• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.9
• /
• pp.3243-3257
• /
• 2021

Deep neural networks provide excellent performance in pattern recognition, audio classification, and image recognition. It is important that they accurately recognize input data, particularly when they are used in autonomous vehicles or for medical services. In this study, we propose a data correction method for increasing the accuracy of an unknown classifier by modifying the input data without changing the classifier. This method modifies the input data slightly so that the unknown classifier will correctly recognize the input data. It is an ensemble method that has the characteristic of transferability to an unknown classifier by generating corrected data that are correctly recognized by several classifiers that are known in advance. We tested our method using MNIST and CIFAR-10 as experimental data. The experimental results exhibit that the accuracy of the unknown classifier is a 100% correct recognition rate owing to the data correction generated by the proposed method, which minimizes data distortion to maintain the data's recognizability by humans.

• Asia Marketing Journal
• /
• v.18 no.1
• /
• pp.75-97
• /
• 2016

Globalization and the emergence of new markets has placed increased emphasis on cross-cultural sales and marketing research. Despite considerable advances in the personal selling literature, little is known about the cross-cultural transferability of key constructs. Given the degree to which well-accepted relational sales behaviors such as ASB and customer oriented selling reflect Western values such as individualism and low uncertainty avoidance, the relative efficacy of such practices in alternative cultural context is less clear. Using a Korean sample, our results confirm the beneficial direct effects of these relational selling practices upon performance. However, we also find strong workgroup interdependence diminishes the effects of ASB and that performance fully mediates the relationship between both sales behaviors and job satisfaction. In contrast to prior results, we find female salespeople and those newer to their position demonstrate stronger ASB.

• Journal of information and communication convergence engineering
• /
• v.21 no.2
• /
• pp.139-144
• /
• 2023

Engineers prefer deep neural networks (DNNs) for solving computer vision problems. However, DNNs pose two major problems. First, neural networks require large amounts of well-labeled data for training. Second, the covariate shift problem is common in computer vision problems. Domain adaptation has been proposed to mitigate this problem. Recent work on adversarial-learning-based unsupervised domain adaptation (UDA) has explained transferability and enabled the model to learn robust features. Despite this advantage, current methods do not guarantee the distinguishability of the latent space unless they consider class-aware information of the target domain. Furthermore, source and target examples alone cannot efficiently extract domain-invariant features from the encoded spaces. To alleviate the problems of existing UDA methods, we propose the mixup regularization in adversarial discriminative domain adaptation (ADDA) method. We validated the effectiveness and generality of the proposed method by performing experiments under three adaptation scenarios: MNIST to USPS, SVHN to MNIST, and MNIST to MNIST-M.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2022.04a
• /
• pp.180-181
• /
• 2022

The higher the addition rate, the less difference in kneading was found. This is because most of the cases are mixed with vinyl gloves by hand when immediately mixed in the field, so there is no problem of mixing dough according to the amount of addition. The mean increase temperature for 20 minutes was 31.6℃ for Plain, 38.1℃ for Al, 33.7℃ for Cu, and 5~7℃ for Al 20% added sample, 38.1℃ for Al 20% added sample, but 36.6℃ for Al 5% added sample. This suggests that it is possible to improve its own heat transferability by adding metal powder to epoxy adhesives.

• Journal of Korean Society of Transportation
• /
• v.32 no.3
• /
• pp.266-279
• /
• 2014

For the evaluation of roadway safety, diverse methods, including before-after studies, simple comparison using historic traffic accident data, methods based on experts' opinion or literature, have been applied. Especially, many research efforts have developed traffic accident prediction models in order to identify critical elements causing accidents and evaluate the level of safety. A traffic accident prediction model must secure predictability and transferability. By acquiring the predictability, the model can increase the accuracy in predicting the frequency of accidents qualitatively and quantitatively. By guaranteeing the transferability, the model can be used for other locations with acceptable accuracy. To this end, traffic accident prediction models using non-linear regression, artificial neural network, and structural equation were developed in this study. The predictability and transferability of three models were compared using a model development data set collected from 90 signalized intersections and a model validation data set from other 33 signalized intersections based on mean absolute deviation and mean squared prediction error. As a result of the comparison using the model development data set, the artificial neural network showed the highest predictability. However, the non-linear regression model was found out to be most appropriate in the comparison using the model validation data set. Conclusively, the artificial neural network has a strong ability in representing the relationship between the frequency of traffic accidents and traffic and road design elements. However, the predictability of the artificial neural network significantly decreased when the artificial neural network was applied to a new data which was not used in the model developing.

