• 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.

• Progress in Medical Physics
• /
• v.23 no.2
• /
• pp.81-90
• /
• 2012

The effect of setup uncertainties on CTV dose and the correlation between setup uncertainties and setup margin were evaluated by Monte Carlo based numerical simulation. Patient specific information of IMRT treatment plan for rectal cancer designed on the VARIAN Eclipse planning system was utilized for the Monte Carlo simulation program including the planned dose distribution and tumor volume information of a rectal cancer patient. The simulation program was developed for the purpose of the study on Linux environment using open source packages, GNU C++ and ROOT data analysis framework. All misalignments of patient setup were assumed to follow the central limit theorem. Thus systematic and random errors were generated according to the gaussian statistics with a given standard deviation as simulation input parameter. After the setup error simulations, the change of dose in CTV volume was analyzed with the simulation result. In order to verify the conventional margin recipe, the correlation between setup error and setup margin was compared with the margin formula developed on three dimensional conformal radiation therapy. The simulation was performed total 2,000 times for each simulation input of systematic and random errors independently. The size of standard deviation for generating patient setup errors was changed from 1 mm to 10 mm with 1 mm step. In case for the systematic error the minimum dose on CTV $D_{min}^{stat{\cdot}}$ was decreased from 100.4 to 72.50% and the mean dose $\bar{D}_{syst{\cdot}}$ was decreased from 100.45% to 97.88%. However the standard deviation of dose distribution in CTV volume was increased from 0.02% to 3.33%. The effect of random error gave the same result of a reduction of mean and minimum dose to CTV volume. It was found that the minimum dose on CTV volume $D_{min}^{rand{\cdot}}$ was reduced from 100.45% to 94.80% and the mean dose to CTV $\bar{D}_{rand{\cdot}}$ was decreased from 100.46% to 97.87%. Like systematic error, the standard deviation of CTV dose ${\Delta}D_{rand}$ was increased from 0.01% to 0.63%. After calculating a size of margin for each systematic and random error the "population ratio" was introduced and applied to verify margin recipe. It was found that the conventional margin formula satisfy margin object on IMRT treatment for rectal cancer. It is considered that the developed Monte-carlo based simulation program might be useful to study for patient setup error and dose coverage in CTV volume due to variations of margin size and setup error.

• Journal of the Society of Disaster Information
• /
• v.16 no.4
• /
• pp.701-711
• /
• 2020

Purpose: In this study, the risk of flammability of a liquid mixture was experimentally confirmed because the purpose of this study was to confirm the increase or decrease of the flammability risk in a mixture of two substances (combustible+combustible) and to present the risk of the mixture. Method: Flash point test method and result processing were tested based on KS M 2010-2008, a tag sealing test method used as a flash point test method for crude oil and petroleum products. The manufacturer of the equipment used in this experiment was Japan's TANAKA. The flash point was measured with a test equipment that satisfies the test standards of KS M 2010 with equipment produced by the company, and LP gas was used as the ignition source and water as the cooling water. In addition, when measuring the flash point, the temperature of the cooling water was tested using cooling water of about 2℃. Results: First of all, in the case of flammable + combustible mixtures, there was little change in flash point if the flash point difference between the two substances was not large, and if the flash point difference between the two substances was low, the flash point tended to increase as the number of substances with high flash point increased. However, in the case of toluene and methanol, the flash point of the mixture was lower than that of the material with a lower flash point. Also, in the case of a paint thinner, it was not easy to predict the flash point of the material because it was composed of a mixture, but as a result of experimental measurement, it was measured between -24℃ and 7℃. Conclusion: The results of this study are to determine the risk of mixtures through experimental studies on flammable mixtures for the purpose of securing the effectiveness of the details of the criteria for determining dangerous goods in the existing dangerous goods safety management method and securing the reliability and reproducibility of the determination of dangerous goods Criteria have been presented, and reference data on experimental criteria for flammable liquids that are regulated in firefighting sites can be provided. In addition, if this study accumulates know-how on differences in test methods, it is expected that it can be used as a basis for research on risk assessment of dangerous goods and as a basis for research on dangerous goods determination.

