• 한국해안·해양공학회논문집
• /
• 제32권6호
• /
• pp.384-395
• /
• 2020

연안에서 천수변형에 의해 발생하는 쇄파는 표사이동, 연안류의 생성, 충격파압의 발생 등과 같은 연안역의 다양한 물리현상과 밀접한 관계를 갖고 있다. 따라서, 연안구조물의 설계 시 쇄파파고 및 쇄파수심과 같은 쇄파지표를 정확하게 예측하는 것이 중요하다. 과거부터 많은 연구자들에 의해 쇄파현상을 규명하고 예측하기 위한 많은 과학적인 노력들이 이루어져 왔다. 대표적인 쇄파에 연구들은 주로 수리모형실험을 통해 쇄파지표 예측을 위한 많은 경험식이 제안되어 왔다. 하지만, 기존의 쇄파지표에 대한 경험식은 일정한 방정식의 가정하에 자료의 통계적 분석을 통해 가정한 방정식의 계수들을 결정하고 있다. 본 논문에서는 회귀 혹은 분류문제와 관련된 다양한 연구분야에 있어서 높은 예측성능을 보여주는 대표적인 선형기반의 지도학습 머신러닝 기법을 적용하였다. 적용된 머신러닝 기법을 기반으로 기존의 쇄파에 대한 실험자료로부터 쇄파지표 예측을 위한 모델을 개발하고, 학습된 모델로부터 쇄파예측을 위한 새로운 선형식을 제시하였다. 새롭게 제안된 쇄파지표식은 단순한 선형식임에도 불구하고 기존의 경험 공식에 비해 유사한 예측성능을 보였다.

• 인터넷정보학회논문지
• /
• 제22권6호
• /
• pp.141-150
• /
• 2021

인공지능(Artificial Intelligence) 기술을 활용하여 인공지능 기반의 전쟁 (AI-enabled warfare)가 미래전의 핵심이 될 것으로 예상한다. 자연어 처리 기술은 이러한 AI 기술의 핵심 기술로 지휘관 및 참모들이 자연어로 작성된 보고서, 정보 및 첩보를 일일이 열어확인하는 부담을 줄이는데 획기적으로 기여할 수 있다. 본 논문에서는 지휘관 및 참모의 정보 처리 부담을 줄이고 신속한 지휘결심을 지원하기 위해 언어 모델 기반의 다출처 정보 통합 (Language model-based Multi-source Information Integration, LAMII) 프레임워크를 제안한다. 제안된 LAMII 프레임워크는 자기지도 학습법을 활용한 언어 모델에 기반한 표현학습과 오토인코더를 활용한 문서 통합의 핵심 단계로 구성되어 있다. 첫 번째 단계에서는, 자기지도 학습 기법을 활용하여 구조적으로 이질적인 두 문장간의 유사 관계를 식별할 수 있는 표현학습을 수행한다. 두 번째 단계에서는, 앞서 학습된 모델을 활용하여 다출처로부터 비슷한 내용 혹은 토픽을 함양하는 문서들을 발견하고 이들을 통합한다. 이 때, 중복되는 문장을 제거하기 위해 오토인코더를 활용하여 문장의 중복성을 측정한다. 본 논문의 우수성을 입증하기 위해, 우리는 언어모델들과 이의 성능을 평가할 때 활용되는 대표적인 벤치마크 셋들을 함께 활용하여 이질적인 문장간의 유사 관계를 예측의 비교 실험하였다. 실험 결과, 제안된 LAMII 프레임워크가 다른 언어 모델에 비하여 이질적인 문장 구조간의 유사 관계를 효과적으로 예측할 수 있음을 입증하였다.

• 한국인공지능학회지
• /
• 제11권4호
• /
• pp.15-20
• /
• 2023

Machine learning is comprised of supervised learning, unsupervised learning and reinforcement learning as the type of data and processing mechanism. In this paper, as input and output are unclear and it is difficult to apply the concrete modeling mathematically, reinforcement learning method are applied for crawling robot in this paper. Especially, Q-Learning is the most effective learning technique in model free reinforcement learning. This paper presents a method to implement a crawling robot that is operated by finding the most optimal crawling method through trial and error in a dynamic environment using a Q-learning algorithm. The goal is to perform reinforcement learning to find the optimal two motor angle for the best performance, and finally to maintain the most mature and stable motion about EV3 Crawling robot. In this paper, for the production of the crawling robot, it was produced using Lego Mindstorms with two motors, an ultrasonic sensor, a brick and switches, and EV3 Classroom SW are used for this implementation. By repeating 3 times learning, total 60 data are acquired, and two motor angles vs. crawling distance graph are plotted for the more understanding. Applying the Q-learning reinforcement learning algorithm, it was confirmed that the crawling robot found the optimal motor angle and operated with trained learning, and learn to know the direction for the future research.

