• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.28 no.2
• /
• pp.385-395
• /
• 2018

With the development of Internet, cyber attack has become a major threat. To detect cyber attacks, intrusion detection system(IDS) has been widely deployed. But IDS has a critical weakness which is that it generates a large number of false alarms. One of the promising techniques that reduce the false alarms in real time is machine learning. However, there are problems that must be solved to use machine learning. So, many machine learning approaches have been applied to this field. But so far, researchers have not focused on features. Despite the features of IDS alerts are important for performance of model, the approach to feature is ignored. In this paper, we propose new feature set which can improve the performance of model and can be extracted from a single alarm. New features are motivated from security analyst's know-how. We trained and tested the proposed model applied new feature set with real IDS alerts. Experimental results indicate the proposed model can achieve better accuracy and false positive rate than SVM model with ordinary features.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.11 no.4
• /
• pp.216-230
• /
• 1999

A Galerkin finite element model for the analysis of harbor oscillation has been developed based on the extended mild-slope equation. Infinite elements are used to accomodate the radiation condition at infinity and joint elements to treat the matching conditions at the harbor entrance which include the energy loss due to flow separation. The numerical tests for rectangular harbors with fully or partially open entrances show that the energy loss at the harbor entrance considerably reduces the the amplification ratios at the innermost parts of the harbors and that the amplification ratios decrease considerably with increasing incident wave heights and jet lengths at the harbor entrance. Application of the model to the Gamcheon harbor show that when the incident wave amplitude is small the amplification ratios rather increase when the entrance energy loss is included than when ignored because of the shift of the resonance periods. Even though the entrance energy loss was insignificant for the measured long-period incident waves, it would be of great importance if the incident waves were large as in the attack of tsunamis. The resonance period of the Helmholtz mode at the Gamcheon Harbor was calculated to be 31 minutes, which agrees well with the measured one between 27 and 33.3 minutes. The measured resonance periods between 9.4 and 12.1 minutes and 5.2 and 6.2 minutes were also calculated by the numerical model as 10.4 minutes and 6.6 or 5.6 minutes, indicating good performance of the model. On the other hand, it was shown that a variety of oscillation modes exists in the Gamcheon Harbor and lateral resonances of considerable amplification ratios also exist at the periods of 3.6 and 1.6 minutes as in the Young-II Bay.

• Journal of the Korean Geotechnical Society
• /
• v.18 no.5
• /
• pp.251-260
• /
• 2002

Geotextile tribes are made of sewn geotextile sheet and hydraulically or mechanically filled with dredged materials. They have been applied in hydraulic and coastal engineering in recent years(shore protection structure, detached breakwater, groins, and jetty). Therefore, it is composed of geotextile and confined fill material. Recently, new preliminary design criteria supported by model and prototype tests, and some stability analysis calculations have been studied. The stability analysis of geotextile tube is composed of geotechnical and hydrodynamic analysis. The stability check points are sliding failure, overturning, bearing capacity failure against the wave attack. In this paper are presented the stability analysis method by empirical equation and 2-D equilibrium analysis for geotextile tube. Also, the hydraulic model tests were performed to verify the theoretical stability analysis with geotextile tube shape, filling ratio, significant wave height, and so on. The results of this study show that the stability of geotextile tube depends on the tube shape, contact area, projection area. The theoretical analysis and hydraulic model test show almost the same results.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.10
• /
• pp.67-76
• /
• 2023

Although deep learning models are making innovative achievements in the field of computer vision, the problem of vulnerability to adversarial examples continues to be raised. Adversarial examples are attack methods that inject fine noise into images to induce misclassification, which can pose a serious threat to the application of deep learning models in the real world. In this paper, we propose a model that detects adversarial examples using differences in predictive values between edge-learned classification models and underlying classification models. The simple process of extracting the edges of the objects and reflecting them in learning can increase the robustness of the classification model, and economical and efficient detection is possible by detecting adversarial examples through differences in predictions between models. In our experiments, the general model showed accuracy of {49.9%, 29.84%, 18.46%, 4.95%, 3.36%} for adversarial examples (eps={0.02, 0.05, 0.1, 0.2, 0.3}), whereas the Canny edge model showed accuracy of {82.58%, 65.96%, 46.71%, 24.94%, 13.41%} and other edge models showed a similar level of accuracy also, indicating that the edge model was more robust against adversarial examples. In addition, adversarial example detection using differences in predictions between models revealed detection rates of {85.47%, 84.64%, 91.44%, 95.47%, and 87.61%} for each epsilon-specific adversarial example. It is expected that this study will contribute to improving the reliability of deep learning models in related research and application industries such as medical, autonomous driving, security, and national defense.

• Journal of Intelligence and Information Systems
• /
• v.13 no.3
• /
• pp.119-140
• /
• 2007

Recommender systems enable a user to decide which information is interesting and valuable in our world of information overload. As the recent studies of distributed computing environment have been progressing actively, recommender systems, most of which were centralized, have changed toward a peer-to-peer approach. Collaborative Filtering (CF), one of the most successful technologies in recommender systems, presents several limitations, namely sparsity, scalability, cold start, and the shilling problem, in spite of its popularity. The move from centralized systems to distributed approaches can partially improve the issues; distrust of recommendation and abuses of personal information. However, distributed systems can be vulnerable to attackers, who may inject biased profiles to force systems to adapt their objectives. In this paper, we consider both effective CF in P2P environment in order to improve overall performance of system and efficient solution of the problems related to abuses of personal data and attacks of malicious users. To deal with these issues, we propose a multi-agent framework for a distributed CF focusing on the trust relationships between individuals, i.e. web of trust. We employ an agent-based approach to improve the efficiency of distributed computing and propagate trust information among users with effect. The experimental evaluation shows that the proposed method brings significant improvement in terms of the distributed computing of similarity model building and the robustness of system against malicious attacks. Finally, we are planning to study trust propagation mechanisms by taking trust decay problem into consideration.

