• 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.

• The Journal of Korean Society for Radiation Therapy
• /
• v.8 no.1
• /
• pp.19-27
• /
• 1996

I. Project Title A Study of Brachytherapy for intraocular tumor II. Objective and Importance of the project The eye enucleation or external-beam radiation therapy that has been commonly used for the treatment of intraocular tumor have demerits of visual loss and in deficiency of effective tumor dose. Recently, brachytherapy using the plaques containing radioisotope-now treatment method that decrease the demerits of the above mentioned treatment methods and increase the treatment effect-is introduced and performed in the countries, Our purpose of this research is to design suitable shape of plaque for the ophthalmic brachytherapy, and to measure absorbed doses of Ir-192 ophthalmic plaque and thereby calculate the exact radiation dose of tumor and it's adjacent normal tissue. III. Scope and Contents of the project In order to brachytherapy for intraocular tumor, 1. to determine the eye model and selected suitable radioisotope 2. to design the suitable shape of plaque 3. to measure transmission factor and dose distribution for custom made plaques 4. to compare with the these data and results of computer dose calculation models IV. Results and Proposal for Applications The result were as followed. 1. Eye model was determined as a 25mm diameter sphere, Ir-192 was considered the most appropriate as radioisotope for brachytherapy, because of the size, half, energy and availability. 2. Considering the biological response with human tissue and protection of exposed dose, we made the plaques with gold, of which size were 15mm, 17mm and 20mm in diameter, and 1.5mm in thickness. 3. Transmission factor of plaques are all 0.71 with TLD and film dosimetry at the surface of plaques and 0.45, 0.49 at 1.5mm distance of surface, respectively. 4. As compared the measured data for the plaque with Ir-192 seeds to results of computer dose calculation model by Gary Luxton et al. and CAP-PLAN (Radiation Treatment Planning System), absorbed doses are within ${\pm}10\%$ and distance deviations are within 0.4mm Maximum error is $-11.3\%$ and 0.8mm, respectively. As a result of it, we can treat the intraocular tumor more effectively by using custom made gold plaque and Ir-192 seeds.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.53-74
• /
• 2018

Since South Korea has an economic structure that has a characteristic which market-dependent on overseas, the tourism industry is considered as a very important industry for the national economy, such as improving the country's balance of payments or providing income and employment increases. Accordingly, the necessity of more accurate forecasting on the demand in the tourism industry has been raised to promote its industry. In the related research, economic variables such as exchange rate and income have been used as variables influencing tourism demand. As information technology has been widely used, some researchers have also analyzed the effect of media on tourism demand. It has shown that the media has a considerable influence on traveler's decision making, such as choosing an outbound destination. Furthermore, with the recent availability of online information searches to obtain the latest information and two-way communication in social media, it is possible to obtain up-to-date information on travel more quickly than before. The information in online media such as blogs can naturally create the Word-of-Mouth effect by sharing useful information, which is called eWOM. Like all other service industries, the tourism industry is characterized by difficulty in evaluating its values before it is experienced directly. And furthermore, most of the travelers tend to search for more information in advance from various sources to reduce the perceived risk to the destination, so they can also be influenced by online media such as online news. In this study, we suggested that the number of online media posting, which causes the effects of Word-of-Mouth, may have an effect on the number of outbound travelers. We divided online media into public media and private media according to their characteristics and selected online news as public media and blog as private media, one of the most popular social media in tourist information. Based on the previous studies about the eWOM effects on online news and blog, we analyzed a relationship between the volume of eWOM and the outbound tourism demand through the panel model. To this end, we collected data on the number of national outbound travelers from 2007 to 2015 provided by the Korea Tourism Organization. According to statistics, the highest number of outbound tourism demand in Korea are China, Japan, Thailand, Hong Kong and the Philippines, which are selected as a dependent variable in this study. In order to measure the volume of eWOM, we collected online news and blog postings for the same period as the number of outbound travelers in Naver, which is the largest portal site in South Korea. In this study, a panel model was established to analyze the effect of online media on the demand of Korean outbound travelers and to identify that there was a significant difference in the influence of online media by each time and countries. The results of this study can be summarized as follows. First, the impact of the online news and blog eWOM on the number of outbound travelers was significant. We found that the number of online news and blog posting have an influence on the number of outbound travelers, especially the experimental result suggests that both the month that includes the departure date and the three months before the departure were found to have an effect. It is shown that online news and blog are online media that have a significant influence on outbound tourism demand. Next, we found that the increased volume of eWOM in online news has a negative effect on departure, while the increase in a blog has a positive effect. The result with the country-specific models would be the same. This paper shows that online media can be used as a new variable in tourism demand by examining the influence of the eWOM effect of the online media. Also, we found that both social media and news media have an important role in predicting and managing the Korean tourism demand and that the influence of those two media appears different depending on the country.

