• Journal of Intelligence and Information Systems
• /
• v.23 no.2
• /
• pp.1-17
• /
• 2017

The deep learning framework is software designed to help develop deep learning models. Some of its important functions include "automatic differentiation" and "utilization of GPU". The list of popular deep learning framework includes Caffe (BVLC) and Theano (University of Montreal). And recently, Microsoft's deep learning framework, Microsoft Cognitive Toolkit, was released as open-source license, following Google's Tensorflow a year earlier. The early deep learning frameworks have been developed mainly for research at universities. Beginning with the inception of Tensorflow, however, it seems that companies such as Microsoft and Facebook have started to join the competition of framework development. Given the trend, Google and other companies are expected to continue investing in the deep learning framework to bring forward the initiative in the artificial intelligence business. From this point of view, we think it is a good time to compare some of deep learning frameworks. So we compare three deep learning frameworks which can be used as a Python library. Those are Google's Tensorflow, Microsoft's CNTK, and Theano which is sort of a predecessor of the preceding two. The most common and important function of deep learning frameworks is the ability to perform automatic differentiation. Basically all the mathematical expressions of deep learning models can be represented as computational graphs, which consist of nodes and edges. Partial derivatives on each edge of a computational graph can then be obtained. With the partial derivatives, we can let software compute differentiation of any node with respect to any variable by utilizing chain rule of Calculus. First of all, the convenience of coding is in the order of CNTK, Tensorflow, and Theano. The criterion is simply based on the lengths of the codes and the learning curve and the ease of coding are not the main concern. According to the criteria, Theano was the most difficult to implement with, and CNTK and Tensorflow were somewhat easier. With Tensorflow, we need to define weight variables and biases explicitly. The reason that CNTK and Tensorflow are easier to implement with is that those frameworks provide us with more abstraction than Theano. We, however, need to mention that low-level coding is not always bad. It gives us flexibility of coding. With the low-level coding such as in Theano, we can implement and test any new deep learning models or any new search methods that we can think of. The assessment of the execution speed of each framework is that there is not meaningful difference. According to the experiment, execution speeds of Theano and Tensorflow are very similar, although the experiment was limited to a CNN model. In the case of CNTK, the experimental environment was not maintained as the same. The code written in CNTK has to be run in PC environment without GPU where codes execute as much as 50 times slower than with GPU. But we concluded that the difference of execution speed was within the range of variation caused by the different hardware setup. In this study, we compared three types of deep learning framework: Theano, Tensorflow, and CNTK. According to Wikipedia, there are 12 available deep learning frameworks. And 15 different attributes differentiate each framework. Some of the important attributes would include interface language (Python, C ++, Java, etc.) and the availability of libraries on various deep learning models such as CNN, RNN, DBN, and etc. And if a user implements a large scale deep learning model, it will also be important to support multiple GPU or multiple servers. Also, if you are learning the deep learning model, it would also be important if there are enough examples and references.

• Journal of Life Science
• /
• v.28 no.2
• /
• pp.240-246
• /
• 2018

Efferent and afferent sympathetic nerves are closely related to the development of hypertension and heart failure. Catheter-based renal sympathetic denervation (RDN) is implemented as a strategy to treat resistant hypertension. We investigated whether RDN procedure causes inflammatory damage on myocardium in the early phase of sympathetic denervation. Twenty-five female swine were divided into 3 groups: normal control (Normal, n=5), sham-operated control (Sham, n=5), and RDN groups (RDN, n=15). The RDN group was further subdivided into 3 subgroups according to the time of sacrifice: immediately (RDN-0, n=5), 1 week (RDN-1, n=5), and 2 weeks (RDN-2, n=5) after RDN. There were no significant changes in the clinical parameters between the normal control and sham-operated group using contrast-media. In the myocardium, inflammatory cytokines, $IL-1{\beta}$ and $TNF-{\alpha}$ increased at the first week, and then decreased at the second week after RDN. Anti-inflammatory cytokine, IL-10 increased immediately, and then decreased at the second week after RDN. Caspase-1 activity and apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) expression increased immediately after RDN until the second week. However, nucleotide-binding oligomerization domain, leucine rich repeat and pyrin domain containing protein 3 (NLRP3) expression did not show any significant differences among the groups. The RDN can cause acute myocardial inflammation through activation of caspase-1 and $IL-1{\beta}$. We should pay attention to protecting against early inflammatory myocardial damage after RDN.

