• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.162-162
• /
• 2017

구글에서 공개한 Tensorflow를 이용한 여러 학문 분야의 연구가 활발하다. 농업 시설환경을 대상으로 한 빅데이터의 축적이 증가함과 아울러 실효적인 정보 획득을 위한 각종 데이터 분석 및 마이닝 기법에 대한 연구 또한 활발한 상황이다. 한편, 타 분야의 성공적인 심층학습기법 응용사례에 비하여 농업 분야에서의 응용은 초기 성장 단계라 할 수 있다. 이는 농업 현장에서 취득한 정보의 난해성 및 완성도 높은 생육/환경 모델링 정보의 부재로 실효적인 전과정 처리 기술 도출에 소요되는 시간, 비용, 연구 환경이 상대적으로 부족하기 때문일 것이다. 특히, 센서 기반 데이터 취득 기술 증가에 따라 비약적으로 방대해진 수집 데이터를 시간 복잡도가 높은 심층 학습 모델링 연산에 기계적으로 단순 적용할 경우 시간 효율적인 측면에서 성공적인 결과 도출에 애로가 있을 것이다. 매우 높은 시간 복잡도를 해결하기 위하여 제시된 하드웨어 가속 기능의 경우 일부 개발환경에 국한이 되어 있다. 일례로, 구글의 Tensorflow는 오픈소스 기반 병렬 클러스터링 기술인 MPICH를 지원하는 알고리즘을 공개하지 않고 있다. 따라서, 본 연구에서는 심층학습 기법 연구에 있어서, 예상 가능한 다양한 자원을 활용하여 최대한 연산의 결과를 빨리 도출할 수 있는 하드웨어적인 접근 방법을 모색하였다. 호스트에서 수행하는 일방적인 학습 알고리즘과 달리 이기종간 심층 학습이 가능하기 위해선 우선, NFS(Network File System)를 이용하여 데이터 계층이 상호 연결이 되어야 한다. 이를 위해서 고속 네트워크를 기반으로 한 NFS의 이용이 필수적이다. 둘째로 제한된 자원의 한계를 극복하기 위한 메모 공유 라이브러리가 필요하다. 셋째로 이기종간 프로세서에 최적화된 병렬 처리용 컴파일러를 이용해야 한다. 가장 중요한 부분은 이기종간의 처리 능력에 따른 작업을 고르게 분배할 수 있는 작업 스케쥴링이 수행되어야 하며, 이는 처리하고자 하는 데이터의 형태에 따라 매우 가변적이므로 해당 데이터 도메인에 대한 엄밀한 사전 벤치마킹이 수행되어야 한다. 이러한 요구조건을 대부분 충족하는 Open-CL ver1.2(https://www.khronos.org/opencl/)를 이용하였다. 최신의 Open-CL 버전은 2.2이나 본 연구를 위하여 준비한 4가지 이기종 시스템에서 모두 공통적으로 지원하는 버전은 1.2이다. 실험적으로 선정된 4가지 이기종 시스템은 1) Windows 10 Pro, 2) Linux-Ubuntu 16.04.4 LTS-x86_64, 3) MAC OS X 10.11 4) Linux-Ubuntu 16.04.4 LTS-ARM Cortext-A15 이다. 비교 분석을 위하여 NVIDIA 사에서 제공하는 Pascal Titan X 2식을 SLI로 구성한 시스템을 준비하였다. 개별 시스템에서 별도로 컴파일 된 바이너리의 이름을 통일하고, 개별 시스템의 코어수를 동일하게 균등 배분하여 100 Hz의 데이터로 입력이 되는 온도 정보와 조도 정보를 입력으로 하고 이를 습도정보에 Linear Gradient Descent Optimizer를 이용하여 Epoch 10,000회의 학습을 수행하였다. 4종의 이기종에서 총 32개의 코어를 이용한 학습에서 17초 내외로 연산 수행을 마쳤으나, 비교 시스템에서는 11초 내외로 연산을 마치는 결과가 나왔다. 기보유 하드웨어의 적절한 활용이 가능한 심층학습 기법에 대한 연구를 지속할 것이다

