• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2006년도 학술발표회 논문집
• /
• pp.51-56
• /
• 2006

Biases embedded in numerical weather precipitation forecasts by the RDAPS model was determined, quantified and corrected. The ultimate objective is to eventually enhance the reliability of reservoir operation by Korean Water Resources Corporation (KOWACO), which is based on precipitation-driven forecasts of stream flow. Statistical post-processing, so called MOS (Model Output Statistics) was applied to RDAPS to improve their performance. The Artificial Neural Nwetwork (ANN) model was applied for 4 cases of 'Probability of Precipitation (PoP) for wet and dry season' and 'Quantitative Precipitation Forecasts (QPF) for wet and dry season'. The reduction on the large systematic bias was especially remarkable. The performance of both networks may be improved by retraining, probably every month. In addition, it is expected that performance of the networks will improve once atmospheric profile data are incorporated in the analysis. The key to the optimal performance of ANN is to have a large data set relevant to the predictand variable. The more complex the process to be modeled by the ANN, the larger the data set needs to be.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2012년도 제24회 한글 및 한국어 정보처리 학술대회
• /
• pp.157-159
• /
• 2012

화자의 의도를 결정하는 문제는 대화 시스템에서 핵심적인 부분이다. 기존의 연구에서는 모델의 간소화를 위해 화자의 의도를 화행과 개념이라는 두 요소로 분리하여 분석하였다. 하지만 두 요소는 서로 밀접하게 관련되어 있기 때문에 모델의 간소화는 의도 분석 성능 저하의 원인이 되었다. 이런 문제점을 해결하기 위해 본 논문에서는 화자 의도 분류를 위한 재학습 방법을 제안한다. 제안된 방법은 화자의 의도를 분석하기 위해 화행 분류 모델과 개념열 분석 모델로 분리하여 분석한다. 학습 단계에서 화행 분류 모델은 개념열 분류 결과를 입력으로 사용하고 개념열 역시 마찬가지로 적용하였다. 목적 지항 대화를 대상으로 한 실험에서 제안된 시스템은 화자 의도 분류에서 최대엔트로피 모델과 지지 벡터 기계의 성능을 효과적으로 향상시켰다.

• 가족자원경영과 정책
• /
• 제2권2호
• /
• pp.93-103
• /
• 1998

The purpose of this study is to explain changes in the mode of labor brought abut with the advance of information society and to suggest policies for the success of teleworking. It is likely that more and more people in advanced countries will be involved in some form of teleworking office workers use computers for handling dat, which is then transmitted over a telecommunications like to client or employer located some distance away. Though teleworking is becoming populer rapidly, it still has many problems to solve. In the conclusion, the paper suggests some measures to be taken for a successful teleworking system. (1) The first step is to establish a special team for the teleworking pilot. And probably the telework center will be most useful type for the pilot. (2) Some protections for teleworkers should be designed against discrimination in employment and occupation. (3) Some institutional arrangements such as tax benefits and incentive system are required for the success of teleworking system. (4) Maintaining competitiveness-faster services and lower cost-should be the first priority. (5) There should be a retraining and education system in the national level for the teleworkers to learn new IT applications.

• Journal of Information Processing Systems
• /
• 제14권5호
• /
• pp.1075-1086
• /
• 2018

In this paper a cross-validation algorithm for training probabilistic neural networks (PNNs) is presented in order to be applied to automatic face identification. Actually, standard PNNs perform pretty well for small and medium sized databases but they suffer from serious problems when it comes to using them with large databases like those encountered in biometrics applications. To address this issue, we proposed in this work a new training algorithm for PNNs to reduce the hidden layer's size and avoid over-fitting at the same time. The proposed training algorithm generates networks with a smaller hidden layer which contains only representative examples in the training data set. Moreover, adding new classes or samples after training does not require retraining, which is one of the main characteristics of this solution. Results presented in this work show a great improvement both in the processing speed and generalization of the proposed classifier. This improvement is mainly caused by reducing significantly the size of the hidden layer.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제4권2호
• /
• pp.236-240
• /
• 2004

