• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.97-97
• /
• 2021

Hydrometeorological variables contain highly complex system for temporal revolution and it is quite challenging to illustrate the system with a temporal linear and nonlinear models. In recent years, deep learning algorithms have been developed and a number of studies has focused to model the complex hydrometeorological system with deep learning models. In the current study, we investigated the temporal structure inside deep learning models for the hydrometeorological variables such as streamflow and climate indices. The results present a quite striking such that each hidden unit of the deep learning model presents different dependence structure and when the number of hidden units meet a proper boundary, it reaches the best model performance. This indicates that the deep dependence structure of deep learning models can be used to model selection or investigating whether the constructed model setup present efficient or not.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.88-97
• /
• 2023

Traffic flow prediction is of great significance in urban planning and traffic management. As the complexity of urban traffic increases, existing prediction methods still face challenges, especially for the fusion of spatiotemporal information and the capture of long-term dependencies. This study aims to use the fusion model of graph neural network to solve the spatio-temporal information fusion problem in traffic flow prediction. We propose a new deep learning model Spatio-Temporal Information Fusion using Graph Neural Networks (STFGNN). We use GCN module, TCN module and LSTM module alternately to carry out spatiotemporal information fusion. GCN and multi-core TCN capture the temporal and spatial dependencies of traffic flow respectively, and LSTM connects multiple fusion modules to carry out spatiotemporal information fusion. In the experimental evaluation of real traffic flow data, STFGNN showed better performance than other models.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.571-585
• /
• 2021

Applications for object detection are expanding as it is automated through artificial intelligence-based processing, such as deep learning, on a large volume of images and videos. High dependence on training data and a non-transparent way to find answers are the common characteristics of deep learning. Attacks on training data and training models have emerged, which are closely related to the nature of deep learning. Privacy, integrity, and robustness for the extracted information are important security issues because deep learning enables object recognition in images and videos. This paper summarizes the security issues that need to be addressed for future applications and analyzes the state-of-the-art security studies related to robustness, privacy, and integrity of object detection for images and videos.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.2
• /
• pp.36.1-36.1
• /
• 2018

Star formation activity of galaxies, along with color and morphology, show significant environmental dependence in local universe, where galaxies in dense environment tend to be more quiescent and redder. However, many studies show that such environmental dependence does not continue at higher redshifts beyond z~1. The question of how the environmental dependence of galactic properties have developed over time is crucial to understanding cosmic galactic evolution. By combining data from Canada-France-Hawaii Telescope Legacy Survey(CFHTLS), Infrared Medium-Deep Survey(IMS), and other surveys, the photometric redshifts of galaxies in CFHTLS W2 field were estimated by fitting spectral energy distribution. The distribution of galaxies was mapped in redshift bins of 0.05 interval from 0.6 to 1.4. For each redshift bin, the number density was mapped. The galaxies in high density regions were grouped into clusters using friend-of-friend method. The color of galaxies were analyzed to study the correlation with redshift as well as environmental difference between field galaxies and cluster member galaxies.

• Journal of Information Processing Systems
• /
• v.20 no.4
• /
• pp.458-476
• /
• 2024

To address the issue of class imbalance in network traffic data, which affects the network intrusion detection performance, a combined framework using transformers is proposed. First, Tomek Links, SMOTE, and WGAN are used to preprocess the data to solve the class-imbalance problem. Second, the transformer is used to encode traffic data to extract the correlation between network traffic. Finally, a hybrid deep learning network model combining a bidirectional gated current unit and deep neural network is proposed, which is used to extract long-dependence features. A DNN is used to extract deep level features, and softmax is used to complete classification. Experiments were conducted on the NSLKDD, UNSWNB15, and CICIDS2017 datasets, and the detection accuracy rates of the proposed model were 99.72%, 84.86%, and 99.89% on three datasets, respectively. Compared with other relatively new deep-learning network models, it effectively improved the intrusion detection performance, thereby improving the communication security of network data.

