• The Korean Journal of Applied Statistics
• /
• v.15 no.2
• /
• pp.251-267
• /
• 2002

Self Organizing Maps(SOM) by Kohonen is very fast algorithm in neural networks. But it doesn't show sure rules of training results. In this paper, we introduce to Bayesian Learning for Self Organizing Maps(BLSOM) which combines self organizing maps with Bayesian learning. So it supports explanatory power of models and improves prediction. BLSOM has global optima anywhere but SOM has not. This is proved by experiment in this paper.

• Journal of Korean Elementary Science Education
• /
• v.35 no.1
• /
• pp.54-64
• /
• 2016

The purpose of this study is to develop instructional strategy which utilizes thinking maps focused on drawing as a measure to enhance science learning motivation, self-directed learning activity and science academic achievement of learners, and to examine the effects of its application. The target unit for this study is 'life cycle of plants' in the fourth grade of elementary school. Two classes of 4th grades of elementary school were selected and divided into two groups. The learners of experimental group have completed thinking map by drawing a picture to express the results to be observed and measured, and used it to arrange the learning contents. The result of this study is as follows. First, it is proven that using thinking maps focused on drawing actually helped improving the motivation of learners to study science. Second, it is proven that this strategy was effective to change their self-directed learning ability in positive ways. Third, it contributed to the improvement of learners' science academic achievement. We found out that the application of this strategy enabled them to enjoy the mapping using drawing, to be immersed in learning, to better recognize the scientific concepts and the structure of learning contents, and to have a positive awareness of the usefulness of thinking maps focused on drawing.

• Journal of the Korean Society of Earth Science Education
• /
• v.3 no.1
• /
• pp.47-54
• /
• 2010

The purpose of this study is finding examine the Thinking Maps and how to use Thinking Maps effectively in Science Education. The result of this study were as follows: First, There are 8 type Maps, Circle Map, Tree Maps, Bubble Map, Double Bubble Map, Flow Map, Multi Flow Map, Brace Map, Bridge Map. Each Maps are useful in the following activities ; Circle Map-Express their thoughts. Tree Map-Activities as like determine the structure, classification, information organization. Bubble Maps-Construction. Double Bubble Map-Comparison of similarities and differences. Flow Map-Set goals, determine the result of changes in time or place. Multi Flow Map-Analysis cause and effect, expectation and reasoning. Brace Map-Analysis whole and part. Bridge Map-Activities need analogies. Second, each element of inquiry has 1~2 appropriate type of Thinking Maps. So student can choose the desired map. Third, the result of analysing of Science Curriculum Subjects, depending on the subject variety maps can be used. Therefore the Thinking Maps can be used for a variety on activities and subject. And student can be selected according to their learning style. So Thinking Maps are effective to improve student's Self-Directed Learning.

• Journal of the Korean Society of Earth Science Education
• /
• v.11 no.3
• /
• pp.172-181
• /
• 2018

The purpose of this study is to investigate the impact on self-directed learning ability and science process skills by utilizing 'Thinking Maps' in a science class. This particular study was proceeded to 5th grader at B elementary school, there was a mutual agreement with a teacher about assigning a research group and a comparison group and it was agreed by students and explaining the reason and purpose of the study. The researchers visited the school and selected 24 students in the research class and 24 students in the comparative class. For a research group, an experimental group, homeroom teacher, proceeded a science class with the application of 'Thinking Maps'. The experimental period was set up as a 40 minutes class unit for 12 weeks. After an experimental group, self-directed learning ability and science process skills were examined, data collection and data analysis were proceeded by order. The following experimental results are as below. First, the application of 'Thinking Maps' method in the class was effective in self-directed learning ability. Second, the application of 'Thinking Maps' method in the class was effective in scientific process skills. Third, the application of 'Thinking Maps' method in the class had a positive cognition from the learners in the experimental group. Based on the discussions and implications of the results of this study, some suggestions in the follow - up study are as follows. First, applying Thinking Maps technique to various science classes to see the effects can also be suggested as one of the new teaching methods. Second, testing the effects of applying different grades of elementary school students using the Thinking Maps technique could also be highlighted as another way of teaching science classes.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.10a
• /
• pp.241-244
• /
• 1996

There is a growing tendency to consider organizational learning as a mechanism for improving organizations and the rate at which organizations learn becomes perceived as a source for attaining competitive advantage. The objective of this research is to present a two-phase(learning efficient, and learning-effective) organizational modeling methodology based on the cognitive-maps and agents concept, and to describe how the result of the modeling can be used in the organizational learning context.

