• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.529-532
• /
• 2023

본 논문은 소셜 네트워크 이용자의 텍스트 데이터를 대상으로, 트랜스포머 계열의 언어모델을 전이학습해 이용자의 MBTI 성격 유형을 분류한 국내 첫 연구이다. Kaggle MBTI Dataset을 대상으로 RoBERTa Distill, DeBERTa-V3 등의 사전 학습모델로 전이학습을 해, MBTI E/I, N/S, T/F, J/P 네 유형에 대한 분류의 평균 정확도는 87.9181, 평균 F-1 Score는 87.58를 도출했다. 해외 연구의 State-of-the-art보다 네 유형에 대한 F1-Score 표준편차를 50.1% 낮춰, 유형별 더 고른 분류 성과를 보였다. 또, Twitter, Reddit과 같은 글로벌 소셜 네트워크 서비스의 텍스트 데이터를 추가로 분류, 트랜스포머 기반의 MBTI 분류 방법론을 확장했다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2003.05c
• /
• pp.2213-2216
• /
• 2003

단일 호스트 환경에 특화되어 설계되어온 기존 침입탐지 시스템(Intrusion Detection System: IDS)은 침입 시 도메인의 보호만을 그 목적으로 하는 수동적인 성격으로써, 새로운 공격 기법에 대한 탐지 및 대응, 그리고 보다 그 규모가 큰 네트워크로의 확장 면에서 구조적인 결함을 가지고 있다. 이러한 IDS의 구조적 문제점의 해결방안으로 액티브 네트워크 기반의 IDS 에 관한 연구가 진행되고 있다. 액티브 네트워크(Active network)란 패킷 스위칭 네트워크 상에 프로그램 가능한 라우터 등인 액티브 노드들을 배치하고, 사용자의 요구에 상응하는 적절한 연산을 위한 데이터와 프로그램으로 구성된 스마트 패킷(smart packet)에 대하여 수행 가능하게 하는 접근 방법이다. 본 논문에서는 이를 기반으로 자율적이며 지능적인 에이전트로 구성된 멀티 에이전트 기술을 액티브 노드에 적용함으로써 기존 IDS 보안메커니즘에서 보다 러 진보된 능동적이고 적극적인 대응을 위한 보안 메커니즘을 제공하여 네트워크 공격에 의한 피해 최소화와 신속한 대응이 가능한 멀티 에이전트 기반 공격 대응 메커니즘을 제시하고, 이를 적용 가능한 액티브 네트워크 기반 프레임 설계를 제안한다.

• Journal of KIISE:Information Networking
• /
• v.30 no.1
• /
• pp.97-116
• /
• 2003

Frequency of attacks on web services and the resulting damage continue to grow as web services become popular. Techniques used in web service attacks are usually different from traditional network intrusion techniques, and techniques to protect web services are badly needed. Unfortunately, conventional intrusion detection systems (IDS), especially those based on known attack signatures, are inadequate in providing reasonable degree of security to web services. An application-level IDS, tailored to web services, is needed to overcome such limitations. The first step in developing web application IDS is to analyze known attacks on web services and characterize them so that anomaly-based intrusion defection becomes possible. In this paper, we classified known attack techniques to web services by analyzing causes, locations where such attack can be easily detected, and the potential risks.

• Proceedings of the Korea Institutes of Information Security and Cryptology Conference
• /
• 2006.06a
• /
• pp.269-272
• /
• 2006

최근 들어, 유비쿼터스 홈네트워크에 대한 관심이 높아지고 있지만, 기능 구현에 초점을 맞추고 있다. 이러한 홈네트워크는 단일 서비스가 아닌 다양한 서비스 집합으로서의 성격이 강하므로 세분화된 보안 요구사항과 제한된 자원에서의 원활한 서비스를 위해 시스템 경량화는 필수적 요소이다. 이에 본 논문에서는 유비쿼터스 홈네트워크 환경에서 요구하는 보안성 및 경량화를 고려한 새로운 침입탐지 모델을 제안한다.

• Korean Journal of Culture and Social Issue
• /
• v.15 no.1
• /
• pp.169-186
• /
• 2009

This study aims to investigate the relationship between personality, especially temperament and subjective well-being. The participants were 681 college students (211 males and 470 females), whose ages ranged from 17 to 37 (M=20.91, SD=2.36). The instruments utilized in the present study were Korean Version of BAS/BIS Scale, The Big Five Locator, Satisfaction with Life Scale, Life Satisfaction Motivation Scale, Life Satisfaction Expectancy Scale, Emotion Frequency Test, and Subjective Happiness Scale. Result indicated that women expected more positive future than men did, while no gender differences were found in any other well-being variables. Correlational analyses revealed that emotional stability and extroversion were closely associated with life satisfaction, happiness, positive and negative emotion, whereas behavioral activation system (BAS) and behavioral inhibition system (BIS) were more closely associated with motivation to live and expectancy of satisfactory life. There was close relationship between conscientiousness and subjective well-being, because they were college students. As a internal factor, personality was better predictor for subjective well-being of female students. For instance, it accounted for around 35% variance of female's subjective happiness. The present findings reiterate the role of personality in quality of life, and it was discussed with characteristics of subjects, situational factors, and previous studies.