• Korea journal of population studies
• /
• v.26 no.1
• /
• pp.1-30
• /
• 2003

This paper attempts to explore the effects of ecological and socioeconomic factors on the level of mortality and the changing trends of such effects during the period of 1990∼2000. For this purpose the population census data and micro-data from the vital statistics for years 1990, 1995 and 2000 were used. As indicators of mortality, the crude death rate(CDR), the standardized death rate(SDR) and the longevity rate were calculated for 170 'Si' s and 'Gun's. Using GIS, this paper first presents the mortality and longevity maps for years 1990, 1995 and 2000. Then ANOVA and regression analyses are carried out in an effort to generalize the effects of ecological and socioeconomic factors on the CDR, the SDR and the longevity rate. When the mortality and longevity maps are examined, three indices of mortality are found to be markedly high in the southwest coastal regions of Cholla-Nam-Do. By contrast, Seoul and Pusan metropolitan areas show substantially low level of mortality and longevity in these indices. It is also found that the regional differentials in the SDR and the longevity rate show a trend of becoming smaller after 1990. The research, however, does not find any linear relationship between the SDR and the longevity rate. The causal mechanisms of the two indices are found to be different. The results of the ANOVA and the regression analysis reveal that the locational factors of both mountainous and farming regions tend to increase the CDR and SDR while both coastal and farming regions disclose a tendency of increasing the longevity rate. The level of statistical significance of these analytical results is found to be weaker when socioeconomic factors such as education, income, marital status, availability of medical care, and sanitary conditions of the region are taken into account. The regional differentials in the mortality level seem to have a clear relationship not only with the socioeconomic factors but also with the age structure influenced by the age selectivity of migration during the past 40 years.

• Agribusiness and Information Management
• /
• v.8 no.1
• /
• pp.17-26
• /
• 2016

India is largest producer of banana in the world producing 29.72 million tonnes from an area of 0.803 million ha with a productivity of 35.7 MT ha-1 and accounted for 15.48 and 27.01 per cent of the world's area and production respectively (www.nhb.gov.in). In India, Tamil Nadu leads other states both in terms of area and production followed by Maharashtra, Gujarat and Andhra Pradesh. In Rayalaseema region of Andhra Pradesh, Kurnool district had special reputation in the cultivation of banana in an area of 5765 hectares with an annual production of 2.01 lakh tonnes in the year 2012-13 and hence, it was purposively chosen for the study. On $23^{rd}$ November 2003, the Government of Andhra Pradesh has commenced a comprehensive project called 'Andhra Pradesh Micro Irrigation Project (APMIP)', first of its kind in the world so as to promote water use efficiency. APMIP is offering 100 per cent of subsidy in case of SC, ST and 90 per cent in case of other categories of farmers up to 5.0 acres of land. In case of acreage between 5-10 acres, 70 per cent subsidy and acreage above 10, 50 per cent of subsidy is given to the farmer beneficiaries. The sampling frame consists of Kurnool district, two mandals, four villages and 180 sample farmers comprising of 60 farmers each from Marginal (<1ha), Small (1-2ha) and Other (>2ha) categories. A well structured pre-tested schedule was employed to collect the requisite information pertaining to the performance of drip irrigation among the sample farmers and Data Envelopment Analysis (DEA) model was employed to analyze the performance of drip irrigation in banana farms. The performance of drip irrigation was assessed based on the parameters like: Land Development Works (LDW), Fertigation costs (FC), Volume of water supplied (VWS), Annual maintenance costs of drip irrigation (AMC), Economic Status of the farmer (ES), Crop Productivity (CP) etc. The first four parameters are considered as inputs and last two as outputs for DEA modelling purposes. The findings revealed that, the number of farms operating at CRS are more in number in other farms (46.66%) followed by marginal (45%) and small farms (28.33%). Similarly, regarding the number of farmers operating at VRS, the other farms are again more in number with 61.66 per cent followed by marginal (53.33%) and small farms (35%). With reference to scale efficiency, marginal farms dominate the scenario with 57 per cent followed by others (55%) and small farms (50%). At pooled level, 26.11 per cent of the farms are being operated at CRS with an average technical efficiency score of 0.6138 i.e., 47 out of 180 farms. Nearly 40 per cent of the farmers at pooled level are being operated at VRS with an average technical efficiency score of 0.7241. As regards to scale efficiency, nearly 52 per cent of the farmers (94 out of 180 farmers) at pooled level, either performed at the optimum scale or were close to the optimum scale (farms having scale efficiency values equal to or more than 0.90). Majority of the farms (39.44%) are operating at IRS and only 29 per cent of the farmers are operating at DRS. This signifies that, more resources should be provided to these farms operating at IRS and the same should be decreased towards the farms operating at DRS. Nearly 32 per cent of the farms are operating at CRS indicating efficient utilization of resources. Log linear regression model was used to analyze the major determinants of input use efficiency in banana farms. The input variables considered under DEA model were again considered as influential factors for the CRS obtained for the three categories of farmers. Volume of water supplied ($X_1$) and fertigation cost ($X_2$) are the major determinants of banana farms across all the farmer categories and even at pooled level. In view of their positive influence on the CRS, it is essential to strengthen modern irrigation infrastructure like drip irrigation and offer more fertilizer subsidies to the farmer to enhance the crop production on cost-effective basis in Kurnool district of Andhra Pradesh, India. This study further suggests that, the present era of Information Technology will help the irrigation management in the context of generating new techniques, extension, adoption and information. It will also guide the farmers in irrigation scheduling and quantifying the irrigation water requirements in accordance with the water availability in a particular season. So, it is high time for the Government of India to pay adequate attention towards the applications of 'Information and Communication Technology (ICT) and its applications in irrigation water management' for facilitating the deployment of Decision Supports Systems (DSSs) at various levels of planning and management of water resources in the country.