• Journal of the Korean Mathematical Society
• /
• v.55 no.3
• /
• pp.553-574
• /
• 2018

In this paper, we study harmonic maps and biharmonic maps on the principal G-bundle in Kobayashi and Nomizu and also the warped product $P=M{\times}_fF$ for a $C^{\infty}$(M) function f on M studied by Bishop and O'Neill, and Ejiri.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.4
• /
• pp.31-39
• /
• 2023

In this paper, we propose an Internal/External Knowledge Distillation (IEKD), which utilizes both external correlations between feature maps of heterogeneous models and internal correlations between feature maps of the same model for transferring knowledge from a teacher model to a student model. To achieve this, we transform feature maps into a sequence format and extract new feature maps suitable for knowledge distillation by considering internal and external correlations through a transformer. We can learn both internal and external correlations by distilling the extracted feature maps and improve the accuracy of the student model by utilizing the extracted feature maps with feature matching. To demonstrate the effectiveness of our proposed knowledge distillation method, we achieved 76.23% Top-1 image classification accuracy on the CIFAR-100 dataset with the "ResNet-32×4/VGG-8" teacher and student combination and outperformed the state-of-the-art KD methods.

• Journal of Korea Multimedia Society
• /
• v.19 no.8
• /
• pp.1310-1319
• /
• 2016

Face detection is the first step in a wide range of face applications. However, detecting faces in the wild is still a challenging task due to the wide range of variations in pose, scale, and occlusions. Recently, many deep learning methods have been proposed for face detection. However, further improvements are required in the wild. Another important issue to be considered in the face detection is the computational complexity. Current state-of-the-art deep learning methods require a large number of patches to deal with varying scales and the arbitrary image sizes, which result in an increased computational complexity. To reduce the complexity while achieving better detection accuracy, we propose a fully convolutional network-based face detection that can take arbitrarily-sized input and produce feature maps (heat maps) corresponding to the input image size. To deal with the various face scales, a multi-scale network architecture that utilizes the facial components when learning the feature maps is proposed. On top of it, we design multi-task learning technique to improve detection performance. Extensive experiments have been conducted on the FDDB dataset. The experimental results show that the proposed method outperforms state-of-the-art methods with the accuracy of 82.33% at 517 false alarms, while improving computational efficiency significantly.

• Journal of Semiconductor Engineering
• /
• v.2 no.1
• /
• pp.125-129
• /
• 2021

As the interest of deep learning, techniques to control the color of images in image processing field are evolving together. However, there is no clear standard for color, and it is not easy to find a way to represent only the color itself like the color-palette. In this paper, we propose a novel color palette extraction system by chroma fine-tuning with reinforcement learning. It helps to recognize the color combination to represent an input image. First, we use RGBY images to create feature maps by transferring the backbone network with well-trained model-weight which is verified at super resolution convolutional neural networks. Second, feature maps are trained to 3 fully connected layers for the color-palette generation with a generative adversarial network (GAN). Third, we use the reinforcement learning method which only changes chroma information of the GAN-output by slightly moving each Y component of YCbCr color gamut of pixel values up and down. The proposed method outperforms existing color palette extraction methods as given the accuracy of 0.9140.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.8 no.2
• /
• pp.116-120
• /
• 2008

Clustering is a method for unsupervised learning. General clustering tools have been depended on statistical methods and machine learning algorithms. One of the popular clustering algorithms based on machine learning is the self organizing map(SOM). SOM is a neural networks model for clustering. SOM and extended SOM have been used in diverse classification and clustering fields such as data mining. But, SOM has had a problem determining optimal number of clusters. In this paper, we propose an improvement of SOM using gap statistic and probability distribution. The gap statistic was introduced to estimate the number of clusters in a dataset. We use gap statistic for settling the problem of SOM. Also, in our research, weights of feature nodes are updated by probability distribution. After complete updating according to prior and posterior distributions, the weights of SOM have probability distributions for optima clustering. To verify improved performance of our work, we make experiments compared with other learning algorithms using simulation data sets.