• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10b
• /
• pp.96-98
• /
• 2000

본 논문에서는 상거래 환경에서 구매자와 비구매자들에 대한 데이터를 학습한 후, 잠재고객들 중에서 구매 확률이 높은 사람을 예측하는 문제에 효율적으로 접근하기 위해 능동적인 데이터 선택 기법을 이용한다. 실험 데이터는 ColL Challenge 2000에서 얻은 데이터로서, 구매자들의 정보보다 비구매자들의 정보가 더 많기 때문에 상당히 균형이 맞지 않는다. 따라서 모든 데이터를 한꺼번에 학습하는 경우에 성능이 좋지 않다. 본 논문에서는 이러한 불균형 분포를 갖는 실제적인 문제에 있어서 성능이 좋지 않다. 본 논문에서는 이러한 불균형 분포를 갖는 실제적인 문제에 있어서 RBF 기반의 신경망을 가지고 능동 학습을 함으로써 기존의 뱃치학습 보다 예측의 정확도를 향상시킬 수 있음을 보인다.

• The Journal of Korean Association of Computer Education
• /
• v.8 no.1
• /
• pp.25-32
• /
• 2005

An analysis of the case study was reported to pioneer the perceptions of teachers, students, and parents for the educational use of tablet computers. The findings showed that the amount of learning, various learning activities, interest, and motivation of learners increased and that the teachers perceived the potentiality of the tablet computer and the necessity of a training for designing mobile-based instructions.

• Journal of KIISE:Software and Applications
• /
• v.34 no.12
• /
• pp.1045-1055
• /
• 2007

Latent topic models are statistical models which automatically captures salient patterns or correlation among features underlying a data collection in a probabilistic way. They are gaining an increased popularity as an effective tool in the application of automatic semantic feature extraction from text corpus, multimedia data analysis including image data, and bioinformatics. Among the important issues for the effectiveness in the application of latent topic models to the massive data set is the efficient learning of the model. The paper proposes an accelerated learning technique for PLSA model, one of the popular latent topic models, by an incremental EM algorithm instead of conventional EM algorithm. The incremental EM algorithm can be characterized by the employment of a series of partial E-steps that are performed on the corresponding subsets of the entire data collection, unlike in the conventional EM algorithm where one batch E-step is done for the whole data set. By the replacement of a single batch E-M step with a series of partial E-steps and M-steps, the inference result for the previous data subset can be directly reflected to the next inference process, which can enhance the learning speed for the entire data set. The algorithm is advantageous also in that it is guaranteed to converge to a local maximum solution and can be easily implemented just with slight modification of the existing algorithm based on the conventional EM. We present the basic application of the incremental EM algorithm to the learning of PLSA and empirically evaluate the acceleration performance with several possible data partitioning methods for the practical application. The experimental results on a real-world news data set show that the proposed approach can accomplish a meaningful enhancement of the convergence rate in the learning of latent topic model. Additionally, we present an interesting result which supports a possible synergistic effect of the combination of incremental EM algorithm with parallel computing.

• The Journal of Korean Philosophical History
• /
• no.56
• /
• pp.209-238
• /
• 2018

The purpose of this research is to analyze characteristics of the method of education or learning reflected in the teacher-student relationship in On Haeng Lok By Toe-gye and explore their relevance to modern education. By writing various works and conversing with his students, Toe-gye devoted himself in the education of the traditional Confucian principles. Specially, he taught his students based on two Confucian educative principles, namely Shim Deuk(心得) and Goong Haeng(躬行). Judging from this, Toe-gye can be seen as someone who tries to fulfill the role of teacher as dictated in the educative principles of the Confucianism. In Confucianism, teacher is responsible for forming a well-rounded view on life in student, rather than simply transmitting knowledge. As such, the teacher was supposed to find a harmonious way to create something new based on what was inherited from the past generation and try to do his best in learning new things himself and teaching his students. These Toe-gye managed to do successfully, earning his students' trust and respect. Being moved and inspired by their teacher, the students continued their intellectual pursuit. This relationship between Toe-gye and his students can be analyzed from the perspective of education method and discussed in terms of cognitive learning and adult learning. In terms of cognitive learning, the education method reflected in the relationship is similar to potential learning, insight learning, and imitation learning. In terms of adult learning, it is similar to self-directed learning and communicative learning.!

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.611-614
• /
• 2024

최근 많은 분야에서 인공지능을 사용한 산업이 각광을 받고 있고 그중 챗-GPT 로 인하여 챗봇에 관한 관심도가 높아져 관련 연구가 많이 진행되고 있다. 특히 질문에 대한 답변을 생성해주는 분야에 대한 연구가 많이 이루어지고 있는데, 질문-답변의 데이터 셋에 대한 학습 방식보다는 질문-답변-배경지식으로 이루어진 데이터 셋에 대한 학습 방식이 많이 연구가 되고 있다. 그러다 보니 배경지식을 어떤 방식으로 모델에게 이해를 해줄 지가 모델 성능에 큰 부분 차지한다. 그리고 최근 연구에 따르면 이러한 배경지식 정보를 이해시키기 위해 잠재 변수 모델링 기법을 활용하는 것이 높은 성능을 갖는다고 하고 트랜스포머 기반 모델 중 생성 문제에서 강점을 보이는 BART(Bidirectional Auto-Regressive Transformer)[1]도 주로 활용된다고 한다. 본 논문에서는 BART 모델에 잠재 변수 모델링 기법 중 잠재 변수를 어텐션에 곱하는 방식을 이용한 모델을 통해 답변 생성 문제에 관한 해결법을 제시하고 그에 대한 결과로 배경지식 정보를 담은 답변을 보인다. 생성된 답변에 대한 평가는 기존에 사용되는 BLEU 방식과 배경지식을 고려한 방식의 BLEU 로 평가한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2000.10a
• /
• pp.183-186
• /
• 2000