• Korean Journal of Community Nutrition
• /
• v.25 no.5
• /
• pp.416-429
• /
• 2020

Objectives: This study examined the personal, interpersonal and community factors related to food sufficiency and variety among Korean adults using data from the 2017 Community Health Survey. Methods: A total of 228,310 adults aged ≥ 19 years were classified into three groups: food sufficiency with variety, food sufficiency without variety and food insufficiency. Personal factors included sociodemographic characteristics, health behavior and health status. Interpersonal factors included social networking and social activities, and community factors included safety, natural environment, living environment, availability of public transportation and health care services. The association of food sufficiency and variety with interpersonal and community factors was assessed using multivariable logistic regression analyses. Results: Of the total sample, the food-sufficiency-without-variety group and food insufficiency group accounted for 31.5% and 3.2%, respectively. The sociodemographic factors associated with food insufficiency and non-variety were women, ≥ 65 years of age, with low education level, low household income, unemployed, single, and living in areas of small population sizes. There were significant differences in health behavior and health status, interpersonal and community factors among the three groups. Multivariable logistic regression analyses conducted after adjusting for confounding factors showed that lack of social networking and social activities and lower satisfaction derived from community environments were associated with the risk of food insufficiency and non-variety. Conclusions: Our results showed that interpersonal and community factors as well as personal factors were related to food sufficiency and variety. Therefore, public policies to help build social networks and participation in social activities, and improve community environment are needed together with food assistance to overcome the problems of food insufficiency and non-variety.

• Korean Journal of Ecology and Environment
• /
• v.46 no.1
• /
• pp.33-40
• /
• 2013

From March 2012 to January 2013, this study was conducted as a part of the project for making a precise electronic ecological zoning map of vegetation on a highly reduced scale of 1 to 5,000 with a view to improving management efficiency of national parks and enlarging the availability of the data produced from the basic research monitoring the resources of national parks. For the research accuracy and rapidity, a vegetation map was specially created for the on-the-site-vegetation research. To make the map more meticulous, we categorized the vegetation database into five groups: broadleaved forest, coniferous forest, mixed forest, rock vegetation and miscellaneous one. After comparing the results of the data built for the vegetation research and the actual research findings, it was made clear that vegetation of both categories was almost the same in case of broad-leaved forest with 72.20% and 78.45% respectively, and also equivalent in other groups like, for example, coniferous forest (16.70%, 13.41%), mixed forest (9.50%, 7.49%) and rock vegetation (0.60%, 0.15%). According to the precise vegetation map produced from the research, the deciduous broad-leaved forest was the most widely prevalent type in the correlated hierarchical classification of vegetation, occupying 65.78% of the overall vegetation. It was followed by mountain valley forest (15.17%), coniferous forest (10.90%), and plantation forest (7.00%) in order. It is particularly noteworthy that Mt. Deogyusan national park has retained a very stable and versatile forest vegetation in the outstanding state since approximately 20% of the mountain turns out to belong to the I grade vegetation conservation classification which contains climax forests, unique vegetation, subalpine vegetation, matured stands which are older than 50 years and etc.