• Transactions of Materials Processing
• /
• v.14 no.7 s.79
• /
• pp.607-612
• /
• 2005

In this paper, warm deep drawing process with local heating and cooling technique was attempted to improve the formability of AZ31 magnesium alloy which is impossibly to form by conventional methods at room temperature by finite element method and experiment. For FE analysis, in first model with considering heat transfer, both die and blankholder were heated to 573K while the punch was kept at room temperature by cooling water. Also distribution of thickness and von Mises stress at room temperature and 498k for warm deep drawing were compared by FEM. Uniaxial tension tests at elevated temperature were done in order to obtain the temperature dependence of material constant under temperature of $293K\~573K$ and cross head velocity of $5\~500mm/min$. The phenomenological model for warm deep drawing process in this work was based on the hardening law and power law strain rate dependency. Deep drawing experiment were conducted at temperatures of room temperature, 373K, 423K, 473K, 498K, 523K, and 573K for the blank and deep drawing tools(holder and die) and at a punch speed of 10mm/min.

• Publications of The Korean Astronomical Society
• /
• v.30 no.2
• /
• pp.413-415
• /
• 2015

How galaxy evolution differs in different environments is one of the intriguing questions in the study of structure formation. While galaxy properties are clearly distinguished in different environments in the local universe, it is still an open issue what causes this environmental dependence of various galaxy properties. To address this question, in this work, we investigate the build-up of passive galaxies over a wide redshift range, from z ~ 2 to z ~ 0.5, focusing on its dependence on galaxy environment. In the UKIDSS/Ultra Deep Survey (UDS) field, we identify high-redshift galaxy cluster candidates within this redshift range. Then, using deep optical and near-infrared data from Subaru and UKIRT available in this field, we analyze and compare the stellar population properties of galaxies in the clusters and in the field. Our results show that the environmental effect on galaxy star-formation properties is a strong function of redshift as well as stellar mass - in the sense that (1) the effect becomes significant at small redshift, and (2) it is stronger for low-mass ($M_{\ast}<10^{10}M_{\odot}$) galaxies. We have also found that galaxy stellar mass plays a more significant role in determining their star-formation property - i.e., whether they are forming stars actively or not - than their environment throughout the redshift range.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.06a
• /
• pp.1757-1760
• /
• 2005

This paper is concerned with spring back after sheet forming of bulk amorphous alloys in the super cooled liquid state. The temperature-dependence and strain-rate dependence of Newtonian/non-Newtonian viscosities as well as the stress overshoot/undershoot behavior of amorphous alloys are reflected in the thermo-mechanical Finite Element simulations. Hemispherical deep drawing operations are simulated for various forming conditions such as punch velocity, die corner radius, friction, blank holder force, clearance and initial forming temperature. Here, spring back by an instantaneous elastic unloading was followed by thermal deformation during cooling and two modes of spring backs are examined in detail. It could be concluded that the superior sheet formability of an amorphous alloy can be obtained by taking the proper forming conditions for loading/unloading.

• Ocean Science Journal
• /
• v.41 no.3
• /
• pp.175-179
• /
• 2006

The Ieodo Ocean Research Station(IORS) is an integrated meteorological and oceanographic observation base which was constructed on the Ieodo underwater rock located at a distance of about 150 km to the south-west of the Mara-do, the southernmost island in Korea. The underwater ambient noise level observed at the IORS was similar to the results of the shallow water surrounding the Korean Peninsula (Choi et al. 2003) and was higher than that of deep ocean (Wenz 1962). The wind dependence of ambient noise was dominant at frequencies of a few kHz. The surface current dependence of ambient noise showed good correlation with the ambient noise in the frequency of 10 kHz. Especially, the shrimp sound was estimated through investigations of waveform and spectrum and its main acoustic energy was about 40 dB larger than ambient noise level at 5 kHz.

• Transactions of Materials Processing
• /
• v.14 no.9 s.81
• /
• pp.751-755
• /
• 2005

This paper is concerned with spring back after sheet forming of bulk amorphous alloys in the super cooled liquid state. The temperature-dependence and strain-rate dependence of Newtonian/non-Newtonian viscosities as well as the stress overshoot/undershoot behavior of amorphous alloys are reflected in the thermo-mechanical Finite Element simulations. Hemispherical deep drawing operations are simulated for various forming conditions such as punch velocity, die comer radius, friction, blank holder force, clearance and initial funning temperature. Here, spring back by an instantaneous elastic unloading was followed by thermal deformation during cooling, and two modes of spring back are examined in detail. It could be concluded that the superior sheet formability of an amorphous alloy can be obtained by taking the proper forming conditions for loading/unloading.