• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.177-189
• /
• 2023

Malware detection is an increasingly important operational focus in cyber security, particularly given the fast pace of such threats (e.g., new malware variants introduced every day). There has been great interest in exploring the use of machine learning techniques in automating and enhancing the effectiveness of malware detection and analysis. In this paper, we present a deep recurrent neural network solution as a stacked Long Short-Term Memory (LSTM) with a pre-training as a regularization method to avoid random network initialization. In our proposal, we use global and short dependencies of the inputs. With pre-training, we avoid random initialization and are able to improve the accuracy and robustness of malware threat hunting. The proposed method speeds up the convergence (in comparison to stacked LSTM) by reducing the length of malware OpCode or bytecode sequences. Hence, the complexity of our final method is reduced. This leads to better accuracy, higher Mattews Correlation Coefficients (MCC), and Area Under the Curve (AUC) in comparison to a standard LSTM with similar detection time. Our proposed method can be applied in real-time malware threat hunting, particularly for safety critical systems such as eHealth or Internet of Military of Things where poor convergence of the model could lead to catastrophic consequences. We evaluate the effectiveness of our proposed method on Windows, Ransomware, Internet of Things (IoT), and Android malware datasets using both static and dynamic analysis. For the IoT malware detection, we also present a comparative summary of the performance on an IoT-specific dataset of our proposed method and the standard stacked LSTM method. More specifically, of our proposed method achieves an accuracy of 99.1% in detecting IoT malware samples, with AUC of 0.985, and MCC of 0.95; thus, outperforming standard LSTM based methods in these key metrics.

• Science of Emotion and Sensibility
• /
• v.21 no.3
• /
• pp.151-164
• /
• 2018

The appropriate interpretation of social cues is a crucial ability for everyday life. While processing socially relevant information, beyond the low-level physical features of the stimuli to emotional information is known to influence human cognition in various stages, from early perception to later high-level cognition, such as working memory (WM). However, it remains unclear how the influence of each type of emotional information on cognitive processes changes in response to what has occurred in the processing stage. Past studies have largely adopted face stimuli to address this type of research question, but we used a unique class of socially relevant motion stimuli, called biological motion (BM), which depicts various human actions and emotions with moving dots to exhibit the effects of anger, happiness, and neutral emotion on task performance in perceptual and working memory. In this study, participants determined whether two BM stimuli, sequentially presented with a delay between them (WM task) or one immediately after the other (perceptual task), were identical. The perceptual task showed that discrimination accuracies for emotional stimuli (i.e., angry and happy) were lower than those for neutral stimuli, implying that emotional information has a negative impact on early perceptual processes. Alternatively, the results of the WM task showed that the accuracy drop as the interstimulus interval increased was actually lower in emotional BM conditions than in the neutral condition, which suggests that emotional information benefited maintenance. Moreover, anger and happiness had distinct impacts on the performance of perception and WM. Our findings have significance as we provide evidence for the interaction of type of emotion and information-processing stage.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_3
• /
• pp.1077-1093
• /
• 2020

Sea Surface Temperature (SST) is an important environmental indicator that affects climate coupling systems around the world. In particular, coastal regions suffer from abnormal SST resulting in huge socio-economic damage. This study used Long Short Term Memory (LSTM) and Convolutional Long Short Term Memory (ConvLSTM) to predict SST up to 7 days in the south sea region in South Korea. The results showed that the ConvLSTM model outperformed the LSTM model, resulting in a root mean square error (RMSE) of 0.33℃ and a mean difference of -0.0098℃. Seasonal comparison also showed the superiority of ConvLSTM to LSTM for all seasons. However, in summer, the prediction accuracy for both models with all lead times dramatically decreased, resulting in RMSEs of 0.48℃ and 0.27℃ for LSTM and ConvLSTM, respectively. This study also examined the prediction of abnormally high SST based on three ocean heatwave categories (i.e., warning, caution, and attention) with the lead time from one to seven days for an ocean heatwave case in summer 2017. ConvLSTM was able to successfully predict ocean heatwave five days in advance.