• Journal of Korea Water Resources Association
• /
• v.43 no.7
• /
• pp.667-679
• /
• 2010

The distributed hydrologic model has been considerably improved due to rapid development of computer hardware technology as well as the increased accessibility and the applicability of hydro-geologic information using GIS. It has been acknowledged that physically-based distributed hydrologic model require significant amounts of data for their calibration, so its application at ungauged catchments is very limited. In this regard, this study was intended to develop a distributed hydrologic model (S-RAT) that is mainly based on conceptually grid-based water balance model. The proposed model shows advantages as a new distributed rainfall-runoff model in terms of their simplicity and model performance. Another advantage of the proposed model is to effectively assess spatio-temporal variation for the entire runoff process. In addition, S-RAT does not rely on any commercial GIS pre-processing tools because a built-in GIS pre-processing module was developed and included in the model. Through the application to the two pilot basins, it was found that S-RAT model has temporal and spatial transferability of parameters and also S-RAT model can be effectively used as a radar data-driven rainfall-runoff model.

• The Journal of the Korean Society for Microbiology
• /
• v.13 no.1
• /
• pp.7-16
• /
• 1978

One hundred and three clinical isolates of enterococci were examined for susceptibility to 8 antibiotics, and transferability and transfer frequency of R plasmid. Ampicillin was the most active, followed in decreasing order by rifampin, amikacin and chloramphenicol, and tetracycline. High-level resistance(${\geq}2,000{\mu}g/ml$) to kanamycin, streptomycin, and gentamicin, known as the most active of the aminoglycosides to enterococcus, was present in 26.2%, 21.4%, and 18.3% of the isolates, respectively. In the drug susceptibility of the species, S. zymogenes was the most resistant and S. durans was the most sensitive to tested antibiotics. We could observed the transferability of enterococcal R plasmid in mixed culture: among the 28 strains which showed multiple drug resistance, 17 strains transferred all or part of their resistance with $2{\times}10^{-4}-2{\times}10^{-6}%$ of transfer frequency to a plasmid-free recipient, S. faecalis strain JH 2-2.

• THE INTERNATIONAL COMMERCE & LAW REVIEW
• /
• v.71
• /
• pp.59-84
• /
• 2016

Among various export financing, Receivable Assignment is very important. Various countries make use of this method. But Korean law system have shortage of legal structure. This paper looks into Receivable Assignment relation to legal structure. And this paper analyze not only detail Korean civil law system about bond and receivable but also comparative other civil law system. Especially, Korean civil law of bond compare Germany civil law of bond or Japanese civil law of bond. In the context, This paper compares Korean civil law system about bond with International standard rule about bond. For example, It is UN convention on the Assignment of Receivables in International Trade and Principles of International Commercial Contract(PICC). This is good for the commercial party in terms of financing and receivable assignment. Thus this paper will establish Korean legal system direction. There are argument on method of making article and modifing article. The purposes of this paper is to examine revitalizing of Receivable Assignment. And this paper deals with improvement of International Commercial Activation.

• Journal of Korea Multimedia Society
• /
• v.12 no.7
• /
• pp.953-961
• /
• 2009

The usage of DRM-applied digital contents reduces user's convenience, thus it leads the many users to avoid using DRM-applied digital contents. In this paper, therefore, we propose the new distribution model in which the users, who purchase the DRM-applied digital contents, can resale freely those contents to the other users. As a result, the proposed model induces the users to distribute contents soundly by including them in one of distribution objects. For the purpose of this, in this paper, we propose that a packaged content and usage rights are transferred to another user by using a method which based on some existing digital ticket or coupon. In addition, the security of the proposed system is discussed for the double transfer, the forgery and the modification of usage rights, and the reproduction of contents and usage rights.

• Journal of applied mathematics & informatics
• /
• v.31 no.3_4
• /
• pp.425-441
• /
• 2013

In literature, several strong designated verifier signature (SDVS) schemes have been devised using elliptic curve bilinear pairing and map-topoint (MTP) hash function. The bilinear pairing requires a super-singular elliptic curve group having large number of elements and the relative computation cost of it is approximately two to three times higher than that of elliptic curve point multiplication, which indicates that bilinear pairing is an expensive operation. Moreover, the MTP function, which maps a user identity into an elliptic curve point, is more expensive than an elliptic curve scalar point multiplication. Hence, the SDVS schemes from bilinear pairing and MTP hash function are not efficient in real environments. Thus, a cost-efficient SDVS scheme using elliptic curve cryptography with pairingfree operation is proposed in this paper that instead of MTP hash function uses a general cryptographic hash function. The security analysis shows that our scheme is secure in the random oracle model with the hardness assumption of CDH problem. In addition, the formal security validation of the proposed scheme is done using AVISPA tool (Automated Validation of Internet Security Protocols and Applications) that demonstrated that our scheme is unforgeable against passive and active attacks. Our scheme also satisfies the different properties of an SDVS scheme including strongness, source hiding, non-transferability and unforgeability. The comparison of our scheme with others are given, which shows that it outperforms in terms of security, computation cost and bandwidth requirement.