• The Korean Journal of Air & Space Law and Policy
• /
• v.30 no.1
• /
• pp.207-244
• /
• 2015

Aviation safety is the stage in which the risk of harm to persons or of property damage is reduced to, and maintained at or below, an acceptable level through a continuing process of hazard identification and risk management. Many accidents and incidents have been taking place since 2014, while there had been relatively safer skies before 2014. International civil aviation community has been exerting great efforts to deal with these emerging issues, thus enhancing and ensuring safety throughout the world over the years. The Preamble of the Chicago Convention emphasizes safety and order of international air transport, and so many Articles in the Convention are related to the safety. Furthermore, most of the Annexes to the Convention are International Standards and Recommended Practices pertaining to the safety. In particular, Annex 19, which was promulgated in Nov. 2013, dealing with safety management system. ICAO, as law-making body, has Air Navigation Commission, Council, Assembly to deliberate and make decisions regarding safety issues. It is also implementing USOAP and USAP to supervise safety functions of member States. After MH 370 disappeared in 2014, ICAO is developing Global Tracking System whereby there should be no loophole in tracking the location of aircraft anywhere in world with the information provided by many stakeholders concerned. MH 17 accident drove ICAO to install web-based repository where information relating to the operation in conflict zones is provided and shared. In addition, ICAO has been initiating various solutions to emerging issues such as ebola outbreak and operation under extreme meteorological conditions. Considering the necessity of protection and sharing of safety data and information to enhance safety level, ICAO is now suggesting enhanced provisions to do so, and getting feedback from member States. It has been observed that ICAO has been approaching issues towards problem-solving from four different dimensions. First regarding time, it analyses past experiences and best practices, and make solutions in short, mid and long terms. Second, from space perspective, ICAO covers States, region and the world as a whole. Third, regarding stakeholders it consults with and hear from as many entities as it could, including airlines, airports, community, consumers, manufacturers, air traffic control centers, air navigation service providers, industry and insurers. Last not but least, in terms of regulatory changes, it identifies best practices, guidance materials and provisions which could become standards and recommended practices.

• Journal of The Korean Association For Science Education
• /
• v.37 no.1
• /
• pp.49-61
• /
• 2017

This study seeks to explore high school students' changing attitudes toward science-related career choices with focus on the scientific career cultural capital perspective. We developed a questionnaire and surveyed female high school students. We used the survey results to explore the scientific career cultural capital factors and characteristics of the students who have chosen a science-related career. We selected twelve participants through science-related career choice type and then conducted in-depth interviews with them. Through the in-depth interviews, we explored how the factors of scientific career cultural capital affect the students' career choices. The result shows that the students developed internal resources through norms of obligation and reciprocity within the structure (family, school, community) based on external resources. The internal resources affected their science-related career choice. In elementary school, the students' interests in science and science-related occupations are the most influential in their choice of career directions. In high school, however, the students also considered career cognition, aptitude cognition, and career-value factors in their career choices. The interests in science were formed mainly due to education activity and daily-life activity. Over time, education activity's effect increased. The interests in science-related occupations were formed mainly due to daily-life activity during elementary school. In high school, the interests in science-related occupations formed due to diverse internal and external resources. Students who maintained a science-related career attracted external resources. Furthermore, they reinforced internal resources through interacting with various external resources. This study has implications for science education concerning students' science-related career choices.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.43 no.5
• /
• pp.121-131
• /
• 2015

Village-concentrated agricultural regional development projects that aim for increased profits are now being aimed at core agricultural and fishing areas. These agricultural and fishing stimulus projects aim to not only increase the net profit of local citizens but also improve their living conditions. As the industry itself has changed, local rural areas have also changed in various ways. One such aspect is the emergence of rural-urban complexes known as "rurban" areas. These naturally occurring rurban regions are created by a combination of complex urban infrastructure while maintaining the insulated rural communities that agricultural areas are usually so readily identified by. "Jeongeupsi Naejangsnagdong", the target of this research, also specifies the unaltered surrounding natural environment of rural areas while at the same time containing complex central living areas typically found in urban areas. This research suggests that the direction of residents' participation in community-level rurban projects could solve the problems found in existing top-down government development projects. This research also suggests rurban area activation plans to improve living conditions through analysis of both local rurban agricultural characteristics and citizen demands. In order to encourage citizen autonomy and self-governing attitudes, citizen-strengthening workshop programs are proposed, such as citizen workshops or pilot activities. This research was carried out by target area analysis, rudimentary planning, development direction setting, detailed project planning, and finally project processing. This procedure established three goals, which are walking environment improvement, community infrastructure establishment, and good living environment establishments, based on actual site research and citizen demands. This research suggests plans to activate community groups that were already established and reflect citizen needs as the main avenues for local businesses. This research is predicted to promote more active and successful growth through autonomy in stimulating these increasingly emerging rurban agricultural regions.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.11 no.3
• /
• pp.1-15
• /
• 2016

