• Geophysics and Geophysical Exploration
• /
• v.26 no.4
• /
• pp.171-184
• /
• 2023

Machine learning (ML)-based cavity detection using a large amount of survey data obtained from vehicle-mounted ground penetrating radar (GPR) has been actively studied to identify underground cavities. However, only simple image processing techniques have been used for preprocessing the ML input, and many conventional seismic and GPR data processing techniques, which have been used for decades, have not been fully exploited. In this study, based on the idea that a cavity can be identified using diffraction, we applied ML-based diffraction separation to GPR data to increase the accuracy of cavity detection using the YOLO v5 model. The original ML-based seismic diffraction separation technique was modified, and the separated diffraction image was used as the input to train the cavity detection model. The performance of the proposed method was verified using public GPR data released by the Seoul Metropolitan Government. Underground cavities and objects were more accurately detected using separated diffraction images. In the future, the proposed method can be useful in various fields in which GPR surveys are used.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.3
• /
• pp.210-218
• /
• 2002

This paper proposes an efficient separation method of the composite images by using independent component analysis(ICA) based on neural networks of the approximate learning algorithm. The Proposed learning algorithm is the fixed point(FP) algorithm based on Secant method which can be approximately computed by only the values of function for estimating the root of objective function for optimizing entropy. The secant method is an alternative of the Newton method which is essential to differentiate the function for estimating the root. It can achieve a superior property of the FP algorithm for ICA due to simplify the composite computation of differential process. The proposed algorithm has been applied to the composite signals and image generated by random mixing matrix in the 4 signal of 500-sample and the 10 images of $512{\times}512-pixel$, respectively The simulation results show that the proposed algorithm has better performance of the learning speed and the separation than those using the conventional algorithm based method. It also solved the training performances depending on initial points setting and the nonrealistic learning time for separating the large size image by using the conventional algorithm.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.577-580
• /
• 2002

자연어의 처리를 위해 반드시 필요한 형태소 분석에는 여러 가지 방법이 있으나 기본적으로 사전을 갖춘 상태에서 가장 가능성 있는 후보를 선택하는 방식을 선택한다. 이러한 방식으로는 사전이 없는 미지의 언어를 분석하기는 불가능하다. 기지의 언어라도 지속적으로 어휘가 변하는 경우나 매우 특별한 분야의 경우에는 필요로 하는 사전이 존재하지 않는다. 본 논문에서는 태그가 없는 단순 말뭉치만을 가지고 자율학습을 이용하여 한국어의 실질 형태소와 형식 형태소를 분리해내는 기법에 대하여 기술한다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.9
• /
• pp.817-821
• /
• 2001

In this paper, a new learning methodology for kernel methods that results in a sparse representation of kernel space from the training patterns for classification problems is suggested. Among the traditional algorithms of linear discriminant function, this paper shows that the relaxation procedure can obtain the maximum margin separating hyperplane of linearly separable pattern classification problem as SVM(Support Vector Machine) classifier does. The original relaxation method gives only the necessary condition of SV patterns. We suggest the sufficient condition to identify the SV patterns in the learning epoches. For sequential learning of kernel methods, extended SVM and kernel discriminant function are defined. Systematic derivation of learning algorithm is introduced. Experiment results show the new methods have the higher or equivalent performance compared to the conventional approach.

• Journal of KIISE:Software and Applications
• /
• v.33 no.12
• /
• pp.999-1007
• /
• 2006

The Backpropagation learning algorithm is known to have slow and false convergence aroused from plateau and local minima. Many substitutes for backpropagation announced so far appear to pay some trade-off for convergence speed and stability of convergence according to parameters. Here, a new algorithm is proposed, which avoids some of those problems associated with the conventional backpropagation problems, especially with local minima, and gives relatively stable and fast convergence with low storage requirement. This is the separate learning algorithm in which the upper connections, hidden-to-output, and the lower connections, input-to-hidden, separately trained. This algorithm requires less computational work than the conventional backpropagation and other improved algorithms. It is shown in various classification problems to be relatively reliable on the overall performance.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10a
• /
• pp.565-567
• /
• 2003