• 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.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.4 no.4
• /
• pp.269-274
• /
• 2004

Current delivery system, in which the analog current produced by a unique DAC circuit is stored into a current-memory circuit and delivered in a time-divided sequence, shows variation of output current as low as 4% in a current source data-driver IC for AM-OLED driven by a current-programmed method without any fuse repairing after fabrication. This driver IC has 54 outputs and can sink constant current as low as 3 ${\mu}A$ with 6-bit analog levels. Such a low current level without variation can hardly be obtained by an ordinary MOS transistor because the current level is in the sub-threshold region and changes exponentially with threshold voltage variation. Thus we adopted a current mirror circuit composed of bipolar transistors to supply well-controlled current within a nano-ampere range.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.40-43
• /
• 1997

This paper presents a 3D object recognition method for generation of 3D environmental map or obstacle recognition of mobile robots. An active light source projects a stripe pattern of light onto the object surface, while the camera observes the projected pattern from its offset point. The system consists of a laser unit and a camera on a pan/tilt device. The line segment in 2D camera image implies an object surface plane. The scaling, filtering, edge extraction, object extraction and line thinning are used for the enhancement of the light stripe image. We can get faithful depth informations of the object surface from the line segment interpretation. The performance of the proposed method has demonstrated in detail through the experiments for varies type objects. Experimental results show that the method has a good position accuracy, effectively eliminates optical noises in the image, greatly reduces memory requirement, and also greatly cut down the image processing time for the 3D object recognition compared to the conventional object recognition.

• Computational Structural Engineering
• /
• v.6 no.2
• /
• pp.103-114
• /
• 1993

An advanced time-domain finite element formulation is presented for the displacement and stress analysis of isotropic, linear viscoelastic problems of a hereditary-type constitutive law. The semidiscrete finite element method with linear time-stepping scheme and an elastic-viscoelastic correspondence principle are used in the theoretical development. An efficient treatment of hereditary integral is introduced to improve the numerical accuracy, to reduce the computation time, and to avoid the use of large memory storage. Two-dimensional numerical examples of plane strain and plane stress are solved under the assumption on the material property of being elastic in dilatation and like three-element Voigt model in distorsion, and compared with the analytical solutions and the past numerical results to show the versatility and efficiency of the proposed method.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2008.04a
• /
• pp.362-367
• /
• 2008

Recently, the use of biaxial hollow slab is increased to reduce noise and vibration of the floor slab. Therefore, an efficient method for the vibration analysis of biaxial hollow slab is required to describe dynamic behavior of biaxial hollow slab. A finite element analysis is one of the method to analyze the biaxial hollow slab. It is necessary to use a refined finite element model for an accurate analysis of a floor slab with an effects of the hollow shape. But it would take a significant amount of computational time and memory if the entire building structure were subdivided into a finer mesh. Thus the proposed method uses equivalent plate model in this study. Dynamic analyses of an example structure subjected to walking loads were performed to verify the efficiency and accuracy of the proposed method.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.3E
• /
• pp.35-42
• /
• 1998

Transient acoustic and elastic wave propagation in inhomogeneous media are studied by using the Fourier method. It is known that the fourier method has advantages in memory requirements and computing speed over conventional methods such as FDM and FEM, because the Fourier method needs less grid points for achieving the same accuracy. To verify the proposed numerical scheme, several examples having analytic solutions are considered, where two different semi-infinite media are in contact along a plane boundary. The comparisons of numerical results by the Fourier method and analytic solutions show good agreements. In addition, the fourier method is applied to a layered half-plane, in which an elastic semi-infinite medium is covered by an elastic layer of finite thickness. It is showed how to derive the analytic solutions by using the Cagniard-de Hoop method. The numerical solutions are in excellent agreements with analytic results.

• East Asian mathematical journal
• /
• v.33 no.5
• /
• pp.495-509
• /
• 2017

With the accuracy of the nonlinearity guaranteed, plenty of time and large memory space are needed when we solve the finite element numerical solution of nonlinear partial differential equations. In this paper, we use the Group Element Method (GEM) to deal with the non-linearity of the BBM-Burgers Equation with Conservation form and perform a numerical analysis for two particular initial-boundary value (the Dirichlet boundary conditions and Neumann-Dirichlet boundary conditions) problems with the Finite Element Method (FEM). Some numerical experiments are performed to analyze the error between the exact solution and the FEM solution in MATLAB.