This study has analyzed predication capabilities leveraging multi-variate model, logistic regression model, and artificial neural network model based on financial information of medium-small sized companies list in KOSDAQ. 83 delisted companies from 2009 to 2012 and 83 normal companies, i.e. 166 firms in total were sampled for the analysis. Modelling with training data was mobilized for 100 companies inlcuding 50 delisted ones and 50 normal ones at random out of the 166 companies. The rest of samples, 66 companies, were used to verify accuracies of the models. Each model was designed by carrying out T-test with 79 financial ratios for the last 5 years and identifying 9 significant variables. T-test has shown that financial profitability variables were major variables to predict a financial risk at an early stage, and financial stability variables and financial cashflow variables were identified as additional significant variables at a later stage of insolvency. When predication capabilities of the models were compared, for training data, a logistic regression model exhibited the highest accuracy while for test data, the artificial neural networks model provided the most accurate results. There are differences between the previous researches and this study as follows. Firstly, this study considered a time-series aspect in light of the fact that failure proceeds gradually. Secondly, while previous studies constructed a multivariate discriminant model ignoring normality, this study has reviewed the regularity of the independent variables, and performed comparisons with the other models. Policy implications of this study is that the reliability for the disclosure documents is important because the simptoms of firm's fail woule be shown on financial statements according to this paper. Therefore institutional arragements for restraing moral laxity from accounting firms or its workers should be strengthened.

• Journal of the Korean Institute of Traditional Landscape Architecture
• /
• v.33 no.4
• /
• pp.26-37
• /
• 2015

This paper aims to review all landscaping elements of Donggwoldo by building and identify the palatial garden landscaping characteristics and landscaping methods in a bid to explore landscaping methods applicable for modern-time gardens of Korea through Succession of Tradition. The research methodology was adopted by which the palatial gardens appearing in Donggwoldo were observed according to garden elements to identify their characteristics. Garden elements in Donggwoldo include oddly shaped stones, ponds, buildings and Madang, borders and areas, and trees. Their characteristics were analyzed, and as a result they are outlined as follows. Location : Buildings in Donggwoldo were located in the optimal areas within the Myungdang (the best location), with the building sites being created by transforming the natural topography positively according to the existing topography and uses. Tree planting : The construction of the buildings involved using the existing trees. There were no specific principle and method of planting trees, and no specific criteria for choosing the kind of tree. Symmetrical planting was adopted and its is considered embracing the viewpoint of making gardens based on the expression of Yin and Yang. Strongly symbolical kinds of trees were also adopted. Bangji : it takes a nearly circular shape in palatial gardens, and such shapes represent conceptual and abstract symbols. They were also frequently used as the place of public entertainment. Pavilions : they did not take a certain standard shape. They had diverse shapes, including a triangle, square, pentagon, hexagon, and octagon and cross. Oddly shaped stones : Oddly shapes stones and stone cases were deployed mainly near the bedroom and the crow prince's residence and in the rear garden. Hwagye : it appeared mainly in the back of the bedroom, the crown prince's residence, the princess's residence, and other women's quarters. Chwibyeong : it was installed for the purpose of drawing the natural energy like a natural inlet instead of being the nature-dividing wall. Korea's garden composition method was very different from the Western and Chinese method. Overall, Chosun palatial garden style was characterized by strict and Confucian features, while the garden construction method adopted the Taoism thought. Yet, the gardens had a carefree aspect.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_3
• /
• pp.1173-1185
• /
• 2019

We, for the first time, retrieved sulfur dioxide (SO₂) vertical column density (VCD) in industrial and volcanic areas from TROPOspheric Monitoring Instrument (TROPOMI) using the Principle component analysis(PCA) algorithm. Furthermore, SO₂ VCDs retrieved by the PCA algorithm from TROPOMI raw data were compared with those retrieved by the Differential Optical Absorption Spectroscopy (DOAS) algorithm (TROPOMI Level 2 SO₂ product). In East Asia, where large amounts of SO₂ are released to the surface due to anthropogenic source such as fossil fuels, the mean value of SO₂ VCD retrieved by the PCA (DOAS) algorithm was shown to be 0.05 DU (-0.02 DU). The correlation between SO₂ VCD retrieved by the PCA algorithm and those retrieved by the DOAS algorithm were shown to be low (slope = 0.64; correlation coefficient (R) = 0.51) for cloudy condition. However, with cloud fraction of less than 0.5, the slope and correlation coefficient between the two outputs were increased to 0.68 and 0.61, respectively. It means that the SO₂ retrieval sensitivity to surface is reduced when the cloud fraction is high in both algorithms. Furthermore, the correlation between volcanic SO₂ VCD retrieved by the PCA algorithm and those retrieved by the DOAS algorithm is shown to be high (R = 0.90) for cloudy condition. This good agreement between both data sets for volcanic SO₂ is thought to be due to the higher accuracy of the satellite-based SO₂ VCD retrieval for SO₂ which is mainly distributed in the upper troposphere or lower stratosphere in volcanic region.

