• Journal of the Korean Magnetics Society
• /
• v.13 no.5
• /
• pp.204-210
• /
• 2003

Electron magnetic resonance (EMR) of paramagnetic Cr$^{3+}$, Mn$^{2+}$, and Fe$^{3+}$ impurity ions in ferroelectric LiNbO$_3$ and LiTaO$_3$ single crystals has been studied. The actual sites location of paramagnetic impurity ions in the crystals was suggested from the experimental results and zero field splitting parameters calculated by superposition model. It turns out that Cr$^{3+}$ ions in LiNbO$_3$ crystal have two resonance centers and enter both the Li$^{+}$ and Nb$^{5+}$ ions. Mn$^{2+}$ and Fe$^{3+}$ impurity ions in LiNbO$_3$ substitute for Nb$^{5+}$ ions. However, both Cr$^{3+}$ and Fe$^{3+}$ ions in LiTaO$_3$ crystal reside at Li$^{+}$ ions.$ +/ ions.+/ ions.

• The KIPS Transactions:PartB
• /
• v.19B no.3
• /
• pp.195-200
• /
• 2012

In Korean, spelling change in various forms must be recovered into base forms in morphological analysis as well as part-of-speech (POS) tagging is difficult without morphological analysis because Korean is agglutinative. This is one of notorious problems in Korean morphological analysis and has been solved by morpheme recovery rules, which generate morphological ambiguity resolved by POS tagging. In this paper, we propose a morpheme recovery scheme based on machine learning methods like Na$\ddot{i}$ve Bayes models. Input features of the models are the surrounding context of the syllable which the spelling change is occurred and categories of the models are the recovered syllables. The POS tagging system with the proposed model has demonstrated the $F_1$-score of 97.5% for the ETRI tree-tagged corpus. Thus it can be decided that the proposed model is very useful to handle morpheme recovery in Korean.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.2
• /
• pp.99-110
• /
• 2023

This study proposed a classification of malicious network traffic using the cyber threat framework(Mitre ATT&CK) and machine learning to solve the real-time traffic detection problems faced by current security monitoring systems. We applied a network traffic dataset called UNSW-NB15 to the Mitre ATT&CK framework to transform the label and generate the final dataset through rare class processing. After learning several boosting-based ensemble models using the generated final dataset, we demonstrated how these ensemble models classify network traffic using various performance metrics. Based on the F-1 score, we showed that XGBoost with no rare class processing is the best in the multi-class traffic environment. We recognized that machine learning ensemble models through Mitre ATT&CK label conversion and oversampling processing have differences over existing studies, but have limitations due to (1) the inability to match perfectly when converting between existing datasets and Mitre ATT&CK labels and (2) the presence of excessive sparse classes. Nevertheless, Catboost with B-SMOTE achieved the classification accuracy of 0.9526, which is expected to be able to automatically detect normal/abnormal network traffic.

• Proceedings of the Korean Radioactive Waste Society Conference
• /
• 2015.10a
• /
• pp.145-146
• /
• 2015

• Journal of Oriental Neuropsychiatry
• /
• v.10 no.1
• /
• pp.1-15
• /
• 1999

The effects of Sahyangsohapwon on the enhancement of learning and memory of AD model rats were studied with Morris water maze and radial arm maze. Sample group was electrolytically lesioned on nbM, and then daily treated with the medicine for two months. Control group with nbM lesion, and sham group with the sham operation were treated the vehicle for same duration. The following results were observed. In the learning trials of Morris water maze, all three groups were improved in learning capacity as trials were repeated, but the sham group showed more prominent improvement in learning compared with the control group(p<0.01). 2. In memory retention test of Morris water maze, the sham group marked more significant improvement statistically in memory retention compared with the control group(p<0.05). 3. In the learning of radial arm maze, the sham group shows better learning capacity significantly compared with the control group(p<0.05). With the experimental results above, Sahyangsohapwon can be supposed to have the improving effects on the learning and memory of AD rats induced by electronical injury of nbM.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.83-86
• /
• 2020

본 논문에서는 네트워크 이상치 탐지를 위하여 정상 데이터만을 활용한 메모리 기반 정상성 학습 모델을 제안한다. 오토인코더를 기반으로 정상 데이터의 특징을 표현하는 프로토타입을 생성할 수 있도록 신경망을 구성하고, 네트워크 데이터의 특성을 반영하여 쿼리의 수를 한 개로 고정하며, 사용되는 프로토타입의 수를 지정한 값으로 고정하여 모든 프로토타입에 정상 데이터의 특징을 반영할 수 있는 학습 방법을 제안한다. 해당 모델을 네트워크 이상치 탐지 데이터 세트인 Kyoto Honeypot, UNSW-NB15, CICIDS-2018에 적용하여 본 결과 Kyoto Honeypot에서는 0.821, UNSW-NB15에서는 0.854, CICIDS-2018에서는 0.981의 AUROC를 달성했다.