• (The) Research of the performance art and culture
• /
• no.20
• /
• pp.251-307
• /
• 2010

Manbojeonseo, a kind of an encyclopedia published several times in Ming Ch'ing dynasty, includes useful information for scholars and common people on daily lives. In 1720, Manbojeonseo was first introduced to Joseon(朝鮮) dynasty by the diplomatic corps visiting Ch'ing dynasty, and widely circulated in the society as an useful information magazine or an individual collection of reference book. Since Manbojeonseo includes the systematically-organized contents of Geumdoron(琴道論, a theory of a heptachord), it could provide a useful reference when the Geumdoron was inserted as the contents of old scores. For an instance, Obultan(五不彈), Tangeumsuji(彈琴須知), and Taeeumgibeop(太音紀法) recorded in Hangeumsinbo(韓琴新譜, 1724) clearly acknowledge Manbojeonseo as their common source. In this paper, the order and the contents of Geumdorons from four different Manbojeonseo are compared. At first, the comparative analysis of Manbojeonseo (1610) edited by Seo Giryong(徐企龍) and Manbojeonseo(1612) edited by Yu Jamyeong(劉子明) are carried out focusing on the contents of the Geumdoron, where both Manbojeonseos contain considerable amount of Geumdoron sections. The tables of the contents in both Manbojeonseos are composed of upper and lower levels classified into 4 large divisions for each. While the contents of the upper level is presumably older and focused more on the theory of the cardinal virtues, the contents of the lower one is relatively new and centered more on the skills for the real play of a heptachord(琴), the lyrics and the musical scores composed of Gamjabo(減字譜). Therefore, it could be said that the upper level is metaphysical while the lower level is physical. One of the differences between those two Manbojeonseos lies in the order and the terminology found in the large divisions. In the case of Manbojeonseo(1612), some terms in the large division represent and theoretically group the detailed descriptions in the small divisions such as 5 demands or 7 taboos in the play of the heptachord. In addition, a few lower divisions were newly added or revised in order to enhance the completeness of Geumhangmun(琴學門, study of a heptachord), and the detailed classification was revised and polished to improve the reasonableness. In Manbojeonseo(1614) composed by the same editor as Manbojeonseo(1610), the contents of the Geumdoron become much briefer than those of Manbojeonseo(1610) and Manbojeonseo(1612). In the case of Manbojeonseo(1739), a new type of the Geumdoron is included called Oeumjeongjobo(五音正操譜) while carrying a similarly brief section of the Geumdoron. Finally, the Geumdorons in Manbojeonseo and several old scores are comparatively analyzed. While the Geumbo(琴譜) owned by Gugagwon(國樂院) and Hangeumsinbo contains relatively old Geumdoron, Yuyeji(遊藝志) and Bangsanhanssigeumbo(芳山韓氏琴譜) adopt practical and relatively new Geumdorons different from the former old scores and similar to Manbojeonseo(1739) considering the order and the contents. In particular, the contents of the Geumdoron in Geumheonakbo(琴軒樂譜) is notably unique containing much of the upper and the lower levels of Manbojeonseo(1612), therefore thought to have actively adopted the contents of new Geumdorons.

• Pediatric Gastroenterology, Hepatology & Nutrition
• /
• v.3 no.2
• /
• pp.181-187
• /
• 2000

