• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.25 no.4
• /
• pp.162-172
• /
• 2021

This paper proposes stochastic methods to find an approximate solution for the L²-Wasserstein least squares problem of Gaussian measures. The variable for the problem is in a set of positive definite matrices. The first proposed stochastic method is a type of classical stochastic gradient methods combined with projection and the second one is a type of variance reduced methods with projection. Their global convergence are analyzed by using the framework of proximal stochastic gradient methods. The convergence of the classical stochastic gradient method combined with projection is established by using diminishing learning rate rule in which the learning rate decreases as the epoch increases but that of the variance reduced method with projection can be established by using constant learning rate. The numerical results show that the present algorithms with a proper learning rate outperforms a gradient projection method.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.8
• /
• pp.39-47
• /
• 2011

Ergezinger's method, one of the layer-by-layer algorithms used for multilyer perceptrons, consists of an output node and can make premature saturations in the output's weight because of using linear least squared method in the output layer. These saturations are obstacles to learning time and covergence. Therefore, this paper expands Ergezinger's method to be able to use an output vector instead of an output node and introduces a learning rate to improve learning time and convergence. The learning rate is a variable rate that reflects the correlation coefficient between new weight and previous weight while updating hidden's weight. To compare the proposed method with Ergezinger's method, we tested iris recognition and nonlinear approximation. It was found that the proposed method showed better results than Ergezinger's method in learning convergence. In the CPU time considering correlation coefficient computation, the proposed method saved about 35% time than the previous method.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.195-206
• /
• 2022

Crack detection is essential for inspection of existing structures and crack segmentation based on deep learning is a significant solution. However, datasets are usually one of the key issues. When building a new dataset for deep learning, laborious and time-consuming annotation of a large number of crack images is an obstacle. The aim of this study is to develop an approach that can automatically select a small portion of the most informative crack images from a large pool in order to annotate them, not to label all crack images. An active learning method with difficulty learning mechanism for crack segmentation tasks is proposed. Experiments are carried out on a crack image dataset of a steel box girder, which contains 500 images of 320×320 size for training, 100 for validation, and 190 for testing. In active learning experiments, the 500 images for training are acted as unlabeled image. The acquisition function in our method is compared with traditional acquisition functions, i.e., Query-By-Committee (QBC), Entropy, and Core-set. Further, comparisons are made on four common segmentation networks: U-Net, DeepLabV3, Feature Pyramid Network (FPN), and PSPNet. The results show that when training occurs with 200 (40%) of the most informative crack images that are selected by our method, the four segmentation networks can achieve 92%-95% of the obtained performance when training takes place with 500 (100%) crack images. The acquisition function in our method shows more accurate measurements of informativeness for unlabeled crack images compared to the four traditional acquisition functions at most active learning stages. Our method can select the most informative images for annotation from many unlabeled crack images automatically and accurately. Additionally, the dataset built after selecting 40% of all crack images can support crack segmentation networks that perform more than 92% when all the images are used.

• The Journal of Korean Association of Computer Education
• /
• v.8 no.1
• /
• pp.55-63
• /
• 2005

Recently, web-based E-learning systems have entered the spotlight by providing new learning environments that break down spatial and temporal limitations. The key to building the web-based E-learning system is in determining how to effectively use the system and to evaluate the degree of learning achieved by the students that use it. In traditional off-line learning systems, we can evaluate students by counting how many questions, designed to evaluate their learning achievement, he or she answers correctly within a predetermined time limit. But this method would make individualized learning, a strong point of E-learning systems, impossible because these systems provide same learning strategy to all students even though they achieve a different level of learning. Therefore, in this paper, I will find any relationships between given test answers using fuzzy implication theory, I call these fuzzy correlations, and then generate evaluation results that are reflected in those relationships. I will compare the differences between this evaluation method and a traditional evaluation method where a student takes a test to evaluate his or her learning achievement after some learning period. Finally, I will discuss how we can use these results in individualized learning.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10a
• /
• pp.91-94
• /
• 1996

This paper presents an approach to estimation of learning gain in iterative learning control for discrete-time affine nonlinear systems. In iterative learning control, to determine learning gain satisfying the convergence condition, we have to know the system model. In the proposed method, the input-output equation of a system is identified by neural network refered to as Piecewise Linearly Trained Network (PLTN). Then from the input-output equation, the learning gain in iterative learning law is estimated. The validity of our method is demonstrated by simulations.