• Journal of the Institute of Convergence Signal Processing
• /
• v.12 no.3
• /
• pp.204-211
• /
• 2011

Embedded systems require intensively complex and high power dissipating modules that have the capability of real-time high performance data processing, wide bandwidth communication, and low power consumption. However, the current battery technology has not been developed as much as meeting the requirements of portable embedded systems for long system lifetime. In this paper, new approach that operates with low energy consumption is proposed to overcome the situation while guaranteeing detection accuracy. The designed method associates a variety of detection algorithms hierarchically to detect events happening around the system. The Change for energy consumption characteristics is shown with change for probabilistic characteristics and those relationships are analytically explained from experiments. Furthermore, several techniques to consume low energy and achieve high detection accuracy are described, depending on the event static or dynamic characteristics.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.4
• /
• pp.44-53
• /
• 2010

Program plagiarism is a significant factor to reduce the quality of education in computer programming. In this paper, we propose the technique of identifying similar or identical programs in order to prevent students from reckless copying their programming assignments. Existing approaches for identifying similar programs are mainly based on fingerprints or pattern matching for text documents. Different from those existing approaches, we propose an approach based on the program structur. Using paring progrmas, we first transform programs into XML documents by representing syntactic components in the programs with elements in XML document, then run X-tree Diff+, which is the change detection algorithm for XML documents, and produce an edit script as a change. The decision of similar or identical programs is made on the analysis of edit scripts in terms of program plagiarism. Analysis of edit scripts allows users to understand the process of conversion between two programs so that users can make qualitative judgement considering the characteristics of program assignment and the degree of plagiarism.

• The KIPS Transactions:PartC
• /
• v.12C no.5 s.101
• /
• pp.649-658
• /
• 2005

Since the Network based attack Is extensive in the real state of damage, It is very important to detect intrusion quickly at the beginning. But the intrusion detection using supervised learning needs either the preprocessing enormous data or the manager's analysis. Also it has two difficulties to detect abnormal traffic that the manager's analysis might be incorrect and would miss the real time detection. In this paper, we propose a traffic attributes correlation analysis mechanism based on self-organizing maps(SOM) for the real-time intrusion detection. The proposed mechanism has three steps. First, with unsupervised learning build a map cluster composed of similar traffic. Second, label each map cluster to divide the map into normal traffic and abnormal traffic. In this step there is a rule which is created through the correlation analysis with SOM. At last, the mechanism would the process real-time detecting and updating gradually. During a lot of experiments the proposed mechanism has good performance in real-time intrusion to combine of unsupervised learning and supervised learning than that of supervised learning.

• Journal of Animal Science and Technology
• /
• v.50 no.2
• /
• pp.273-278
• /
• 2008

There have been studies about dog’s personality and behavior, which is helpful to breed dogs as guide or companion. In this study, a questionnaire was developed using 54 Jindo dogs, which considered ten items about aggressiveness and sociability. The scores were analyzed by principle component analysis (PCA), after accounting for four variables: age, gender, growing place, and coat-colors. Our results from the PCA indicated three principle components, which classified ‘aggressiveness’, ‘sociability’ and unknown factor. The four variables did not significantly affect aggressiveness(P>0.05). However, there was a relationship between coat-color and sociability, i.e., the Jindo dogs with fawn color were more sociable than the white ones(P<0.1).

• Smart Media Journal
• /
• v.11 no.2
• /
• pp.18-24
• /
• 2022

Problems caused by malicious comments occur on many social media. In particular, YouTube, which has a strong character as a medium, is getting more and more harmful from malicious comments due to its easy accessibility using mobile devices. In this paper, we designed and implemented a YouTube malicious comment detection system to identify malicious comments in YouTube contents through LSTM-based natural language processing and to visually display the percentage of malicious comments, such commentors' nicknames and their frequency, and we evaluated the performance of the system. By using a dataset of about 50,000 comments, malicious comments could be detected with an accuracy of about 92%. Therefore, it is expected that this system can solve the social problems caused by malicious comments that many YouTubers faced by automatically generating malicious comments statistics.