• 인터넷정보학회논문지
• /
• 제22권2호
• /
• pp.1-9
• /
• 2021

전통적으로 대부분의 악성코드는 도메인 전문가에 의해 추출된 특징 정보를 활용하여 분석되었다. 하지만 이러한 특징 기반의 분석방식은 분석가의 역량에 의존적이며 기존의 악성코드를 변형한 변종 악성코드를 탐지하는 데 한계를 가지고 있다. 본 연구에서는 도메인 전문가의 개입 없이도 변종 악성코드의 패밀리를 분류할 수 있는 ResNet-Variational AutoEncder 기반 변종 악성코드 분류 방법을 제안한다. Variational AutoEncoder 네트워크는 입력값으로 제공되는 훈련 데이터의 학습 과정에서 데이터의 특징을 잘 이해하며 정규 분포 내에서 새로운 데이터를 생성하는 특징을 가지고 있다. 본 연구에서는 Variational AutoEncoder의 학습 과정에서 잠재 변수를 추출을 통해 악성코드의 중요 특징을 추출할 수 있었다. 또한 훈련 데이터의 특징을 더욱 잘 학습하고 학습의 효율성을 높이기 위해 전이 학습을 수행했다. ImageNet Dataset으로 사전학습된 ResNet-152 모델의 학습 파라미터를 Encoder Network의 학습 파라미터로 전이했다. 전이학습을 수행한 ResNet-Variational AutoEncoder의 경우 기존 Variational AutoEncoder에 비해 높은 성능을 보였으며 학습의 효율성을 제공하였다. 한편 변종 악성코드 분류를 위한 방법으로는 앙상블 모델인 Stacking Classifier가 사용되었다. ResNet-VAE 모델의 Encoder Network로 추출한 변종 악성코드 특징 데이터를 바탕으로 Stacking Classifier를 학습한 결과 98.66%의 Accuracy와 98.68의 F1-Score를 얻을 수 있었다.

• 대한수학회보
• /
• 제60권6호
• /
• pp.1697-1704
• /
• 2023

We study a multidimensional nonlinear variational sine-Gordon equation, which can be used to describe long waves on a dipole chain in the continuum limit. By using the method of characteristics, we show that a solution of a nonlinear variational sine-Gordon equation with certain initial data in a multidimensional space has a singularity in finite time.

• 한국멀티미디어학회논문지
• /
• 제22권11호
• /
• pp.1251-1258
• /
• 2019

Intelligent tutoring system enables users to effectively learn by utilizing various artificial intelligence techniques. For instance, it can recommend a proper curriculum or learning method to individual users based on their learning history. To do this effectively, user's characteristics need to be analyzed and classified based on various aspects such as interest, learning ability, and personality. Even though data labeled by the characteristics are required for more accurate classification, it is not easy to acquire enough amount of labeled data due to the labeling cost. On the other hand, unlabeled data should not need labeling process to make a large number of unlabeled data be collected and utilized. In this paper, we propose a semi-supervised learning method based on feedback variational auto-encoder(FVAE), which uses both labeled data and unlabeled data. FVAE is a variation of variational auto-encoder(VAE), where a multi-layer perceptron is added for giving feedback. Using unlabeled data, we train FVAE and fetch the encoder of FVAE. And then, we extract features from labeled data by using the encoder and train classifiers with the extracted features. In the experiments, we proved that FVAE-based semi-supervised learning was superior to VAE-based method in terms with accuracy and F1 score.

• 한국지구물리탐사학회:학술대회논문집
• /
• 한국지구물리탐사학회 2004년도 대한지구물리학회.한국지구물리탐사학회 공동학술대회 초록집
• /
• pp.103-113
• /
• 2004

We investigate the time-variational characteristics of power spectrum and transfer function of geomagnetic field by robust estimation technique. In the case of power spectrums of geomagnetic field, there are some the periodic fluctuations related with solar activity, Meanwhile, the geomagnetic transfer function shows so considerable weak time-variational fluctuation that the estimations of transfer function seem to be comparatively stable in time-variant view.

• 한국전산구조공학회논문집
• /
• 제27권1호
• /
• pp.37-43
• /
• 2014

동역학의 새로운 변분이론인 확장 해밀턴 이론은 수학물리학을 비롯한 공학에 있어 초기치-경계치 문제해석에 광범위하게 적용될수 있는 기반을 제공하는 것으로 본 논문에서는 이 이론을 기반으로 선형탄성 단자유도계에 적용한 새로운 수치해석법을 제안하였다. 곧, 변분이론의 특성을 감안해, 전체 time-step에 대한 수치해를 한번에 산정하는 해석법을 제안하였고, 주요 예제를 통해 이 해석법의 특성을 살펴보았다. 에너지 보존 시스템의 경우(비감쇠 시스템에 외력이 작용치 않는 경우), time-step에 관계없이 에너지와 모멘텀이 보존되는 symplecticity property를 가지고 있음을 확인할 수 있었고, 감쇠 시스템인 경우, time-step이 점점 작아질수록 정확한 해에 빠르게 수렴하는 것을 확인하였다.