• Proceedings of the Korean Powder Metallurgy Institute Conference
• /
• 2001.04a
• /
• pp.14-14
• /
• 2001

WC-Co 계의 입성장 억제는 현재 초경합금 분야에서 공학적으로 가장 중요한 이슈들 중의 하나이다 VC를 비롯한 입방정 탄화물이나 $Cr_3C_2$ 등의 여러 가지 탄화 물이 혼합되어 입성장 억제에 이용되는데 입성장 억제의 효과는 대략적으로 용해되는 탄화물의 양에 의존하고 있는 것으로 추정된다. 보다 효율적으로 입성장 억제를 실현하려면 입성장 기구를 명확히 할 필요가 있다. 최 등[1]은 VC가 WC 입자 표면 에서의 edge energy를 증가시켜서 2차원 핵생성의 에너지 장벽을 올리게 하고 이에 따라 입성장이 억제된다는 모델을 제안하였다. 이러한 모텔을 입증하기 위해서는 이론적으로만 예측이 가능한 edge energy보다 좀 더 명확한 물리적인 변수가 제시되어야 할 것이다. 여기서는 또 다른 계인 NbC- TiC-Co 계에서 NbC와 TiC의 버에 따른 입성장 거동과 입자의 형상간의 관계로부터 업성장과 edge energy 그리 고 edge energy와 입자 형상간의 관계를 알아보고 이로부터 좀 더 구체적인 의미 에서의 입성장 모텔과 입성장 억제기구를 제시하고자 한다.

• Smart Media Journal
• /
• v.13 no.6
• /
• pp.24-34
• /
• 2024

If the future can be predicted from network traffic data, which is a time series, it can achieve effects such as efficient resource allocation, prevention of malicious attacks, and energy saving. Many models based on statistical and deep learning techniques have been proposed, and most of these studies have focused on improving model structures and learning algorithms. Another approach to improving the prediction performance of the model is to obtain a good-quality data. With the aim of obtaining a good-quality data, this paper applies a dense sampling technique that augments time series data to the application of network traffic prediction and analyzes the performance improvement. As a dataset, UNSW-NB15, which is widely used for network traffic analysis, is used. Performance is analyzed using RMSE, MAE, and MAPE. To increase the objectivity of performance measurement, experiment is performed independently 10 times and the performance of existing sparse sampling and dense sampling is compared as a box plot. As a result of comparing the performance by changing the window size and the horizon factor, dense sampling consistently showed a better performance.

• Journal of the Korean Vacuum Society
• /
• v.20 no.4
• /
• pp.300-306
• /
• 2011

We report the surface distribution of metal (Ag, Au) nanoparticles grown on polarity-patterned ferroelectric substrates by photochemical reaction. Single crystal periodically polarity-patterned $LiNbO_3$(PPLN) was used as a ferroelectric substrate. The nanoparticles were grown by ultra-violet (UV) light exposure of the PPLN in the aqueous solutions including metas. The surface distribution of the grown nanoparticles were measured by atomic force microscopy and identification of the orientation of the polarity of the ferroelectric surface was performed by piezoelectric force microscopy. The Ag- and Au-nanoparticles grown on +z polarity regions are larger and denser than that on -z polarity regions. In particlur, the largest and denser Ag-nanoparticles were grwon on the polarity boundary regions of the PPLN while Au-nanoparticles were not specifically grown on the boundary regions. Thus, we found that the size and position of metal nanoparticles grown on ferroelectric surfaces can be controlled by UV-exposure time and polarity pattern structures. Also, we discuss the difference of the surface distribution of the metal nano-particles depending on the polarity of the ferroelectric surfaces in terms of surface band structures, reduced work fucntion, and inhomogeneous electric field distribution.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2011.02a
• /
• pp.445-445
• /
• 2011

본 연구에서는 편극 패턴된 강유전체 단결정 $LiNbO_3$ 기판에 광화학적 반응에 의해 금속(Au, Ag, Cu)나노입자를 표면에 선택적으로 성장하였다. 강유전체는 자발편극성의 특성을 지니고 있기 때문에 선택적으로 전압을 가하여 편극성의 역전에 의해 표면의 편극성을 선택적으로 패터닝이 가능하다. 본 연구에서는 주기적으로 양의 편극 영역과 음의 편극 영역이 패턴된 $LiNbO_3$ 기판을 사용하였다. 표면의 편극성은 압전소자반응현미경법(PFM)을 이용하여 확인하였으며, 극성은 R-V curve로 확인하였다. 금속입자는 금속입자를 포함하는 용액에 기판을 넣고 자외선을 조사하여 성장시켰다. 성장된 금속입자의 표면 분포 및 분석은 AFM을 이용하여 측정하였다. Ag 입자를 성장시킨 결과, (-z)편극 영역보다 (+z)편극영역에서 보다 많은 금속 나노입자들이 환원반응을 일으켜 나노입자를 형성하였으며, 경계영역 (inversion domain boundary)에 가장 많은 나노구조체가 형성되었다. Au 입자의 경우, (+z)편극영역이 (-z)편극영역의 표면보다 더 많은 입자가 형성되었지만 Ag입자처럼 편극영역의 경계에서 많이 증착되는 경향성은 보이지 않았다. Cu 입자의 경우 광화학반응을 거의 일으키지 않았으며, 편극영역에 따른 증착 경향성도 보이지 않았다. 이와 같은 결과를 증착된 금속 나노입자의 편극에 따른 표면분포를 강유전체 표면 극성에 따른 표면 밴드구조와, 각 입자가 지닌 환원전위와 전자친화도에 관련된 모델로 설명할 것이다.