Purpose: The purpose of our study is to provide useful information for the prevalence of obesity by the standard weight for height and the relationship of the body image to dieting methods for weight control in children. Methods: The survey was performed by an anthropometry and a special questionnaire on children in grades 5 to 6 at elementary schools from May to June 1997 in Kwangju city. Results: 1) Selected individuals were categorized into obese, normal, thin by the standard weight for height and body image. The prevalence of obesity by the standard weight for height was 19.5%, normal was 70.5%, thin was 10.1% in males and respectively 10.5%, 66.3%, 23.1% in females. The prevalence of obesity in regards to body image was 18.4%, normal was 58.4%, thin was 23.2% in males and respectively 24.5%, 58.6%, 16.9% in females. 2) The obese body image was 6.8% in normal and thin groups in males and 19.4% in females. The body image of obese children who do not perceive themselves as obese was 32.6% in males and 28.7% in females. 3) The standard body image was 88.8% as realized by themselves, 77.9% when they compared themselves to friends, 62.7% as told by parents, 56.8% as told by friends, and 29.9% when they compared themselves to celebrities. 4) The source of information for dieting treatment was 44% from radio and television, 30% from books, newspapers, magazines, 17% from family and 9% from friends and seniors. 5) The prevalence of dieting according to the standard weight for height was 18.0% in thin, 18.7% in normal and 36.7% in obese group in males and respectively 17.8%, 22.4%, 46.3% in females. There was a statistically significant correlation between males and females (p<0.0005). 6) The prevalence of dieting up to 2 weeks in the thin and normal groups as recognized by weight centile was 17.3% according to body image. The prevalence of dieting up to 2 weeks in the obese groups was 37.2% according to body image (p<0.001). Conclusion: We suggest that many students perceived their body size incorrectly and tried weight control. The incorrect perception of body size seems to cause the problem of failure to thrive, malnutrition etc. It is necessary for further study to find and prevent side effects by unnecessary weight control and to provide proper health education and management about obesity.

• Journal of Preventive Medicine and Public Health
• /
• v.35 no.3
• /
• pp.245-254
• /
• 2002

Objective : To test if the intake of $H_2$ receptor antagonists ($H_2$-RAs) increases the risk of gastric cancer in the elderly. Methods : The source population for this study was drawn from the responders to a questionnaire survey administered to the Korea Elderly Pharmacoepidemiological Cohort (KEPEC), who were beneficiaries of the Korean Medical Insurance Corporation, were at least 65 years old, and residing in Busan in 1999. The information on $H_2$-RAs exposure was obtained from a drug prescription database compiled between inn. 1993 and Dec. 1994. The cases consisted of 76 gastric cancer patients, as confirmed from the KMIC claims data, the National Cancer Registry and the Busan Cancer Registry. The follow-up period was from Jan. 1993 to Dec. 1998. Cancer free controls were randomly selected by 1:4 individual matching, which took in to consideration the year of birth and gender. Information on confounders was collected by a mail questionnaire survey. The odds ratios, and their 95% confidence intervals, were calculated using a conditional logistic regression model. Results : After adjusting for a history of gastric ulcer symptoms, medication history, and body mass index, the adjusted OR (aOR) was 4.6 (95% CI=1.72-12.49). The odds ratio of long term use (more than 7 days) was 2.3 (95% CI=1.07-4.82). The odds ratio of short term use was 4.6 (95% CI=1.26-16.50). The odds ratio of parenteral use was 4.4 195% CI=1.16-17.05) and combination use between the oral and parenteral routes (aOR, 16.8; 95% CI=1.21-233.24) had the high risk of gastric cancer. The aOR of cimetidine was 1.7 (95% CI=1.04-2.95). The aOR of ranitidine was 2.0 (95% CI=1.21-3.40). The aOR of famotidine was 1.7 (95% CI=0.98-2.80). Conclusion : The intake of $H_2$-RAs might increase the risk of gastric cancer through achlorhydria in the elderly.

• Korean Journal of Heritage: History & Science
• /
• v.54 no.3
• /
• pp.126-149
• /
• 2021