• Korean journal of communication and information
• /
• v.45
• /
• pp.296-340
• /
• 2009

The power of media has always been problematic in the countries of full press freedom. Originally, the media used to be an effective vehicle for communication within human beings. However, it exerts an overwhelming power toward human society. Through applying the well-known four dog models in terms of media function, this study attempts to examine how the press media in South Korea transformed themselves into another powerful independent organization or institution after regime shift in 1987. The whole editorials of four sampled newspapers were analyzed through frame analysis model. The ChosunIlbo, known as a conservative and pro-government paper, shows to take the role of supporting chaebol policies under Roh TaeWoo Administration. However, it criticizing sharply against the chaebol policies of Roh MooHyun Administration. The JoongangIlbo, known as a pro-chaebol paper, appears anti-government position through the entire four administrations in terms of chaebol policies. Particularly, it reveals hostile editorial coverage during the Roh MooHyun Administration. However, KyunghyangShinmun, currently known as a liberal paper, viewed somewhat complicated positions (see text in more detail) because of its ownership turbulence during the past twenty years. On the other hand, Hangyoreh, regarded as a progressive paper, keeps in supportive attitude consistently against the four sampled administrations as far as regulating each government policies for chaebols.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.6
• /
• pp.629-635
• /
• 2019

In Deep Learning models derivative is implemented by error back-propagation which enables the model to learn the error and update parameters. It can find the global (or local) optimal points of parameters even in the complex models taking advantage of a huge improvement in computing power. However, deliberately generated data points can 'fool' models and degrade the performance such as prediction accuracy. Not only these adversarial examples reduce the performance but also these examples are not easily detectable with human's eyes. In this work, we propose the method to detect adversarial datasets with random noise addition. We exploit the fact that when random noise is added, prediction accuracy of non-adversarial dataset remains almost unchanged, but that of adversarial dataset changes. We set attack methods (FGSM, Saliency Map) and noise level (0-19 with max pixel value 255) as independent variables and difference of prediction accuracy when noise was added as dependent variable in a simulation experiment. We have succeeded in extracting the threshold that separates non-adversarial and adversarial dataset. We detected the adversarial dataset using this threshold.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.44 no.7
• /
• pp.552-561
• /
• 2016

In this experimental study, the dynamic stability derivatives of a tailless lambda-shape UAV are estimated from time history data of aerodynamic moments measured from the internal balance while the test model is forced to oscillate at given frequencies and amplitudes. A 3-axis forced oscillation apparatus is designed to induce decoupled roll, yaw, pitch oscillations respectively. The results show that the roll damping derivatives remain stable at the entire range of angle of attack tested, whereas the pitch damping derivatives become unstable beyond $15^{\circ}$ angle of attack. The amplitude and frequency have little impact on roll damping derivatives while the smaller amplitude and frequency of oscillation improves the pitch stability. The yaw damping derivative values are fairly small as expected for a tailless configuration. The results indicate that the proposed methodology and test apparatus area valid for estimating the dynamic stability derivatives of a tailless UAV.

• Korean Journal of Clinical Laboratory Science
• /
• v.50 no.2
• /
• pp.126-135
• /
• 2018

This study examined whether the difference in the middle cerebral artery (MCA) velocities can predict the prognosis of stroke and whether the prognostic impact differs among stroke subtypes. Transient ischemic attack (TIA) or acute ischemic stroke patients, who underwent a routine evaluation and transcranial Doppler (TCD), were included in this study. The MCA asymmetry index was calculated using the relative percentage difference in the mean flow velocity (MFV) between the left and right MCA: (|RMCA MFV-LMCA MFV|/mean MCA MFV)${\times}100$. The stroke subtypes were determined using the TOAST classification. Poor functional outcomes were defined as a mRS score ${\geq}3$ at 3 months after the onset of stroke. A total of 988 patients were included, of whom 157 (15.9%) had a poor functional outcome. Multivariable analysis showed that only the MCA asymmetry index was independently associated with a poor functional outcome. ROC curve analysis showed that adding the MCA asymmetry index to the prediction model improved the discrimination of a poor functional outcome from acute ischemic stroke (from 88.6% [95% CI, 85.2~91.9] to 89.2% [95% CI, 85.9~92.5]). The MCA asymmetry index has an independent prognostic value for predicting a poor short-term functional outcome after an acute cerebral infarction. Therefore, TCD may be useful for predicting a poor functional outcome in patients with acute ischemic stroke.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.28 no.4
• /
• pp.847-857
• /
• 2018

As the number of malicious code increases steeply, cyber attack victims targeting corporations, public institutions, financial institutions, hospitals are also increasing. Accordingly, academia and security industry are conducting various researches on malicious code detection. In recent years, there have been a lot of researches using machine learning techniques including deep learning. In the case of research using Convolutional Neural Network, ResNet, etc. for classification of malicious code, it can be confirmed that the performance improvement is higher than the existing classification method. However, one of the characteristics of the target attack is that it is custom malicious code that makes it operate only for a specific company, so it is not a form spreading widely to a large number of users. Since there are not many malicious codes of this kind, it is difficult to apply the previously studied machine learning or deep learning techniques. In this paper, we propose a method to classify malicious codes when the amount of samples is insufficient such as targeting type malicious code. As a result of the study, we confirmed that the accuracy of 97% can be achieved even with a small amount of data by applying the Memory Augmented Neural Networks model.