• Journal of the Korean Society for Library and Information Science
• /
• v.24
• /
• pp.9-51
• /
• 1993

The purpose of this study is twofold: to investigate into general characteristics of the abstracting services in Korea and to discuss general directions of development of the abstracting services in the country. This study is designed to achieve the purpose by gathering and analysing data related to the abstracting journals published in the past ten years and by comparing the results with similar data gathered by the investigator in 1984. The major conclusions made in this study is summarised as follows. (1) Researchers and professionals working in limited numbers of subject fields are benefited by abstracting services of recent achievements in research and development in Korea. Those in most of the fields have essentially no abstracting services of such achievements. Even many researchers and professionals in the limited numbers of the fields that have some elementary abstracting services are not informed of research results in their fields because the abstracting journals are scattered in many narrow subjects and in many cases, the abstracting journals only cover publications in some specific forms and kinds. (2) Abstracting journals of general subjects, which are supposed to be of more or less help to the researchers in the subject fields that have no abstracting journals of their own, have rapidly increased in number in the past ten years. Most of such abstracting journals carry thesis and dissertation abstracts, and the rest those of research papers published in specific places, in specific forms, by specific institutes, and of reports of research projects sponsored by specific foundations. These abstracting journals are not of the kind that comprehensively provide general readers with current awareness of publications of research results in Korea. (3) Most of the abstracting journals existing in Korea are published by institutions of higher education and research institutes, and the rest by commercial publishers, industrial firms, libraries, information centers, government agencies, research foundations, learned societies, etc. Those which publish many titles are small in number and those publish one or two titles are large in number. The former is largely made up of institutions of higher education and research institutes. (4) Ten years ago, there was not a single publishing house that produced abstracting journals. Three commercial publishing houses now produce abstracting journals. As this change occurs, centers of excellence are founded and competitive elements are introduced in abstracting services. This change, in turn, is expected to improve quality of the other abstracting journals in Korea. (5) The abstracting journals published in Korea are classified by type into those of dissertations, research papers, journal articles, patent specifications in that descending order. The fact that Master's and doctoral dissertation abstracts are dominating in Korea is due to the irrational practice of publishing those abstracts at many institutions. (6) Most of the abstracting journals existing in Korea are published by national or government-supported research institutes in order to publicise their own research outputs. Their coverage of literature is normally narrow, and naturally their value to users is limited. (7) The abstracting journals published in Korea increased in number at the rate of $77.8-100\%$ every five years in the past twenty-five years. Most of the abstracting journals that ceased to be published during the period survived for two years. (8) Korean is the desirable language for the abstracting journals designed to be distributed within Korea. About half of the abstracting journals published in Korea is printed in Korean and the other half in foreign languages, and in Korean with foreign languages. All the abstracting journals in foreign languages are printed in English xcept one, which is printed in Japanese. (9) Some twenty percent of the abstracting journals in Korea is published monthly, bimonthly, and quarterly. Others are published annually, biannually, and irregularly. The latter may not function properly as a current-awareness tool due to long intervals between their issues. It is particularly undesirable that about half of the abstracting journals in Korea is published irregularly. Most of the abstracting journals published in Korea are distributed freely to individuals and institutions selected by the publishers. (10) The abstracting journals published by the use of computers increased drastically in the past ten years. The abstracting journals produced by the conventional type-setting method will probably disappear In Korea in another ten years to come. Automation of the production of abstracting journals does not simply mean technical, economic improvement of publishing processes but availability of machine-readable databases that can be used for other purposes, including the generation of other publications and the provision of machine literature searching capabilities. Necessary steps should be taken for this important development that is occurring in the abstracting services in Korea.