• Journal of Science Education
• /
• v.38 no.3
• /
• pp.569-584
• /
• 2014

The Purpose of this study was an analysis of high school student's understanding level about concepts of special relativity through in-depth interview. The 8 participants were 10th grade students in H high school in Ulsan city, who were interviewed and analyzed in the results of the interview about basic concepts of special relativity using achievement checklist in 6 situations(principle of constancy of light velocity, principle of relativity, relativity of simultaneity, garage paradox, rocket paradox). As results of the checklist, the participants showed high achievement in the content level of simple phenomena and simple concepts related to special relativity. But they showed low achievement in the concept level for fundamental understanding of special relativity. As results of the interview, it was found that the participants decided the order of events depending on their intuition and had a difficulty to apply the coordinate system to real situation, even though they mathematically understood it. In addition, some participants who could not understand the inertial coordinate system explained paradoxes of relativity depending on their intuition and had learner's chaos. Finally, though high school students usually being in formal operational stage, some students had difficulty to draw phenomena of space and time in two dimensional plane.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.5
• /
• pp.593-600
• /
• 2019

The similarity in tones between speakers can lower the performance of speaker verification. To improve the performance of speaker verification systems, we propose a multi-task learning technique using deep neural network to learn speaker information and age information. Multi-task learning can improve generalization performances, because it helps deep neural networks to prevent hidden layers from overfitting into one task. However, we found in experiments that learning of age information does not work well in the process of learning the deep neural network. In order to improve the learning, we propose a method to dynamically change the objective function weights of speaker identification and age estimation in the learning process. Results show the equal error rate based on RSR2015 evaluation data set, 6.91 % for the speaker verification system without using age information, 6.77 % using age information only, and 4.73 % using age information when weight change technique was applied.

• Journal of the Korean Society for Marine Environment & Energy
• /
• v.15 no.1
• /
• pp.9-18
• /
• 2012

Annual power consumption of our country is positioned in the upper percentile in the world, and because the proportion of fossil power generation is high, which ranks the 10th $CO_2$ emission country. In this regard, government has established and is implementing the National Energy Basic Plan to realize to get out of fossilization in energy supply while focusing on securing the technology for renewable energy as well as its commercialization in order to reduce greenhouse gas. Resource recovery technology for deep seawater thermal energy which is one of renewable energies is newly getting attention domestically as well as in overseas for securing resources and environmental improvement as a core technology for multilateral use of marine resources for low carbon and green growth. Economic feasibility analysis was conducted for the research and development as follows on the use of ocean thermal energy conversion and seawater air conditioning. First, in the case of power generation using deep seawater and warm discharge water from ocean thermal energy conversion plant of 1MW level, it is judged that the economic feasibility is insufficient but the feasibility will be significantly improved if we consider not only power generation but also drinking water and certified emission reduction by developing the power plant to the size for commercialization. Second, the economic feasibility for the use of deep seawater as air conditioning for the power plant of 1,000RT level turned out to be very good. Especially, when we consider certified emission reduction, it will be possible to secure sufficient economic feasibility. When we use it in connection with ocean thermal energy conversion, water conversion and agricultural and fishery use, it is judged that economic ripple effect will be significant and therefore it will be necessary to conduct research and development for early commercialization, distribution and diffusion of deep seawater energy.

• Journal of the Korea Convergence Society
• /
• v.13 no.4
• /
• pp.643-651
• /
• 2022

The mobile augmented reality(MAR) market is rapidly growing with following the sudden rise of the metaverse. However, research on improving user experience in mobile augmented reality environment is lacking. This study aims to develop the usability principles of an AR-based smartphone app. The research conducts following methods. 1)The characteristics of MAR were understood to identify usability elements of AR apps. 2)Usability principles were collected through analysis of case studies on the usability of AR and the 14 usability principles were developed through in-depth interviews with experts for the purpose of chunking. 3)To verify the 14 developed usability principles, evaluation items were composed based on the usability principles and usability evaluation survey was conducted. Consequently, significant differences were found in immersivity, presence, efficiency, satisfaction, playfulness and It was verified that the developed usability principle is useful in developing AR based smartphone apps. This study is meaningful in that it conducted the basic research of mobile AR usability principles