• 한국환경과학회지
• /
• 제9권5호
• /
• pp.429-436
• /
• 2000

In order to study on the variational characteristics of water quality and chlorolphyll a concentration the water samples were collected daily or three times a week during the period from April 1990 to November 1991 at Kukdong port located in the northern Kamak bay of Southern Korea I made an analysis on biological factor as chlorophyll a concentration as well as physico-chemical factors such as water temperature salinity sigma-t dissolved oxygen, nutrients (ammonia, nitrite, nitrate, phosphate, and silicate) N/P ratio and chemical oxygen demand. In Northern Kamak bay seasonal variations in physical factors such as water temperature salinity and sigma-t were very marked. On the other hand chemical factors such as nutrients concentration and COD were not so. Chemical factors, in particular silicate were influenced by input of freshwater. And the roles of silicate on the seasonal succession of phytoplankton species composition was very low. Phytoplankton biomass as measured by chlorophyll a concentration was very high all the year round and it was controlled by the combination of several factors especially of N/P ratio determined by dissolved inorganic nitrogen.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2008년도 춘계학술대회 논문집
• /
• pp.831-834
• /
• 2008

The spectral element model is known to provide very accurate structural dynamic characteristics, while reducing the number of degree-of-freedom to resolve the computational and cost problems. Thus, the spectral element model with variational method for an axially moving Bernoulli-Euler beam subjected to axial tension is developed in the present paper. The high accuracy of the spectral element model is the verified by comparing its solutions with the conventional finite element solutions and exact analytical solutions. The effects of the moving speed and axial tension the vibration characteristics, wave characteristics, and the static and dynamic stabilities of a moving beam are investigated.

• 정보과학회 논문지
• /
• 제45권2호
• /
• pp.157-164
• /
• 2018

딥러닝의 급속한 발전은 패턴인식 분야의 성능을 혁신했으며 몇몇 문제에서는 인간 수준을 넘어서는 결과들을 보여주고 있다. 데이타를 분류하는 패턴인식과 달리 본 논문에서는 주어진 몇개의 한국어 문장으로부터 비슷한 문장들을 생성하는 문제를 다룬다. 이를위해 생성모델 중의 하나인 Variational Auto-Encoder 기반의 모델을 한국어 생성에 맞게 개선하고 적용하는 방법들을 논의한다. 첫째, 교착어인 한국어의 특성상 띄어쓰기를 기준으로 단어 생성시 단어의 개수가 너무 많아 이를 줄이기 위해 조사 및 어미들을 분리할 필요가 있다. 둘째, 한국어는 어순이 비교적 자유롭고 주어 목적어 등이 생략되는 경우가 많아 기존의 단방향 인코더를 양방향으로 확장한다. 마지막으로, 주어진 문장들을 기반으로 비슷하지만 새로운 문장들을 생성하기 위해 기존 문장들의 인코딩된 벡터표현들로부터 새로운 벡터를 찾아내고, 이 벡터를 디코딩하여 문장을 생성한다. 실험 결과를 통해 제안한 방법의 성능을 확인한다.

• 한국환경과학회지
• /
• 제9권1호
• /
• pp.1-11
• /
• 2000

A study on the variational characteristics of water quality and phytoplankton biomass by principal component analysis(PCA) was carried out in Kogum-sudo from February to October in 1993. We analyzed PCA on biological factors such as chlorophyll a and phytoplankton cell numbers for centric and pennate diatoms, phytoflagellates, and total phytoplankton as well as physico-chemical factors as water temperature, salinity, transparency, dissolved oxygen(DO), saturation of DO, apparent oxygen utilization (AOU), chemical oxygen demand(COD), nutrient (ammonia, nitrite, nitrate, phosphate and silicate), N/P ratio and suspended solid(SS). The source of nutrients supply depended on the mineralization of organic matters and inputs of seawater from outside rather than runoff of freshwater. The phytoplankton biomass was changed within short interval period by nutrients change. And it was controlled by the combination of several environmental factors, especially of light intensity, ammonia and phosphate. The marine environmental characteristics were determined by the mineralization of organic matters in winter, by runoff of freshwater including high nutrients concentration in spring, by ammonia uptake and high phytoplankton productivity in summer, and phosphate supplied input seawater from outside of Kogeum-sudo in autumn. And Kogum-sudo was separated with 2 regions by score distributions of PCA. That is to say, one region was middle parts of straits which was characterized by the mixing seawater and the accumulated organic matters, other one region was Pungnam Bay and the water around Kogum Island which was done by high phytoplankyon biomass and productivity year-round.

• 시스템엔지니어링학술지
• /
• 제19권2호
• /
• pp.145-156
• /
• 2023

Trajectory of ballistic missile is defined by inherent flight dynamics, which decided range and maneuvering characteristics. It is crucial to predict range and maneuvering characteristics of ballistic missile in KAMD (Korea Air and Missile Defense) to minimize damage due to ballistic missile attacks, Nowadays, needs for applying AI(Artificial Intelligence) technologies are increasing due to rapid developments of DNN(Deep Neural Networks) technologies. To apply these DNN technologies amount of data are required for superviesed learning, but trajectory data of ballistic missiles is limited because of security issues. Trajectory data could be considered as multivariate time series including many variables. And augmentation in time series data is a developing area of research. In this paper, we tried to augment trajectory data of ballistic missiles using recently developed methods. We used TimeVAE(Time Variational AutoEncoder) method and TimeGAN(Time Generative Adversarial Networks) to synthesize missile trajectory data. We also compare the results of two methods and analyse for future works.