• Journal of Korea Multimedia Society
• /
• v.21 no.2
• /
• pp.173-180
• /
• 2018

This paper proposes a method to improve the performance of face recognition via super-resolution method using sparse representation and deep learning from low-resolution facial images. Recently, there have been many researches on ultra-high-resolution images using deep learning techniques, but studies are still under way in real-time face recognition. In this paper, we combine the sparse representation and deep learning to generate super-resolution images to improve the performance of face recognition. We have also improved the processing speed by designing in parallel structure when applying sparse representation. Finally, experimental results show that the proposed method is superior to conventional methods on various images.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.11
• /
• pp.5427-5445
• /
• 2019

While deep neural networks have achieved remarkable performance in representation learning, a huge amount of labeled training data are usually required by supervised deep models such as convolutional neural networks. In this paper, we propose a new representation learning method, namely generative adversarial networks (GAN) based bagging deep convolutional autoencoders (GAN-BDCAE), which can map data to diverse hierarchical representations in an unsupervised fashion. To boost the size of training data, to train deep model and to aggregate diverse learning machines are the three principal avenues towards increasing the capabilities of representation learning of neural networks. We focus on combining those three techniques. To this aim, we adopt GAN for realistic unlabeled sample generation and bagging deep convolutional autoencoders (BDCAE) for robust feature learning. The proposed method improves the discriminative ability of learned feature embedding for solving subsequent pattern recognition problems. We evaluate our approach on three standard benchmarks and demonstrate the superiority of the proposed method compared to traditional unsupervised learning methods.

• Education of Primary School Mathematics
• /
• v.8 no.1
• /
• pp.23-32
• /
• 2004

It is not too much to say that problem solving is still the focus of school mathematics though the trend of mathematics education for ten year from the one of 1980 is problem solving and the one of mathematics education for ten year from the one of 1990 is standards and constructivism. There are so many crucial clues or methods in good problem solving but I think that one of them is a representation. So, the purpose of this study is to investigate what is the meaning of representation in general and why representation is so important in elementary mathematics learning, Moreover, I have analyzed the gifted children's thinking of representation which is appeared in the previous internet home task of 40 gifted children who are selected through the examination of 1st, 2nd with paper and pencil and 3rd with practical skill and interview and finally I have presented some examples of children's representation how they use representation to model, investigate and understand special concept more easily in elementary school mathematics class.

• Communications of Mathematical Education
• /
• v.29 no.2
• /
• pp.215-239
• /
• 2015

A goal of this study is figuring out how fraction learning centered on various representation activities influences the fraction comprehension and mathematical attitudes. The study focused on 33 4th-grade students of B elementary school in Seoul. In the study, 15 fraction learning classes comprising enactive, iconic, and symbolic representations took place over 6 weeks. After the classes, the ratio of the students who achieved relational understanding increased and the students averagely recorded 90 pt or more on the fraction comprehension test I, II and III. Two-dependent samples t-test was conducted to analyze a significant difference in mathematical attitudes between pre-test and post-test. On the test result, there was the meaningful difference with 0.01 level of significance. To conclude, the fraction learning centered on various representation activities improves students' relational understanding and fraction understanding. In addition, the fraction learning centered on various representation activities gives positive influences on mathematical attitudes since it increases learning orientation, self-control, interests, value cognition, and self-confidence of the students and decreases fears of the students.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.55-61
• /
• 2021

The character motion control based on physics simulation using reinforcement learning continue to being carried out. In order to solve a problem using reinforcement learning, the network structure, hyperparameter, state, action and reward must be properly set according to the problem. In many studies, various combinations of states, action and rewards have been defined and successfully applied to problems. Since there are various combinations in defining state, action and reward, many studies are conducted to analyze the effect of each element to find the optimal combination that improves learning performance. In this work, we analyzed the effect on reinforcement learning performance according to the state representation, which has not been so far. First we defined three coordinate systems: root attached frame, root aligned frame, and projected aligned frame. and then we analyze the effect of state representation by three coordinate systems on reinforcement learning. Second, we analyzed how it affects learning performance when various combinations of joint positions and angles for state.

• Journal of The Korean Association of Information Education
• /
• v.13 no.4
• /
• pp.479-487
• /
• 2009

In the existing curriculum of the Elementary computer Education emphasizes students' problem-solving ability and knowledge of informatics. However, current computer education focus on using application program. In order to raise students' problem-solving ability and logical thinking ability, it is necessary to learning about computer science education. Thereupon, this study applied unplugged learning method to the elementary student. To apply the play-based unplugged learning method to the area of information representation. As a result, unplugged learning method produced higher academic achievement than the lecture model. Also it was more positive in the affective area than the lecture model.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.12
• /
• pp.157-165
• /
• 2011

As the contents of information were reinforced the rules of computer science, it was necessary to develop the new teaching and learning plans and various education materials that encourage students and easy to access. So, in this study developed a wide ran of play activities and educational materials without using a computer for the area of information representation and applied to 68 students, second grade of middle school for 4 weeks. After the class, the effect on the learning achievement was verified by the t-test. As a result, it was shown that there was a significant difference between learning with play activities materials and the traditional lecture-type. It means that learning with play activities materials in the information representation influences a positive effect to the learning achievement.

• The Mathematical Education
• /
• v.45 no.3 s.114
• /
• pp.263-273
• /
• 2006

The purpose of this paper is to research the factors and the effects of immediacy in mathematics teaching and learning and mathematical problem solving. The factors of immediacy are visualization, functional fixedness and representatives. In special, students can apprehend immediately the clues and solution using the visual representation because of its properties of finiteness and concreteness. But the errors sometimes originate from visual representation which come from limitation of the visual representation. It suggests that students have to know conceptual meaning of the visual representation when they use the visual representation. And this phenomenon is the same in functional fixedness and representatives which are the factors of immediacy The methods which overcome the errors of immediacy is that problem solvers notice the limitation of the factors of immediacy and develop the meta-cognitive ability. And it means we have to emphasize the logic and the intuition in mathematical teaching and learning. Clearly, we can't solve all mathematical problems using only either the logic or the intuition.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.842-846
• /
• 1993

Because of complexity, neural network is difficult to learn. So if internal representation[1] can be performed successfully, it is possible to use perceptron learning rule. As a result, learning is easier. Therefore the method of internal representations applied to the "XOR" problem, and the "spirals" problem. And then using the above results, the structure of neural network for computing is embodied.mputing is embodied.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.11 no.3
• /
• pp.135-142
• /
• 2011

Unlike using the sequence-based representation for a chromosome in previous genetic algorithms for Bayesian structure learning, we proposed a matrix representation-based genetic algorithm. Since a good chromosome representation helps us to develop efficient genetic operators that maintain a functional link between parents and their offspring, we represent a chromosome as a matrix that is a general and intuitive data structure for a directed acyclic graph(DAG), Bayesian network structure. This matrix-based genetic algorithm enables us to develop genetic operators more efficient for structuring Bayesian network: a probability matrix and a transpose-based mutation operator to inherit a structure with the correct edge direction and enhance the diversity of the offspring. To show the outstanding performance of the proposed method, we analyzed the performance between two well-known genetic algorithms and the proposed method using two Bayesian network scoring measures.