• 한국멀티미디어학회논문지
• /
• 제22권1호
• /
• pp.1-9
• /
• 2019

Single image rain removal is a typical inverse problem which decomposes the image into a background scene and a rain streak. Recent works have witnessed a substantial progress on the task due to the development of convolutional neural network (CNN). However, existing CNN-based approaches train the network with synthetically generated training examples. These data tend to make the network bias to the synthetic scenes. In this paper, we present an unsupervised framework for removing rain streaks from real-world rainy images. We focus on the natural phenomena that static rainy scenes capture a common background but different rain streak. From this observation, we train siamese network with the real rain image pairs, which outputs identical backgrounds from the pairs. To train our network, a real rainy dataset is constructed via web-crawling. We show that our unsupervised framework outperforms the recent CNN-based approaches, which are trained by supervised manner. Experimental results demonstrate that the effectiveness of our framework on both synthetic and real-world datasets, showing improved performance over previous approaches.

• Educational Technology International
• /
• 제18권2호
• /
• pp.131-157
• /
• 2017

The purposes of the study were to investigate the participation of artists in paper circuit computing learning and to conduct an in-depth study on the formation and development of practical knowledge. To do this, we selected as research participants six artists who participated in the learning program of an art museum, and used various methods such as pre-open questionnaires, participation observation, and individual interviews to collect data. The collected data were analyzed based on community of practice theory. Results showed that the artists participated in the learning based on a desire to use new technology or find a new work production method for interacting with their audiences. In addition, the artists actively formed practical knowledge in the curriculum and tried to apply paper circuit computing to their works. To continuously develop the research, participants formed a study group or set up a practical goal through planned exhibitions. The results of this study can provide implications for practical approaches to, and utilization of, paper circuit computing.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2006년도 학술발표회 논문집
• /
• pp.354-359
• /
• 2006

The analysis of large-scale water resources systems is often complicated by the presence of multiple reservoirs and diversions, the uncertainty of unregulated inflows and demands, and conflicting objectives. Reinforcement learning is presented herein as a new approach to solving the challenging problem of stochastic optimization of multi-reservoir systems. The Q-Learning method, one of the reinforcement learning algorithms, is used for generating integrated monthly operation rules for the Keum River basin in Korea. The Q-Learning model is evaluated by comparing with implicit stochastic dynamic programming and sampling stochastic dynamic programming approaches. Evaluation of the stochastic basin-wide operational models considered several options relating to the choice of hydrologic state and discount factors as well as various stochastic dynamic programming models. The performance of Q-Learning model outperforms the other models in handling of uncertainty of inflows.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1999년도 제14차 학술회의논문집
• /
• pp.292-297
• /
• 1999

Evolving in artificial agent is an extremely difficult problem, but on the other hand, a challenging task. At present the studies mainly centered on single agent learning problem. In our case, we use simulated soccer to investigate multi-agent cooperative learning. Consider the fundamental differences in learning mechanism, existing reinforcement learning algorithms can be roughly classified into two types-that based on evaluation functions and that of searching policy space directly. Genetic Programming developed from Genetic Algorithms is one of the most well known approaches belonging to the latter. In this paper, we give detailed algorithm description as well as data construction that are necessary for learning single agent strategies at first. In following step moreover, we will extend developed methods into multiple robot domains. game. We investigate and contrast two different methods-simple team learning and sub-group loaming and conclude the paper with some experimental results.

• 한국산학기술학회논문지
• /
• 제14권11호
• /
• pp.5455-5464
• /
• 2013

본 연구는 조절초점이 학업성취에 영향을 미치는 과정에서 자기조절학습과 학업소진이 어떠한 역할을 하는지 알아보고자 하였다. 이를 위해 312명의 대학생을 대상으로 설문을 실시하였으며, SPSS 18.0과 Amos 8.0으로 분석하였다. 연구결과, 자기조절학습은 조절초점의 하위변인인 향상초점과 학업소진을 매개하는 것으로 나타났고, 학업소진은 자기조절학습과 학업성취를 매개하는 것으로 나타났다. 즉 조절초점은 자기조절학습을 기반으로 한 학업소진을 통해 실제 학업성취에 영향을 미치는 것으로 나타났다. 이와 같은 결과를 중심으로 학업성취를 향상시키기 위한 개입방법으로 자기조절학습을 높이고 학업소진을 줄일 수 있는 상담 및 교육의 필요성을 제언하였고, 제한점과 후속 연구의 필요성을 제안하였다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제16권2호
• /
• pp.81-86
• /
• 2016

