• ETRI Journal
• /
• 제22권3호
• /
• pp.10-18
• /
• 2000

Although an n-th order cross-entropy (nCE) error function resolves the incorrect saturation problem of conventional error backpropagation (EBP) algorithm, performance of multilayer perceptrons (MLPs) trained using the nCE function depends heavily on the order of nCE. In this paper, we propose an adaptive learning rate to markedly reduce the sensitivity of MLP performance to the order of nCE. Additionally, we propose to limit error signal values at out-put nodes for stable learning with the adaptive learning rate. Through simulations of handwritten digit recognition and isolated-word recognition tasks, it was verified that the proposed method successfully reduced the performance dependency of MLPs on the nCE order while maintaining advantages of the nCE function.

• 한국환경교육학회지:환경교육
• /
• 제12권1호
• /
• pp.365-377
• /
• 1999

The concept of environmental education contains from teaching environmental pollutions and the importance of environment to the ecological relationship between human and the nature. The ultimate goal of environmental eduaction, however, is to build the environmental-friendly and responsible behavoir. One of the best way to achieve this goal is ecological learning centers, where students can observe and analyse threes, plants and animals, and they learn the principle of the environmental succession with feeling and understanding. Students internalize environmental awareness through experiencing the nature. In this paper, we would like to introduce the diverse types of ecological learning centers in germany and their programs focusing on the ecocenters in berlin, with an intention of adopting practical programs in korean school system.

• 한국정보기술학회 영문논문지
• /
• 제10권2호
• /
• pp.61-72
• /
• 2020

Recently, researches that are using deep learning technology in various fields are being conducted. The fields include geographic map processing. In this paper, I propose a method to infer where the map area included in the image is. The proposed method generates and learns images including a map, detects map areas from input images, extracts character strings belonging to those map areas, and converts the extracted character strings into coordinates through geocoding to infer the coordinates of the input image. Faster R-CNN was used for learning and map detection. In the experiment, the difference between the center coordinate of the map on the test image and the center coordinate of the detected map is calculated. The median value of the results of the experiment is 0.00158 for longitude and 0.00090 for latitude. In terms of distance, the difference is 141m in the east-west direction and 100m in the north-south direction.

• 한국군사과학기술학회지
• /
• 제27권3호
• /
• pp.346-354
• /
• 2024

Deep learning models require high computing power due to a substantial amount of computation. It is difficult to use them in devices with limited computing environments, such as coastal surveillance equipments. In this study, a lightweight model is constructed by analyzing the weight changes of the convolutional layers during the training process based on MobileNet and then pruning the layers that affects the model less. The performance comparison results show that the lightweight model maintains performance while reducing computational load, parameters, model size, and data processing speed. As a result of this study, an effective pruning method for constructing lightweight deep learning models and the possibility of using equipment resources efficiently through lightweight models in limited computing environments such as coastal surveillance equipments are presented.

• 한국멀티미디어학회논문지
• /
• 제25권5호
• /
• pp.757-768
• /
• 2022

In this paper, we proposed a system that visualizes a hologram device in 3D by utilizing the CT image segmentation function based on artificial intelligence deep learning. The input axial CT medical image is converted into Sagittal and Coronal, and the input image and the converted image are divided into 3D volumes using ResUNet, a deep learning model. In addition, the volume is created by segmenting the tumor region in the segmented liver image. Each result is integrated into one 3D volume, displayed in a medical image viewer, and converted into a video. When the converted video is transmitted to the hologram device and output from the device, a 3D image with a sense of space can be checked. As for the performance of the deep learning model, in Axial, the basic input image, DSC showed 95.0% performance in liver region segmentation and 67.5% in liver tumor region segmentation. If the system is applied to a real-world care environment, additional physical contact is not required, making it safer for patients to explain changes before and after surgery more easily. In addition, it will provide medical staff with information on liver and liver tumors necessary for treatment or surgery in a three-dimensional manner, and help patients manage them after surgery by comparing and observing the liver before and after liver resection.

• 한국전문물리치료학회지
• /
• 제7권3호
• /
• pp.81-89
• /
• 2000

The purpose of this study was to determine the effect of electrical stimulation biofeedback on motor learning of quadriceps muscle isometric exercise in 3 patients who have undergone total knee replacement surgery. A multiple baseline design across subjects was used. The electrical stimulation biofeedback was provided with each patient during quadriceps isometric exercise, which last 10 to 14 sessions with 10 repetitions each sessions. After training patients received 4 retention tests. Maximum muscle activity was measured pre- and post- electrical stimulation biofeedback training and retention test to evaluate the effect of biofeedback training. Maximum isometric muscle activity of quadriceps was increased after electrical stimulation biofeedback training in all subjects. The results indicate that a electrical stimulation biofeedback training is a useful method to improve motor learning of quadriceps isometric exercise in total knee replacement.