The reign of King (later Emperor) Gojong of the Joseon dynasty (later the Korean Empire), which lasted from 1863 to 1907, was a period of turmoil caused by political and social instability at home and military incursions by foreign powers. It was also a period in which Joseon was proclaimed as the Korean Empire and, accordingly, the royal seals of the dynasty underwent a major change. Records show that some 135 seals were produced during the reign of Gojong. The present study of the makers and styles of the royal seals aims to reveal a wealth of information on the seals, which typically had handles in the shape of either a turtle or dragon. Among the seal makers of this era, Jeon Heung-gil was particularly highly regarded as a Golden Royal Seal Artisan (Geumbojang) because he was highly skilled at making both turtle and dragon handles. Kim Eun-seok, a master Jade Royal Seal Artisan (Okbojang), also excelled in the production of turtle and dragon handles for his jade seals. Another master Jade Royal Seal Artisan, Yi Jung-ryeo, is noteworthy because he developed a new style of jade royal seals, which eventually became the dominant style after 1890, when Kim Eun-seok was not active. Furthermore, after the 1890s, his style was also applied to the production of the gold royal seals and developed as the dominant style. Regarding the dragon handles adopted after the proclamation of the Korean Empire, both the golden and the jade royal seals were made in the same style by the same artisans in the service of the Joseon dynasty. They adopted the style of Chinese imperial seals when they began making seals with dragon handles for the Korean Empire, although the basic shapes of both handle and dragon were copied from those made during the Joseon period. As a ceremonial object symbolizing the authority and legitimacy of the royal or imperial family, the style of the royal seals was influenced mainly by changes in the political situation at home and abroad. As Gojong's reign was a period in which more royal seals were made than in the reign of any other ruler, the seals originating from his reign constitute a richer source of information about the efforts of the dynasty to preserve the tradition while effectively dealing with the changes of the new era.

• Korean Journal of Heritage: History & Science
• /
• v.49 no.2
• /
• pp.4-19
• /
• 2016

Jangkei(狀啓) made to the Royal Court by Yi Sun-sin during the Japanese invasions of Korea is handed down under the names of Jangcho(狀草), Keicho(啓草), Keibon(啓本) and others depending on copying patterns of those times and later times as it was copied out by a third person. In particular, "YimjinJangcho(壬辰狀草)" which Yi drew up during his service as the director of the naval forces in Jeolla Jwasooyeong is known as the most popular Jangkei. "ChungMinKongKeicho" which has been re-located recently after loss is a national treasure level cultural property as valuable as "YimjinJangcho" and should be treated as a model of Yi Sun-sin's other Jangkeis by next generations. As of now, however it is not confirmed if it is a totally new book related to Yi Sun-sin or is supplementary to the lost Jangkei, this study decided to ascertain relevant information through a bibliographic discussion on the question. "Chungmin(忠愍)" was the title that was used after the death of Yi Sun-sin, and "ChungMinKongKeicho" was completed when Jangkei was copied in 1662. 12 books that would not be found in YimjinJangcho are included in the book and such books are also present in the Jangkei supplement which has been known lost so far. What should be especially focused on here is that the forms and contents of these (11) photographs that Japanese shot from "ChungMinKongKeicho" in 1928 turned out to be completely identical to those of the original copy. The point that Korean History Compilation Committee added the 12 books to Jangkei as referring to the book as "One Keicho(啓草) partially copied(抄寫) in separation" and that Cho Sung-do categorized the 12 books into a supplement and others can be solid proofs to make the Jangkei supplement called "ChungMinKongKeicho". In terms of "ChungMooKongKeicho", since it consists of 62 books in total, it is not reasonable to see the book as Jangkei supplement which has the extra 12 more books for itself. "ChungMooKongKeibon" in "ChungMooKongYusa" was written with a total of 16 books. In the body, Yidumun is only clearly present, and the three books in the later part are same with the original copy of "ChungMooKongKeicho". "YimjinJangcho" by Korean History Compilation Committee has been the only book in which Yidumun was observed so far but now, it is assumed that the publication date of "ChungMooKongKeibon" goes before that of the former. The counterargument to the opinion that "ChungMinKongKeicho" is the supplement to Jangkei is based on Lee Eun-sang's comment "One page of a log in the Jangkei copy supplement." At first Seol Ui-sik introduced a piece photo of the rough draft of "MoosulIlki" in a drawing form through "Nanjung Ilkicho by Yi Sun-sin" in 1953. Lee Eun-sang also added two pages of the handwritten Yilkicho in the Jangkeichobon supplement to "MoosulIlki" and for the second time, the phrase "One page of a log written during the last 10 days after the Jangkei copy supplement" and "Supplement" were used. Those views are originated from the comment "One photograph of the rough draft of "MoosulIlki"" which Seol Ui-sik introduced without knowledge of the exact source. Lee Eun-sang said, "One page of a log in the Jangkei copy supplement" because Lee mistook "ChungMooKongYusa" for a book related to Jangkei. Since it is the wrong argument different from the actual situation of the original copy, if it has to be corrected, it should be rephrased "One page of a log in ChungMooKongYusa." After all, the source of the counterargument is the mistake because there has never been the Jangkei supplement with one page of a log included. All the Jangkeis other than "YimjinJangcho" can be said as the Jangkei supplements but still, they are separated from the other Jangkeis for the extra 12 more books are present in the commonly-called Jangkei supplement. Due to that reason, the argument on how "ChungMinKongKeicho" with the 12 books added is the popular Jangkei supplement should be considered more reasonable.