• 한국디지털정책학회:학술대회논문집
• /
• 2005.06a
• /
• pp.605-615
• /
• 2005

This paper represent a method of U-Learning based on advanced e-Learning. Ubiquitous computing configuration and advanced Information technology. As we know well, the 21th century is called knowledge based informational society. Many scholar stress that the improved 21th century's educational paradigm be able to success based on advanced educational paradigm. Therefore, we discuss the material for e-Learning fields including with necessity, vision, law, quality authorization etc. Also, we discuss the relational technologies including with meta data, standardization, identification etc. Finally, we propose a method for constructing the U-Learning based on advanced e-Learning and Ubiquitous computing configuration.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.96.3-96
• /
• 2001

In this paper the self-organized and fuzzy inference used stochastic gradient ascent method is proposed. Fuzzy rule and fuzzy set increase as occasion demands autonomously according to the observation information. And two rules(or two fuzzy sets)becoming to be similar each other as progress of learning are unified. This unification causes the reduction of a number of parameters and learning time. Using fuzzy inference and making a rule with an appropriate state division, our proposed method makes it possible to construct a robust reinforcement learning system.

• International journal of advanced smart convergence
• /
• v.7 no.1
• /
• pp.24-32
• /
• 2018

An outlier detection method using mixed prediction model has been described in this paper. The mixed prediction model consists of time-series model and regression model. The parameter estimation of the prediction model was performed using supervised learning and a genetic algorithm is adopted for a learning method. The experiments were performed in artificial and real data set. The prediction performance is compared with the existing prediction methods using artificial data. Outlier detection is conducted using the real sensor measurements in a dam. The validity of the proposed method was shown in the experiments.

• Journal of KIISE:Software and Applications
• /
• v.31 no.11
• /
• pp.1431-1438
• /
• 2004

Reinforcement learning learns policies for accomplishing a task's goal by experience through interaction between agent and environment. Q-learning, basis algorithm of reinforcement learning, has the problem of curse of dimensionality and slow learning speed in the incipient stage of learning. In order to solve the problems of Q-learning, new function approximation methods suitable for reinforcement learning should be studied. In this paper, to improve these problems, we suggest Fuzzy Q-Map algorithm that is based on online fuzzy clustering. Fuzzy Q-Map is a function approximation method suitable to reinforcement learning that can do on-line teaming and express uncertainty of environment. We made an experiment on the mountain car problem with fuzzy Q-Map, and its results show that learning speed is accelerated in the incipient stage of learning.

• Journal of the Korean School Mathematics Society
• /
• v.3 no.1
• /
• pp.131-148
• /
• 2000

Above all, the ability to solve problems must be emphasized as a basic skill of mathematics, but it is neglected when we teach. In this study, learning task means ［same meaning］ ［same form］ ［same technique］, so I tried to extend mathematical scholastic ability of the students as an extensional problem solving that is a basic element of mathematics. The purpose of this study is the investigation of level type learning, using the basic learning elements to extend thinking ability. From the constructed hypothesis as follows and then implement it. I selected basic learning elements from an analyzed textbook and then task learning material was created for each level type learning. The problem solving ability will be extended through the level type learning of the small group, using the level type learning task material. The conclusions this study are as follows. The level type learning in small group learning, using and making level type learning material, having basic learning elements in analysed text are. Basic learning content is understood clearly and deeply, so, fundamentally, it is effective in achieving the problem solving in mathematics. It is an effective method to achieve the meta-cognitive faculty because achieved the expected method of solving problems and resulted in the true learning of content.