멀티미디어 컨텐츠는 다양한 매체적 속성으로 학습 흥미를 높일 뿐 아니라. 잠재적인 학습성취도를 중지시킬 것이라고 기대한다. 특히, 학습자들의 상호작용성이 두드러진 사이버 환경속에서 멀티미디어 매체의 기대효과는 상당히 크다. 본 논문에서는 새로운 형태의 교수-학습 매체로 등장한 웹 컨텐츠의 형태에 따른 학습자의 학습성취도를 분석하고, 학습자의 내재적 학습동기를 유발하는 요소를 고려한 새로운 학습 패러다임을 지원하는 웹 컨텐츠를 설계 및 구현한다.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.10
• /
• pp.193-201
• /
• 2020

This study is to evaluate the effectiveness of problem-solving ability by applying a team-based learning model to the classes of humanities and social science students, and to conduct a structural model analysis on the relationship between sub-factors. Team-based learning was conducted six times in six teams with 30 students in the second and third grades of the humanities and social sciences. The problem solving ability score of the target students was significantly higher after team-based learning and was statistically significant. There was no problem in normality with the latent variables, which are the sub-factors of problem solving ability, and the factor load value was statistically significant at the .001 level in the confirmatory factor analysis of the observed variables for the latent variables, which was a valid model. A good level of fitness was also shown in the verification of the fitness of the research model. As a result, it was analyzed that latent variables of cause analysis, problem clarification, planning execution, performance evaluation, and alternative development had an indirect or direct influence on each other.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.8
• /
• pp.355-364
• /
• 2023

As advanced cyber threats continue to increase in recent years, it is difficult to detect new types of cyber attacks with existing pattern or signature-based intrusion detection method. Therefore, research on anomaly detection methods using data learning-based artificial intelligence technology is increasing. In addition, supervised learning-based anomaly detection methods are difficult to use in real environments because they require sufficient labeled data for learning. Research on an unsupervised learning-based method that learns from normal data and detects an anomaly by finding a pattern in the data itself has been actively conducted. Therefore, this study aims to extract a latent vector that preserves useful sequence information from sequence log data and develop an anomaly detection learning model using the extracted latent vector. Word2Vec was used to create a dense vector representation corresponding to the characteristics of each sequence, and an unsupervised autoencoder was developed to extract latent vectors from sequence data expressed as dense vectors. The developed autoencoder model is a recurrent neural network GRU (Gated Recurrent Unit) based denoising autoencoder suitable for sequence data, a one-dimensional convolutional neural network-based autoencoder to solve the limited short-term memory problem that GRU can have, and an autoencoder combining GRU and one-dimensional convolution was used. The data used in the experiment is time-series-based NGIDS (Next Generation IDS Dataset) data, and as a result of the experiment, an autoencoder that combines GRU and one-dimensional convolution is better than a model using a GRU-based autoencoder or a one-dimensional convolution-based autoencoder. It was efficient in terms of learning time for extracting useful latent patterns from training data, and showed stable performance with smaller fluctuations in anomaly detection performance.

• Proceedings of the KGS Conference
• /
• 2002.11a
• /
• pp.133-140
• /
• 2002

1980년 후반부터 지리에서 과학적 탐구를 지원하는 학습도구로써 GIS의 잠재성을 모색하기 시작하였다. 이후 지리교육에서 GIS의 유용성, 지리교육에 GIS를 적용하는데 해결해야할 문제들, 지리교육에 적용할 때 필요한 접근방법, 다양한 학습 모형 개발에 대한 연구가 진행되었다. 그리고 최근 Internet GIS 기술이 개발되면서 Web을 통한 학습 자료의 개발이 이루어지고 있다.(중략)

• The Journal of Korean Association of Computer Education
• /
• v.22 no.1
• /
• pp.31-46
• /
• 2019

The purpose of this study is to analyze the self-directed learning types of software education learners and to characterize them according to each type. To do this, 429 middle school students participating in online software education at K university were surveyed and a latent class analysis to analyze self-directed learning types was conducted. As a result, the self-directed learning types of the software education learners were classified into 'highest level of self-directed learning type (class 1)', 'self learning style recognition type (class 2)', 'self learning style preference type (class 3)', and 'lack of self-directed learning type(class 4)'. Also, the level of software learning achievement according to self-directed learning type of software education learners was found to be the highest at 'highest level of self-directed learning type (class 1)' and lowest at 'self learning style preference type (class 3)'. Based on these results, we suggested the strategic implications for software education.