• 대한간호학회지
• /
• 제35권6호
• /
• pp.1036-1043
• /
• 2005

Purpose: The purpose of this study was to develop a vital signs e-book for undergraduate student nurses and evaluate the content, system and student satisfaction. Method: This study was done in three stages, the development of a vital signs e-book, implementation and evaluation. The subjects were 73 undergraduate student nurses in Y university. Result: Thirty one learning objectives were used to create the contents. A set of 5 chapters and 18 subsections were defined after validation from nurse educators. The e-book is available at http://123.134.207.23/ebook/vitalsigns. Analysis of the questionnaires showed a mean score for content, system and students satisfaction of $3.17\pm73,\;3.11\pm79,\;and\;2.96\pm.74$ respectively out of a possible 4 points. Conclusion: Nurse educators should provide quality and effective web-based courses that meet undergraduate student nurses' learning needs and they should incorporate web-based learning into traditional teaching to meet the demands of nursing education.

• ETRI Journal
• /
• 제28권1호
• /
• pp.59-66
• /
• 2006

Learning behaviors of a radial basis function network (RBFN) using a singular value decomposition (SVD) and stochastic gradient (SG) algorithm, together named RBF-SVD-SG, for odor sensing systems are analyzed, and a fast training method is proposed. RBF input data is from a conducting polymer sensor array. It is revealed in this paper that the SG algorithm for the fine-tuning of centers and widths still shows ill-behaving learning results when a sufficiently small convergence coefficient is not used. Since the tuning of centers in RBFN plays a dominant role in the performance of RBFN odor sensing systems, our analysis is focused on the center-gradient variance of the RBFN-SVD-SG algorithm. We found analytically that the steadystate weight fluctuation and large values of a convergence coefficient can lead to an increase in variance of the center-gradient estimate. Based on this analysis, we propose to use the least mean square algorithm instead of SVD in adjusting the weight for stable steady-state weight behavior. Experimental results of the proposed algorithm have shown faster learning speed and better classification performance.

• 한국IT서비스학회지
• /
• 제14권1호
• /
• pp.115-128
• /
• 2015

As the life expectancy of human beings gets longer and our society changes into highly competitive arena, the implementation of online adult learning is growing, and therefore the learners in self-regulated scaffolding learning environments is becoming an important topic. This study is to investigate the main effects of scaffolding and self-regulation and the interaction effect on discussion participation and comprehension in online learning environments. To do this, ninety-nine adults taking online learning courses with the open university in Korea were investigated. Adult learners were divided into one of the four groups (no scaffolding, conceptual, strategic, and conceptual and strategic scaffoldings). Regarding self-regulation, learners were divided into two groups (low and high self-regulated) based on the mean score of subjective report of self-regulated learning. The results are as follows : First, 'strategic scaffolding' is more effective than 'conceptual scaffolding' in discussion participation (F=2.772, p < .05) and comprehension test (F=7.156, p < .05). Second, high self-regulated learners more actively participate than low self-regulated learners in discussion (F=6.230, p < .05), and achieve higher scores (F=4.863, p < .05). Third, there is no interaction effect between scaffolding strategies and the level of self-regulation. The theoretical and practical implications of these findings are discussed.

• 공학교육연구
• /
• 제25권2호
• /
• pp.22-31
• /
• 2022

This study was conducted to investigate the effect of learning style and lecture attitude on academic achievement in non-face-to-face lectures. Among students attending H University in the first semester of 2020. 1,880 students voluntarily participated in the survey. The questionnaire consists of 33 questions, and actual GPA was used to determine academic achievement. In this study, ANOVA and regression were utilized to find out the factors affecting academic achievement. The results are as follows. First, in the correlation between learning style types, the higher the avoidant learning style, the lower the independent, cooperative, and competitive types. Second, the higher the independent learning type, the higher the positive lecture attitude and GPA, and the higher the type of cooperative learning, the higher the negative lecture attitude. Third, the positive attitudes of 4th graders were higher than that of 1st graders. Fourth, there was no difference in learning style between colleges. Fifth, independent/competitive learning styles and non-face-to-face lecture attitudes had a significant effect on academic achievement in non-face-to-face lectures. In conclusion, in order to narrow the academic gap, instructors must identify the type of learning style. And it is necessary to consider students' preferences for non-face-to-face lectures.