• Journal of Food Hygiene and Safety
• /
• v.30 no.3
• /
• pp.217-226
• /
• 2015

This study was conducted to establish a standardized process for developing food safety health indicators. With this aim, we proposed a standardized process, accessed the validity of the suggested process by performing simulations, and provided a method to utilize the indicators. Developing process for domestic environmental health indicators was benchmarked to propose a standardized process for developing food safety health indicators, and DPSEEA framework was applied to the development of indicators. The suggested standardized process consists of an exploitation stage and a management stage. In the exploitation stage, a total of 6 procedures (initial indicators suggestion, candidate indicators selection, data availability assessment, feasibility assessment, pilot study, and final indicator selection) are conducted, and the indicators are routinely calculated and officially announced in the management stage. The exploitation stage is operated by an interaction between a task force team who manages the overall process, and an advisory committee (minimum of 4 in academia, 2 in research, 4 in specialists of Ministry of Food and Drug Safety) who reviews and performs evaluations on the indicators. The standardized process was simulated with 45 initial indicators, and total of 4 indicators (17 detailed indicators) were selected: 'Proportion of domestic fruit/vegetable receiving 'acceptable' in the evaluation of pesticide/herbicide residues', 'Food-borne disease outbreaks', 'Food-borne legal infectious disease incidence', 'Salmonellosis incidence'. Synthetic food safety health index was derived by calculating percent difference with the data from 2010 to 2012. Results showed that when comparing the year 2010 to 2011, and 2011 to 2012, the overall food safety status improved by 10.37% and 9.87%, respectively. In addition, the contribution of indicators to the overall food safety status can be determined by looking into the individual indicators, and the synthetic index may be illustrated to enhance the ease of interpretation to the public and policy makers. In overall, food health safety indicators can be useful in many ways and therefore, attention should be drawn to conduct further studies and establish related legislations.

• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.27-65
• /
• 2020

Many companies on information and communication technology make public their own developed AI technology, for example, Google's TensorFlow, Facebook's PyTorch, Microsoft's CNTK. By releasing deep learning open source software to the public, the relationship with the developer community and the artificial intelligence (AI) ecosystem can be strengthened, and users can perform experiment, implementation and improvement of it. Accordingly, the field of machine learning is growing rapidly, and developers are using and reproducing various learning algorithms in each field. Although various analysis of open source software has been made, there is a lack of studies to help develop or use deep learning open source software in the industry. This study thus attempts to derive a strategy for adopting the framework through case studies of a deep learning open source framework. Based on the technology-organization-environment (TOE) framework and literature review related to the adoption of open source software, we employed the case study framework that includes technological factors as perceived relative advantage, perceived compatibility, perceived complexity, and perceived trialability, organizational factors as management support and knowledge & expertise, and environmental factors as availability of technology skills and services, and platform long term viability. We conducted a case study analysis of three companies' adoption cases (two cases of success and one case of failure) and revealed that seven out of eight TOE factors and several factors regarding company, team and resource are significant for the adoption of deep learning open source framework. By organizing the case study analysis results, we provided five important success factors for adopting deep learning framework: the knowledge and expertise of developers in the team, hardware (GPU) environment, data enterprise cooperation system, deep learning framework platform, deep learning framework work tool service. In order for an organization to successfully adopt a deep learning open source framework, at the stage of using the framework, first, the hardware (GPU) environment for AI R&D group must support the knowledge and expertise of the developers in the team. Second, it is necessary to support the use of deep learning frameworks by research developers through collecting and managing data inside and outside the company with a data enterprise cooperation system. Third, deep learning research expertise must be supplemented through cooperation with researchers from academic institutions such as universities and research institutes. Satisfying three procedures in the stage of using the deep learning framework, companies will increase the number of deep learning research developers, the ability to use the deep learning framework, and the support of GPU resource. In the proliferation stage of the deep learning framework, fourth, a company makes the deep learning framework platform that improves the research efficiency and effectiveness of the developers, for example, the optimization of the hardware (GPU) environment automatically. Fifth, the deep learning framework tool service team complements the developers' expertise through sharing the information of the external deep learning open source framework community to the in-house community and activating developer retraining and seminars. To implement the identified five success factors, a step-by-step enterprise procedure for adoption of the deep learning framework was proposed: defining the project problem, confirming whether the deep learning methodology is the right method, confirming whether the deep learning framework is the right tool, using the deep learning framework by the enterprise, spreading the framework of the enterprise. The first three steps (i.e. defining the project problem, confirming whether the deep learning methodology is the right method, and confirming whether the deep learning framework is the right tool) are pre-considerations to adopt a deep learning open source framework. After the three pre-considerations steps are clear, next two steps (i.e. using the deep learning framework by the enterprise and spreading the framework of the enterprise) can be processed. In the fourth step, the knowledge and expertise of developers in the team are important in addition to hardware (GPU) environment and data enterprise cooperation system. In final step, five important factors are realized for a successful adoption of the deep learning open source framework. This study provides strategic implications for companies adopting or using deep learning framework according to the needs of each industry and business.