• Korean Journal of Childcare and Education
• /
• v.9 no.2
• /
• pp.263-283
• /
• 2013

This study analyzed picture book "Who Pooped on my Head?" by using semiotic theories and considered involved meaning of the text. Because characteristics of children's development and picture book should be in harmony, it is necessary to understand the text in-depth. As a result of analysis, the text helps psychological stability and contains the meaning and value that guides children into social value or norm. Analyzing the text by using Greimas' Semiotic square, the meaning that following bowel movement standard is socially established behavior was involved. It seems that giving up instinct and accepting standard are process to extend thought. In conclusion, this study confirmed that the meaning and value that support cognitive and emotional development was contained in the picture book, "Who Pooped on my Head?".

• The Journal of the Korea Contents Association
• /
• v.13 no.11
• /
• pp.608-616
• /
• 2013

Debate on the introduction of musical 'stage quota' has been triggered to protect domestic original musical market. In the in-depth interviews with experts, most of the interviewees were in favor of the introduction of this system, ensuring mainly cultural diversity. The proposed model by proponents is that large-scaled national and public theaters(1,000 seats or more) must perform original musicals at least from 20% to 30% of total musical performance days per year set by law until its competitiveness should be ensured. Opponents proposed that government should design musical promotion and funding policy. The study implies that stage quota should carried out since the system is essential, and should be institutionalized through legislation. It is also suggested that government should consider and consult the stakeholders to draft this system reflecting answerers' limitations such as target, quota rate, period to minimize side effects like shrinkage of musical industry.

• Journal of Digital Convergence
• /
• v.18 no.2
• /
• pp.427-437
• /
• 2020

This article researched the changes of terrestrial local broadcasting companies in a crisis of management. According to the following: First, the structure of the internal organization of local broadcasting companies has been changed to put more emphasis on 'management'. Second, most of the interviewees responded that this change was an inevitable choice. However, some responded that their 'role identity' was changing in a negative direction. Third, these changes have led to the emergence of programs that could undermine public interest. It also showed an increase in the production of 'funding-based' programs and 'mobile first' programs.

• Journal of Internet Computing and Services
• /
• v.19 no.4
• /
• pp.45-52
• /
• 2018

A single-image super-resolution is a process of restoring a high-resolution image from a low-resolution image. Recently, the super-resolution using the deep neural network has shown good results. In this paper, we propose a neural network structure that improves speed and performance over conventional neural network based super-resolution methods. To do this, we analyze the conventional neural network based super-resolution methods and propose solutions. The proposed method reduce the 5 stages of the conventional method to 3 stages. Then we have studied the optimal width and depth by experimenting on the width and depth of the network. Experimental results have shown that the proposed method improves the disadvantages of the conventional methods. The proposed neural network structure showed superior performance and speed than the conventional method.

• The Journal of the Korea Contents Association
• /
• v.22 no.3
• /
• pp.285-295
• /
• 2022

The purpose of this study is to prepare primary data for providing smooth and systematic rehabilitation sports services by confirming the opinions of instructors experienced in the field of disability sports before the implementation of Article 15 'Rehabilitation sports' of the 'Act on Guarantee of Right to Health and Access to Health Services for Persons with Disabilities. In-depth interviews were conducted as a research method, and qualitative analysis was conducted on the contents of the interview. The in-depth interview is unstructured, allowing disability sports instructor to freely present their opinions on difficulties experienced while operating the program, necessary matters for rehabilitation sports implementation. We transcribed the recording data of the in-depth interview and coded the opinions through the thematic analysis method among qualitative research methods. As a result, the final 104 opinions were used and classified into 9 categories; rehabilitation sports goal, training rehabilitation sports instructor, evaluation items and educational guidelines, relationships with program participants, facilities utilization, effects, the timing of provision, and doctor roles.