Stroke-based composite HMMs with articulation states are proposed to deal with 3D spatio-temporal trajectory gestures. The direct use of 3D data provides more naturalness in generating gestures, thereby avoiding some of the constraints usually imposed to prevent performance degradation when trajectory data are projected into a specific 2D plane. Also, the decomposition of gestures into more primitive strokes is quite attractive, since reversely concatenating stroke-based HMMs makes it possible to construct a new set of gesture HMMs without retraining their parameters. Any deterioration in performance arising from decomposition can be remedied by a partial tuning process for such composite HMMs.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.484-484
• /
• 2000

The problem of determining the proper size of an neural network is recognized to be crucial, especially for its practical implications in such important issues as learning and generalization. Unfortunately, it usually is not obvious what size is best: a system that is too snail will not be able to learn the data while one that is just big enough may learn the slowly and be very sensitive to initial conditions and learning parameters. One popular technique is commonly known as pruning and consists of training a larger than necessary network and then removing unnecessary weights/nodes. In this paper, a new pruning method is developed, based on the penalty-term methods. This method makes the neural network good for the generalization and reduces the retraining time after pruning weights/nodes.

• 한국통신학회논문지
• /
• 제28권7C호
• /
• pp.705-711
• /
• 2003

This paper studies the application of a fuzzy-ARTMAP neural network to digital communications channel equalization. This approach provides new solutions for solving the problems, such as complexity and long training, which found when implementing the previously developed neural-basis equalizers. The proposed fuzzy-ARTMAP equalizer is fast and easy to train and includes capabilities not found in other neural network approaches; a small number of parameters, no requirements for the choice of initial weights, automatic increase of hidden units, no risk of getting trapped in local minima, and the capability of adding new data without retraining previously trained data. In simulation studies, binary signals were generated at random in a linear channel with Gaussian noise. The performance of the proposed equalizer is compared with other neural net basis equalizers, specifically MLP and RBF equalizers.

• 한국지능시스템학회논문지
• /
• 제2권3호
• /
• pp.50-57
• /
• 1992

Neural networks have become more popular recently and are now being applied to numerous fiedls. One of the major applications of neural networks is image recognition. Various image recognition system have been proposed so far, but there is no definite solution yet. In this paper, we propose a design of Fuzzy Neural Network Image Recognizer(FNNIR). Our model uses a fuzzy neural network model, named SONN[KIM90]. This model returns the information of the number of clusters and cluster and cluster center values for a given image data ste. Unlike the well-kinwn backpropagation technique, we do not need retraining for new data. Our newly designed image recongitionsystem FNNIR that uses fuzzy merger is proposed and experimented for a sample color image.

• 한국건축시공학회:학술대회논문집
• /
• 한국건축시공학회 2015년도 춘계 학술논문 발표대회
• /
• pp.217-218
• /
• 2015

There are various standards and manuals on the quantity takeoff for construction estimation. However, a lot of time and money has been spent on retraining because of difference between practice and education. Therefore, the main purpose of this paper is to suggest improvement to analyze problems and limitations of the existing educational content for quantity takeoff. In order to achieve research objective, comparative analysis for quantity takeoff contents is carried out. This study will be extended to assist in effective quantity takeoff in building projects by supplementing the existing issues with quantity takeoff contents and developing a BIM-based quantity takeoff multimedia contents that use diverse media to increase the understanding on the subject for a non-professional.

• 한국환경교육학회지:환경교육
• /
• 제13권1호
• /
• pp.14-22
• /
• 2000

This attempts to find out training system of environment education teacher in Korea. The results are summarized as follows. The primary and secondary school have focused on environment education and the environment course was designated as a subject, but only 12% of the 2741 middle school chose the environment subject in 1998. The environment education course is not popular among students. The environment education is an interdisciplinary subject, which is composed of natural science, social studies, earth science, and medical science, that is why the subject is so unsystematic and complicated that appropriate teaching methods and contents for school classes are not able to be developed. Moreover, material and manuals in environment education for students and teachers are limited. While the contents of environment education is composed of field experience learning and experiment learning, but lecture-centered instruction is emphasized in school because of materials, time and experts. Over 300 environmental education teachers are annually produced, but the ratio of employment low. is, Therefore, a retraining program for environment education teacher needs to be developed.