataset, such as removing numbers and special characters from text data. To evaluate the proposed methods, we compared the performances of Word2Vec and GloVe word embeddings, which used all the words. We showed that one of the proposed methods is better than the embeddings with all the words. By removing unimportant words, we can get better performance. However, if we removed too many words, it showed that the performance was lowered. For future research, it is required to consider diverse ways of preprocessing and the in-depth analysis for the co-occurrence of words to measure similarity values among words. Also, we only applied the proposed method with Word2Vec. Other embedding methods such as GloVe, fastText, ELMo can be applied with the proposed methods, and it is possible to identify the possible combinations between word embedding methods and elimination methods.