• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.6
• /
• pp.847-856
• /
• 1995

When a human recognizes nouns in a sentence, s/he associates them with the hyper concepts of onus. For computer to simulate the human's word recognition, it should build the knowledge base (WordNet)for the hyper concepts of words. Until now, works for the WordNet haven't been performed in Korea, because they need lots of human efforts and time. But, as the power of computer is radically improved and common MRD becomes available, it is more feasible to automatically construct the WordNet. This paper proposes the method that automatically builds the WordNet of Korean nouns by using the descripti on of onus in Korean MRD, and it proposes the rules for extracting the hyper concepts (hypernyms)by analyzing structrual characteristics of Korean. The rules effect such characteristics as a headword lies on the rear part of sentences and the descriptive sentences of nouns have special structure. In addition, the WordNet prototype of Korean Nouns is developed, which is made by combining the hypernyms produced by the rules mentioned above. It extracts the hypernyms of about 2,500 sample words, and the result shows that about 92per cents of hypernyms are correct.

• Journal of Digital Convergence
• /
• v.14 no.5
• /
• pp.325-331
• /
• 2016

In driving at evening or night, we are not able to recognize the speed bump and so stop suddenly. It could result in accidents. And also, we have a restriction of street light installation in farm road because it could be harmful to the crops and driver could not recognize the walking people. It needs to develop the speed bump with light and streetlight to be non harmful to the crops. So, we develop both the speed bump and streetlight with LED which could be non harmful to the crops and be increased recognition of walking people in farm road. For LED lighting power, we use the solar cells, and piezoelectric elements. It has automatic on/off according to power saving rates without illumination sensor. Minimization of circuit elements and design of minimum resisters and low power LED was used for power saving in assuring 3-days.

• Journal of KIISE
• /
• v.43 no.3
• /
• pp.370-379
• /
• 2016

Recently personal media were produced in a variety of ways as a lot of smart devices have been spread and services using these data have been desired. Therefore, research has been actively conducted for the media analysis and recognition technology and we can recognize the meaningful object from the media. The system using the media ontology has the disadvantage that can't classify the media appearing in the video because of the use of a video title, tags, and script information. In this paper, we propose a system to automatically classify video using the objects shown in the media data. To do this, we use a description logic-based reasoning and a rule-based inference for event processing which may vary in order. Description logic-based reasoning system proposed in this paper represents the relation of the objects in the media as activity ontology. We describe how to another rule-based reasoning system defines an event according to the order of the inference activity and order based reasoning system automatically classify the appropriate event to the category. To evaluate the efficiency of the proposed approach, we conducted an experiment using the media data classified as a valid category by the analysis of the Youtube video.

• Korean Journal of Remote Sensing
• /
• v.27 no.1
• /
• pp.33-41
• /
• 2011

In the information communication technology, it is world-widely apparent that trend movement from internet web to smartphone app by users demand and developers environment. So it needs kinds of appropriate technological responses from geo-spatial domain regarding this trend. However, most cases in the smartphone app are the map service and location recognition service, and uses of geo-spatial contents are somewhat on the limited level or on the prototype developing stage. In this study, app for extraction of corner point features using geo-spatial imagery and their linkage to database system are developed. Corner extraction is based on Harris algorithm, and all processing modules in database server, application server, and client interface composing app are designed and implemented based on open source. Extracted corner points are applied LOD(Level of Details) process to optimize on display panel. Additional useful function is provided that geo-spatial imagery can be superimposed with the digital map in the same area. It is expected that this app can be utilized to automatic establishment of POI (Point of Interests) or point-based land change detection purposes.

• Journal of Soil and Groundwater Environment
• /
• v.10 no.1
• /
• pp.65-74
• /
• 2005

National groundwater monitoring stations have been managed throughout the country by Korea Water Resources Corporation (KOWACO) in order to monitor variations in quantity and quality of groundwater resources. A multi-sensor installed in each monitoring station well measures groundwater level, water temperature and electrical conductivity every six hours and the logged data are automatically transmitted to a host computer in KOWACO. Meanwhile despite regular station inspection and replacement of deteriorate or broken devices, abnormal values or outliers often occur due to intrinsic limitations of automatic monitoring and transmission. Thus prompt recognition and measures to these values are essentially required to reduce disturbance and missing period of the data. In this study, time and frequency of outlier occurrence were analyzed for the water level data obtained from national groundwater monitoring stations within the Han river basin in 2000. The analysis results indicated that the most prominent patterns of the outliers were rapid decline for water level, no variation for temperature and steep decline for electrical conductivity. This study provided a sample criterion for determining the outlier for each parameter.

• Journal of Korean Library and Information Science Society
• /
• v.51 no.2
• /
• pp.355-379
• /
• 2020

Digital Library System(DLS) for the school library is a key access tool for school library materials. The purpose of this study was to find ways to improve the accessibility of materials through analysis of students' information search behavior in DLS. Data were collected through recording of 42 participants' DLS search process, and questionnaire. As a result, the search success rate and search satisfaction were found to be lower when the main purpose of DLS is simple leisure reading, information needs are relatively ambiguous, and when user experiences the complicated situations in the search process. The satisfaction level of search time sufficiency was the highest, and the search result satisfaction was the lowest. Besides, there was a need to improve DLS, such as integrated search of other library collection information, the recommendation of related materials, the print output of collection location, voice recognition through mobile apps, and automatic correction of search errors. Through this, the following can be suggested. First, DLS should complement the function of providing career information by reflecting the demand of education consumers. Second, improvements to DLS functionality to the general information retrieval system level must be made. Third, an infrastructure must be established for close cooperation between school library field personnel and DLS management authorities.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.3
• /
• pp.2123-2130
• /
• 2015