• Korean Security Journal
• /
• no.31
• /
• pp.73-100
• /
• 2012

Republic of Korea for the last 2012 nuclear security summit in Seoul in three days from March 26, has held great success. This industry watchers at home and abroad through professional meetings and private organizations to promote global nuclear security was intended to draw, for the preparation of this event the most important one of the things that should be addressed at the meeting was over the top guard. The national guard of heads of state and nationally significant as the important issues should be dealt with. So who is responsible for national security summit national guard for a successful guard against participants should have a higher understanding, these stories ever to be dealt with important security subjects is the psychological state of cognition. This study is currently a source of honor to escort the psychological state of subjects, and whether the search as I looked at cognition. Qualitative research methods to conduct research in-depth interview technique was used for one of the transcribed data analysis program, a non-quantitative data were analyzed using NVivo 8. As a result of this study, analysis of the problem compared with the usual methods 39 (72.2%), fine-grained observation and 15 (27.7%), through the media how cognition 41 (73.2%), marginal help of 15 (26.7%) were. The usual comparison with how honored Issue sources to detect security of those words and actions and facial expressions, etc. Security of subjects over the state of mind cognition that he, detailed observation of the national guard to security subjects the psychological state of cognition which are pre-Thorough Information subjects that can have all the information about security and the state can be obtained by saying that. Media coverage through a cognitive approach national guard of the media coverage to understand the security subjects of the political situation to understand the prerequisite that were marginal for help through the cognitive approach of security subjects surrounding the execution of security workers, Secretary in charge of protocol and security, and be propagated through the selection techniques to be utilized was that.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.41 no.2
• /
• pp.102-109
• /
• 2021

In this study, the effect of forage sources in the total mixed ration (TMR) on in vitro goat rumen fermentation was investigated. Rice straw (RS), Italian ryegrass (IRG), timothy (TIM), and alfalfa (ALF) were used as forage sources. Each forage source was mixed with a commercial goat concentrate diet in the ratio of 1:1. Total 4 TMR were prepared. Rumen simulated in vitro fermentation using goat rumen fluid collected from the slaughterhouse was conducted until 72^th. For fermentation parameters, gas production (GP), volatile fatty acids (VFAs), and ammonia nitrogen (NH₃-N) were examined. All assays were performed at 24^th, 48^th, and 72^th h of incubation individually. Contents of crude protein and non-fibrous carbohydrate were greater in the order of RS < IRG < TIM < ALF. Significant treatment effects were found in valerate and NH₃-N at 24^th h of incubation (p<0.05). ALF showed the greatest contents of them and RS was the lowest. At 48^th incubation, a significant effect was detected at GP (p<0.05) and RS was greater than others. However, GP of RS was lower than others at 72^th. Significant effects on Total VFA, butyrate, and valerate productions were found at 72^th h of incubation (p<0.05). ALF showed the greatest production. Methane production from all treatments was not significantly different for each incubation time (p>0.05). The present study provided primary information on how goat rumen fermentation responds to different nutrient contents and forage sources of TMR. And the information could be used for the design or optimizing economical diet formulation for goats.