• Education of Primary School Mathematics
• /
• v.17 no.1
• /
• pp.1-16
• /
• 2014

The purpose of this study is to analyze whether analogy distance and mathematical knowledge affect on transfer problems solving with different analogy distance. To conduct the study, transfer problems were classified into multiple categories: mathematical word problem based on rates, science word problem based on rates, and real-life problem based on rates with different analogy distance. Then analysed there are differences in participants' transfer ability and which mathematical knowledge contributes to the solution on over the three transfer problem. The study demonstrated a statistical significant difference(.05) in participants' three transfer problem solving and a gradual decrease of the participants' success rates of on transfer problems solving. Moreover, conceptual knowledge influenced transfer problem solving more than factual knowledge about rates. The study has an important implications in that it provided new direction for study about transfer of learning, and also show a good mathematics instruction on where teachers will put the focus in mathematical lesson to foster elementary students' transfer ability.

• Journal of Internet Computing and Services
• /
• v.20 no.4
• /
• pp.91-102
• /
• 2019

In autonomous driving systems, the ability to classify pedestrians in images captured by cameras is very important for pedestrian safety. In the past, after extracting features of pedestrians with HOG(Histogram of Oriented Gradients) or SIFT(Scale-Invariant Feature Transform), people classified them using SVM(Support Vector Machine). However, extracting pedestrian characteristics in such a handcrafted manner has many limitations. Therefore, this paper proposes a method to classify pedestrians reliably and effectively using CNN's(Convolutional Neural Network) deep features and transfer learning. We have experimented with both the fixed feature extractor and the fine-tuning methods, which are two representative transfer learning techniques. Particularly, in the fine-tuning method, we have added a new scheme, called M-Fine(Modified Fine-tuning), which divideslayers into transferred parts and non-transferred parts in three different sizes, and adjusts weights only for layers belonging to non-transferred parts. Experiments on INRIA Person data set with five CNN models(VGGNet, DenseNet, Inception V3, Xception, and MobileNet) showed that CNN's deep features perform better than handcrafted features such as HOG and SIFT, and that the accuracy of Xception (threshold = 0.5) isthe highest at 99.61%. MobileNet, which achieved similar performance to Xception and learned 80% fewer parameters, was the best in terms of efficiency. Among the three transfer learning schemes tested above, the performance of the fine-tuning method was the best. The performance of the M-Fine method was comparable to or slightly lower than that of the fine-tuningmethod, but higher than that of the fixed feature extractor method.

• The Journal of the Korea Contents Association
• /
• v.17 no.9
• /
• pp.661-671
• /
• 2017

The purpose of this study is to identify the correlations among learning self-efficacy, confidence in performance, perception of importance and transfer intention for core basic nursing skill in nursing students. The subjects of this study were 2nd grade students at a nursing university. The collected data were analyzed using SPSS 21.0 program. As a result, the transfer intention had a correlation with the learning self-efficacy (r=.49, p<.001), confidence in performance (r=.30, p=.006), perception of the importance (r=.31, p=.005). The results of this study suggest that further research is necessary to verify the causal relationship between the transfer intention and the related variables in order to develop an effective education program for promoting the transfer intention.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.102-104
• /
• 2017

최근 Convolutional neural network (CNN)을 도입하여, SAR 영상의 목표물 인식 알고리즘이 높은 성능을 보여주었다. SAR 영상은 4 종류의 polarization 정보로 구성되어있다. 기계와 신호처리의 비용으로 인하여 일부 데이터는 적은 수의 polarization 정보를 가지고 있다. 따라서 우리는 SAR 영상 data 를 멀티모달 데이터로 해석하였다. 그리고 우리는 이러한 멀티모달 데이터에 잘 작동할 수 있는 콘볼루션 신경망을 제안하였다. 우리는 데이터가 포함하는 모달의 수에 반비례 하도록 scale factor 구성하고 이를 입력 크기조절에 사용하였다. 입력의 크기를 조절하여, 네트워크는 특징맵의 크기를 모달의 수와 상관없이 일정하게 유지할 수 있었다. 또한 제안하는 입력 크기조절 방법은 네트워크의 dead filter 의 수를 감소 시켰고, 이는 네트워크가 자신의 capacity 를 잘 활용한다는 것을 의미한다. 또 제안된 네트워크는 특징맵을 구성할 때 다양한 모달을 활용하였고, 이는 네트워크가 모달간의 상관관계를 학습했다는 것을 의미한다. 그 결과, 제안된 네트워크의 성능은 입력 크기조절이 없는 일반적인 네트워크보다 높은 성능을 보여주었다. 또한 우리는 전이학습의 개념을 이용하여 네트워크를 모달의 수가 많은 데이터부터 차례대로 학습시켰다. 전이학습을 통하여 네트워크가 학습되었을 때, 제안된 네트워크는 특정 모달의 조합 경우만을 위해 학습된 네트워크보다 높은 성능을 보여준다.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.295-299
• /
• 2022