웹 기반 학습(WBI)은 인터넷의 빠르게 확산되어 가고 있다. 또한, 학습자의 학습효과를 높이기 위한 많은 연구와 기술들이 적용되었다. 그러나, 대부분의 웹 기반 학습 시스템은 문제은행에서 반복적인 학습내용의 제공이나 학습의 패턴을 주는 학습자료 제공자 측면에서의 일방적인 학습방법을 택하고 있다. 시스템의 개발방법에서도 웹 기반 학습시스템은 전통적인 순차적 프로세스 개발방법에 따라 구현됨으로서 개발의 비효율성 밀 재사용, 유지보수등이 어려웠다. 본 연구에서는 문제은행에서 사용되는 문제추출 컴포넌트를 설계 및 구현하였다. 문제추출은 학습자의 학습효과를 높이기 위하여 문항난이도를 분석하여 학습자에게 적절한 문제가 출제될 수 있도록 하였다. 또한, 본 연구의 시스템에서 중요한 비즈니스 로직이 되는 난이도 분석 및 적용부분은 EJB컴포넌트로 구현함으로서 구현로직의 분리 및 재사용성, 유지보수성을 높였다.

• 한국산학경영학회:학술대회논문집
• /
• 2005.11a
• /
• pp.13-25
• /
• 2005

기업의 신용평가는 기업의 위험도를 측정하여 어음, 사채 및 대출금 등의 회수 가능성을 평가하는 것이다. 이러한 기업의 신용평가 결과는 해당 기업의 채권 수익률이나 주가 등에 영향을 미치고, 또한 금융기관, 투자자 및 거래처 등이 대출 결정, 투자 결정, 신용판매 등의 의사결정을 내리는데 영향을 미친다. 본 논문에서는 보다 정확한 기업 신용 평가를 위해 다집단 분류 문제를 이집단 분류 문제화하는 신경망 분리 모형을 제안한다. 또한, 본 논문에서 제안한 신경망 분리 모형의 우수성을 검증하기 위해 기존의 일반적인 신경회로망, 판별분석 모형과 비교한다. 실험 결과 신경회로망을 분리시켜 학습을 단순화시키는 방법이 기존의 방법에 비해 우수한 결과를 보였다.

• The KIPS Transactions:PartB
• /
• v.8B no.6
• /
• pp.675-684
• /
• 2001

This paper describes a method for stemming of Korean language by using unsupervised learning from raw corpus. This technique does not require a lexicon or any language-specific knowledge. Since we use unsupervised learning, the time and effort required for learning is negligible. Unlike heuristic approaches that are theoretically ungrounded, this method is based on widely accepted statistical methods, and therefore can be easily extended. The method is currently applied only to Korean language, but it can easily be adapted to other agglutinative languages, since it is not language-dependent.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.11
• /
• pp.41-49
• /
• 2019

We are to build an unsupervised machine learning-based language model which can estimate the amount of information that are in need to process words consisting of subword-level morphemes and syllables. We are then to investigate whether the reading times of words reflecting their morphemic and syllabic structures are predicted by an information-theoretic measure such as surprisal. Specifically, the proposed Morfessor-based unsupervised machine learning model is first to be trained on the large dataset of sentences on Sejong Corpus and is then to be applied to estimate the information-theoretic measure on each word in the test data of Korean words. The reading times of the words in the test data are to be recruited from Korean Lexicon Project (KLP) Database. A comparison between the information-theoretic measures of the words in point and the corresponding reading times by using a linear mixed effect model reveals a reliable correlation between surprisal and reading time. We conclude that surprisal is positively related to the processing effort (i.e. reading time), confirming the surprisal hypothesis.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.271-274
• /
• 2002

본 논문에서는 초기 부호책 설계 방법으로 가장 널리 사용되는 이분 미소분리 방법의 성능 개선과 설계시간을 단축하기 위한 새로운 알고리즘을 제안한다. 성능 개선을 위해 학습벡터의 소속수가 최소인 부호백터를 제거하고, 최대인 부호벡터를 미소분리하여 대체하는 방법을 적용하고, 모든 부호벡터와의 거리오차론 구하여 학습벡터의 소속 여부를 결정하는 기존 방법과는 달리, 전단계와 현재 단계의 소속 부호벡터와의 거리오차를 가지고 소속 여부를 결정함으로써 설계시간을 크게 단축할 수 있다.