• Journal of Venture Innovation
• /
• v.2 no.2
• /
• pp.95-115
• /
• 2019

In the context of today's business environment, not only is the nation or company's credit rating considered very important in our recent society, but it is also becoming important in international transactions. Likewise, at this point of time when the importance and reliability of credit evaluation are becoming important at home and abroad, this study analyzes financial ratios related to corporate profitability, safety, activity, financial growth, and profit growth to study the impact of financial indicators on enterprise value and corporate taxes on credit evaluation. To proceed with this, the financial ratio of 465 companies of KOSPI securities listed in 2017 was calculated and the impact of enterprise value and corporate taxes on credit evaluation was analyzed. Especially, this further study tried to derive a reliable and consistent conclusion by analyzing the financial data of KOSPI securities listed companies for eight years from 2011, which is the first year of K-IFRS introduction, to 2018. Research has shown that the significance levels among variables that show the profitability, safety, activity, financial growth, and profit growth of each financial ratio were significant at the 99% level, except for the profit growth. Validation of the research hypothesis found that while the profitability of KOSPI-listed companies significantly affects corporate value and income tax, indicators such as safety ratio and growth ratio do not significantly affect corporate value and income tax. Activity ratio resulted in significant effects on the value of enterprise value but not significant impacts on income taxes. In addition, it was found that the enterprise value has a significant effect on the company's credit and corporate income taxes, and that corporate income taxes also have a significant effect on the corporate credit evaluation, and this also shows that there is a mediating function of corporate tax. And as a result of further study, when looking at the financial ratio for eight years from 2011 to 2018, it was found that two variables, KARA and LTAX, are significant at a 1% significant level to KISC, whereas LEVE variables is not significant to KISC. The limitation of this study is that credit rating score and financial score cannot be said to be reliable indicators that investors in the capital market can normally obtain, compared to ranking criteria for corporate bonds or corporate bills directly related to capital procurement costs of enterprise. Above all, it is necessary to develop credit rating score and financial score reflecting financial indicators such as business cash flow or net assets market value and non-financial indicators such as industry growth potential or production efficiency.

• Applied Chemistry for Engineering
• /
• v.9 no.3
• /
• pp.457-461
• /
• 1998

The vesicle system of DPPC(dipalmitoylphosphaticylcholine)/Chol(Cholesterol) has been modified by incorporating various mole fractions of flourinated surfactant($C_8F_{17}(CH_2)_2OCO-CH_2CH(SO_3Na)COO(CH_2)_2C_8F_{17}$. Sodium bis(1H,1H,2H,2H-heptadecaflurododecyl)-2-sulfosuccinate, FS)/fluorinated fatty acid salt ($C_7F_{15}COONH_4$, ammoniumpentadecaflurooctyrate, FFS), and their physicochemical properties have been investigated in an attempt to enhance the stability of phospholipid vesicle system. The ${\zeta}$-potential measurement by use of Zetamaster sub-micron Particle Electrophoresis Analyzer (Malvern Co.) showed that a charged homogeneous DPPC/Chol/FS vesicle has been formed owing to the incorporated FFS effect on the membrane, playing a role as a cosurfactant in the bilayer between DPPC and FS components. With increase in the concentration of FFS, it was found that the particle size and also surface charge of the DPPC/Chol/FS vesicle decreased. The stability of DPPC/Chol/FS/FFS liposome was found to be enhanced significantly compared to that of DPPC/Chol/FS according to the dispersity change as a function of time. The release rate of dye molecule of Methylene Blue from the DPPC/Chol/FS/FFS vesicle was determined to be slower than that of DPPC/Chol/FS system, and it may be attributed to the increase in microviscosity of the hydrophobic region in the bilayer. The affinfinity of DPPC/Chol/FS/FFS vesicles to albumin was found to be slightly lowered compared to that of DPPC/Chol/FS. Based on these findings, it was confirmed that a more stable and homogeneous vesicle system of DPPC/Chol/FS could be prepared by addition of FFS, acting as a cosurfactant in the aggregate formation.