Term weighting is a popular technique that effectively weighs the term features to improve accuracy in document classification. While several successful term weighting algorithms have been suggested, none of them appears to perform well consistently across different data domains. In this paper we propose several reasonable methods to combine different term weight vectors to yield a robust document classifier that performs consistently well on diverse datasets. Specifically we suggest two approaches: i) learning a single weight vector that lies in a convex hull of the base vectors while minimizing the class prediction loss, and ii) a mini-max classifier that aims for robustness of the individual weight vectors by minimizing the loss of the worst-performing strategy among the base vectors. We provide efficient solution methods for these optimization problems. The effectiveness and robustness of the proposed approaches are demonstrated on several benchmark document datasets, significantly outperforming the existing term weighting methods.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제50권11호
• /
• pp.522-531
• /
• 2001

Recently conventional artificial intelligence(AI) approaches have been employed to build action selectors for the autonomous mobile robot(AMR). However, in these approaches, the decision making process to choose an action from multiple competence modules is still an open question. Many researches have been focused on the reactive planning systems such as the biological immune system. In this paper, we attempt to construct an action selector for an AMR based on the artificial immune network and internet. The information from vision sensors is used for antibody. We propose a learning method for artificial immune network using evolutionary algorithm to produce antibody automatically. The internet environment for an AMR action selector shows the usefulness of the proposed learning artificial immune network application.

• 한국멀티미디어학회논문지
• /
• 제11권6호
• /
• pp.875-882
• /
• 2008

This research proposes an alternative approach to machine learning based ones for text categorization. For using machine learning based approaches for any task of text mining, documents should be encoded into numerical vectors; it causes two problems: huge dimensionality and sparse distribution. Although there are various tasks of text mining such as text categorization, text clustering, and text summarization, the scope of this research is restricted to text categorization. The idea of this research is to avoid the two problems by encoding a document or documents into a table, instead of numerical vectors. Therefore, the goal of this research is to improve the performance of text categorization by proposing approaches, which are free from the two problems.

• 국제학술발표논문집
• /
• The 6th International Conference on Construction Engineering and Project Management
• /
• pp.583-587
• /
• 2015

Scenario-based learning (SBL) has been used in a variety of training situations across different disciplines. Despite its seemly widespread use in construction management discipline, very few attempts have been made to explore its effectiveness and the respective students' learning experience. Using a survey research design, this study aims to investigate students' perceptions on SBL approach in construction management courses. The specific objectives are: (i) to identify the characteristics of a favourable SBL environment, and (ii) to explore the students' learning experience and effectiveness of the SBL approach. The results show that the four characteristics of a favourable SBL environment are: effective team formulation, constant engagement with lecturer, working in a group, and incorporation of motivational incentive for participation. The students really appreciated the opportunities to apply concepts learnt in the lectures in their SBL group work. Also, they perceived that the SBL approach is effective in developing their reflective and critical thinking skills, analytic and problem-solving skills and their ability to work as a team. These findings should facilitate more critical approaches to similar form of teaching methods.

• 기본간호학회지
• /
• 제11권3호
• /
• pp.248-255
• /
• 2004

Purpose: This study was done to compare the effectiveness of web based learning versus traditional education for learning foley-catheterization in Basic Nursing. Method: This study was a quasi-experimental research. The participants were 60 students who were taking Basic Nursing at A nursing college (3 years) in C city. Thirty students each were assigned to the experimental or control group. Data were collected between October 20 and November 4, 2003. The data were analyzed by descriptive statistics, t-test and ANCOVA. Results: The mean score for competence in foley-catheterization practice in the experimental group was 48.63 and in the control group, 44.67. This result was statistically significant (t=7.655, p=.001). The mean score for knowledge in the experimental group was 63.0, while fur the control group, 64.0. This result was not statistically significant (t=-.330, p=.743). The mean score for confidence in learning in the experimental group was 26.70 for the pre-test and 30.73 for the post-test, and in the control group 27.93 and 28.37 respectively, but this result was not statistically significant (F=.858, p=.358). Conclusion: The Web based learning was found to be effective in nursing practice but not nursing knowledge. It is necessary to continue to develop approaches to teaching nursing and to evaluate these approaches with further research.