For such automatic monitoring robot or a robot cleaner that is utilized indoors, if it deviates from someone by replacement or, or of a mobile robot such as collisions with unexpected object direction or planned path, based on the planned path There is a need to come back to, it is necessary to tough self-position estimation ability of mobile robot in this, which is also associated with resolution of the kidnap problem of conventional mobile robot. In this study, the case of a mobile robot, operates indoors, you want to take advantage of the low cost of the robot. Therefore, in this paper, by using the acquisition device to an external image information such as the CCTV which is installed in a room, it acquires the environment image and take advantage of marker recognition of the mobile robot at the same time and converted it absolutely spatial coordinates it is, we are trying to solve the self-position estimation of the mobile robot in the room and kidnap problem and actual implementation methods potential field to try utilizing robotic systems. Thus, by implementing the method proposed in this study to the actual robot system, and is promoting the relevant experiment was to verify the results.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11c
• /
• pp.2265-2268
• /
• 2002

인터넷 사용자들의 기하급수적인 증가와 웹 페이지의 폭발적인 증가로 인하여 정보공유를 위한 인터넷에서 효율적으로 원하는 정보를 발견하고 이용하기에 매우 힘든 상황에 처해있다. 따라서, 이러한 문제점들을 근본적으로 해결하기 위한 노력의 일환으로 기계가 이해하고 추론할 수 있는 시맨틱 웹이 등장하였다. 시맨틱 웹과 관련된 여러 가지 기술들 중 시맨틱 웹 서비스는 사용자에게 현재의 인터넷 환경에서 제공할 수 있는 서비스보다 향상된 고품질의 서비스를 제공하는 것을 목적으로 삼고 있다. 이러한 시맨틱 웹 서비스는 웹 서비스의 발견, 실행 및 구성으로 구성된다. 본 논문에서는 시맨틱 웹 서비스를 자동화하기 위한 노력의 일환으로서 시맨틱 웹에서 웹 서비스를 구성하는 문제에 대하여 언급한다. 시맨틱 웹 서비스 구성문제는 사용자의 요구사항을 충족시키기 위하여 다양한 웹 서비스들을 조합하는 문제이다. 그러나, WSFL, X-LANG 및 BPEL4WS 그리고, DAML-S와 같은 웹 서비스 구성문제에 대한 일련의 노력들에서는 사용자 요구사항에 대한 검증이나 서비스의 품질에 대한 사항들을 확인 및 제공할 수 있는 방법들이 없다. 따라서, 본 논문에서는 시맨틱 웹 서비스 구성문제와 관련된 상기와 같은 문제점들을 해결할 수 있는 방법론을 제시한다. 본 논문에서 제시된 방법론에서는 시맨틱 웹 구성문제를 제약만족문제로 변환함으로써 제약만족문제에 있어 늘리 알려진 다양한 알고리즘들을 활용할 수 있는 장점들이 있을 뿐만 아니라 사용자들의 요구사항에 대한 검증과 서비스의 품질을 확인할 수 있는 장점들이 있다.의 위상변화에 대한 적응성을 높일 수 있도록 한다. SQL Server 2000 그리고 LSF를 이용하였다. 그리고 구현 환경과 구성요소에 대한 수행 화면을 보였다.ool)을 사용하더라도 단순 다중 쓰레드 모델보다 더 많은 수의 클라이언트를 수용할 수 있는 장점이 있다. 이러한 결과를 바탕으로 본 연구팀에서 수행중인 MoIM-Messge서버의 네트워크 모듈로 다중 쓰레드 소켓폴링 모델을 적용하였다.n rate compared with conventional face recognition algorithms. 아니라 실내에서도 발생하고 있었다. 정량한 8개 화합물 각각과 총 휘발성 유기화합물의 스피어만 상관계수는 벤젠을 제외하고는 모두 유의하였다. 이중 톨루엔과 크실렌은 총 휘발성 유기화합물과 좋은 상관성 (톨루엔 0.76, 크실렌, 0.87)을 나타내었다. 이 연구는 톨루엔과 크실렌이 총 휘발성 유기화합물의 좋은 지표를 사용될 있고, 톨루엔, 에틸벤젠, 크실렌 등 많은 휘발성 유기화합물의 발생원은 실외뿐 아니라 실내에도 있음을 나타내고 있다.>10)의 $[^{18}F]F_2$를 얻었다. 결론: $^{18}O(p,n)^{18}F$ 핵반응을 이용하여 친전자성 방사성동위원소 $[^{18}F]F_2$를 생산하였다. 표적 챔버는 알루미늄으로 제작하였으며 본 연구에서 연구된 $[^{18}F]F_2$가스는 친핵성 치환반응으로 방사성동위원소를 도입하기 어려운 다양한 방사성의 약품개발에 유용하게 이용될 수 있을 것이다.었으나

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.12
• /
• pp.507-516
• /
• 2019

Costs invested in road maintenance and road development are on the rise. However, due to accidents such as portholes and ground subsidence, the risks to the drivers' safety and the material damage caused by accidents are also increasing. Following this trend, we have developed a system that determines road damage, according to the magnitude of vibration generated without directly intervening the driver when driving. In this paper, we implemented the system using a remote control car (RC car) simulator due to the limitation of the environment in which the actual vehicle is not available in the process of developing the system. In addition, we attached a vibration sensor and GPS sensor to the body of the RC car simulator to measure the vibration value and location information generated by the movement of the vehicle in real-time while driving, and transmitting the corresponding data to the server. In this way, we implemented a system that allows external users to check the damage of roads and the maintenance of the repaired roads based on data more easily than the existing systems. By using this system, we can perform early prediction of road breakage and pattern prediction based on the data. Further, for the RC car simulator, commercialization will be possible by combining it with business in other fields that require flatness.