자원이 부족한 언어 환경에서 사전학습 언어모델 학습을 위한 대용량의 코퍼스를 구축하는데는 한계가 존재한다. 본 논문은 이러한 한계를 극복할 수 있는 Cross-lingual Post-Training (XPT) 방법론을 적용하여 비교적 자원이 부족한 한국어에서 해당 방법론의 효율성을 분석한다. 적은 양의 한국어 코퍼스인 400K와 4M만을 사용하여 다양한 한국어 사전학습 모델 (KLUE-BERT, KLUE-RoBERTa, Albert-kor)과 mBERT와 전반적인 성능 비교 및 분석 연구를 진행한다. 한국어의 대표적인 벤치마크 데이터셋인 KLUE 벤치마크를 사용하여 한국어 하위태스크에 대한 성능평가를 진행하며, 총 7가지의 태스크 중에서 5가지의 태스크에서 XPT-4M 모델이 기존 한국어 언어모델과의 비교에서 가장 우수한 혹은 두번째로 우수한 성능을 보인다. 이를 통해 XPT가 훨씬 더 많은 데이터로 훈련된 한국어 언어모델과 유사한 성능을 보일 뿐 아니라 학습과정이 매우 효율적임을 보인다.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.5
• /
• pp.117-122
• /
• 2024

Many studies on software defect prediction have been conducted, but it has been difficult to use them due to a lack of training data. Cross-project defect prediction is a technique to solve this problem, where a prediction model learned with sufficient training data from existing source project is used to predict defects in the target project. Before learning, domain adaptation techniques, a type of transfer learning, are used to minimize the difference in data distribution between the two projects. In this paper, we produced new prediction models using W-BDA and MEDA and compared their performance with existing models using TCA and BDA. As a result of the evaluation experiment, MEDA showed irregular and poor performance compared to other models, but BDA showed better performance than TCA, and W-BDA showed slightly better performance than BDA.

• Fire Science and Engineering
• /
• v.30 no.3
• /
• pp.117-123
• /
• 2016

To accomplish the purpose, the current study drew factors affecting the transfer of education and training through a review of domestic and overseas literature, and aimed to empirically investigate whether these factors actually affect the transfer of education and training of fire officers. The results showed that significant variables affecting the degree of perception on the transfer of education and training were in the order of work relationship, learning culture, peer support, self-efficacy, learning motivation, learning ability, and teaching method.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.12 no.6
• /
• pp.1-12
• /
• 2017

This paper is brought to asses the training effect of KS-QFD boot camp for the companies in the early growth stage. In particular, the focus of research falls on measuring transfer intension of the participants from the early stage companies older than three years old, motivating effect of applying knowledges acquired from KS-QFD training camp into their real business case. KS-QFD program is presented to help company in the early stage companies over three years old of boosting up their sales volume more than 5 times than now for the next 18 months by this training. The training program of KS-QFD is ultimately to design more practical and helpful program to real business and spread out. The research establish model by setting the learner readiness and perceived content validity by doing training design as independent variables, self-efficacy of learner as mediating variable, and transfer intension as dependant variable. Research results shows the following outcomes. First, learner readiness does not have directly effect on transfer intension under keeping statistical significance. But as the parameter of self-efficacy, it has perfect mediating effect. Second, research proves that perceived content validity have directly impact on learning transfer intension of mediating by self-efficacy partially. This research contributes on proving that learning by doing KS-QFD boot camp enable the participants to build up their self-efficacy and lead to enhance transfer intension. In more steps, the research validates that KS-QFD training camp have delivered very practical and helpful on-site knowledge to the participants.

• The Journal of Korean Association of Computer Education
• /
• v.19 no.6
• /
• pp.55-68
• /
• 2016

Informatics curriculum revised 2015 proposed the use of block type and text type of programming language by organizing problem solving and the programming unit in a spiral. The purpose of this study is to find out whether the algorithms helps programming learning and whether there is a positive transition effect in block type programming learning to text type programming trailing learning. For 15 elementary school students was conducted block type and text type programming learning. As a result of the research, it is confirmed that writing the algorithm in a limited way can interfere with the learner's expression of thinking, but the block type programming learning has a positive transition to the text type programming learning. This study is meaningful that it suggested a plan for the programming education which is sequential from elementary school.

• Journal of Internet Computing and Services
• /
• v.22 no.2
• /
• pp.1-9
• /
• 2021

Traditionally, most malicious codes have been analyzed using feature information extracted by domain experts. However, this feature-based analysis method depends on the analyst's capabilities and has limitations in detecting variant malicious codes that have modified existing malicious codes. In this study, we propose a ResNet-Variational AutoEncder-based variant malware classification method that can classify a family of variant malware without domain expert intervention. The Variational AutoEncoder network has the characteristics of creating new data within a normal distribution and understanding the characteristics of the data well in the learning process of training data provided as input values. In this study, important features of malicious code could be extracted by extracting latent variables in the learning process of Variational AutoEncoder. In addition, transfer learning was performed to better learn the characteristics of the training data and increase the efficiency of learning. The learning parameters of the ResNet-152 model pre-trained with the ImageNet Dataset were transferred to the learning parameters of the Encoder Network. The ResNet-Variational AutoEncoder that performed transfer learning showed higher performance than the existing Variational AutoEncoder and provided learning efficiency. Meanwhile, an ensemble model, Stacking Classifier, was used as a method for classifying variant malicious codes. As a result of learning the Stacking Classifier based on the characteristic data of the variant malware extracted by the Encoder Network of the ResNet-VAE model, an accuracy of 98.66% and an F1-Score of 98.68 were obtained.