• 한국인터넷방송통신학회논문지
• /
• 제24권1호
• /
• pp.83-88
• /
• 2024

본 연구에서는 Railroad surface 데이터를 활용하여 Semi-Supervised learning방식으로 railroad surface의 defect를 검출해내는 방안을 제안한다. Resnet50에 ImageNet으로 pretrained된 모델을 이용한다. Label이 없는 데이터에서 무작위로 데이터를 선정, 선정한 데이터에 label을 부여한 뒤 이 데이터로 모델을 학습시킨다. 학습된 모델을 이용하여 나머지 데이터의 결과값을 예측한 후, 그 예측값이 일정한 threshold보다 큰 것을 골라내고, threshold보다 큰 값들을 값이 큰 순서대로 정렬하여, 일정한 크기만큼 training data에 추가한다. 이 때, 각 class에 속할 확률이 높은 쪽으로 pseudo-labeling을 수행한다. 초기에 label이 부여된 데이터 개수에 따른 전체적인 class 분류 성능을 확인하는 실험 또한 진행하였고, 전체 training data대비 10% 미만의 labeled data로 최대 98%의 정확도를 얻는 성능을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권1호
• /
• pp.180-194
• /
• 2021

For small image registration, feature-based approaches are likely to fail as feature detectors cannot detect enough feature points from low-resolution images. The classic FFT approach's prediction accuracy is high, but the registration time can be relatively long, about several seconds to register one image pair. To achieve real-time and high-precision rigid registration for small images, we apply deep neural networks for supervised rigid transformation prediction, which directly predicts the transformation parameters. We train deep registration models with rigidly transformed CIFAR-10 images and STL-10 images, and evaluate the generalization ability of deep registration models with transformed CIFAR-10 images, STL-10 images, and randomly generated images. Experimental results show that the deep registration models we propose can achieve comparable accuracy to the classic FFT approach for small CIFAR-10 images (32×32) and our LSTM registration model takes less than 1ms to register one pair of images. For moderate size STL-10 images (96×96), FFT significantly outperforms deep registration models in terms of accuracy but is also considerably slower. Our results suggest that deep registration models have competitive advantages over conventional approaches, at least for small images.

• Journal of Advanced Marine Engineering and Technology
• /
• 제26권6호
• /
• pp.695-704
• /
• 2002

Generally, induction motor controller requires rotor speed sensor for commutation and current control, but it increases cost and size of the motor. So in these days, various researches including speed sensorless vector control have been reported and some of them have been put to practical use. In this paper a new speed estimation method using neural networks is proposed. The optimal neural network structure was tracked down by trial and error, and it was found that the 8-16-1 neural network has given correct results for the instantaneous rotor speed. Supervised learning methods, through which the neural network is trained to learn the input/output pattern presented, are typically used. The back-propagation technique is used to adjust the neural network weights during training. The rotor speed is calculated by weights and eight inputs to the neural network. Also, the proposed method has advantages such as the independency on machine parameters, the insensitivity to the load condition, and the stability in the low speed operation.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제1권1호
• /
• pp.55-60
• /
• 2012

최근 봇넷(Botnet)은 탐지 기술을 피하기 위하여 C&C(Command and Control)서버 접속시 DNS(Domain Name System) 서비스를 이용하고 있다. DNS 서비스를 이용한 비정상 행위에 대응하기 위해서 DNS 트래픽 기반의 분석 연구가 필요하다. 본 논문에서는 좀비PC의 C&C서버 도메인주소 질의와 같은 DNS트래픽 기반의 비정상 도메인 분류(Classification)를 위해서 DNS트래픽 수집 및 지도학습(Supervised Learning)에 대해 연구한다. 특히, 본 논문에서는 PCA(Principal Component Analysis) 주성분분석 기술을 통해 DNS 기반의 분류시스템에서의 효과적인 분석 성분들을 구성할 수 있다.

• 한국안전학회지
• /
• 제36권1호
• /
• pp.56-63
• /
• 2021

The narrative texts of industrial accident reports help to identify accident risk factors. They relate the accident triggers to the sequence of events and the outcomes of an accident. Particularly, a set of related keywords in the context of the narrative can represent how the accident proceeded. Previous studies on text analytics for structuring accident reports have been limited to extracting individual keywords without context. We proposed a context-based analysis using a Natural Language Processing (NLP) algorithm to remedy this shortcoming. This study aims to apply Word2Vec of the NLP algorithm to extract adjacent keywords, known as word embedding, conducted by the neural network algorithm based on supervised learning. During processing, Word2Vec is conducted by adjacent keywords in narrative texts as inputs to achieve its supervised learning; keyword weights emerge as the vectors representing the degree of neighboring among keywords. Similar keyword weights mean that the keywords are closely arranged within sentences in the narrative text. Consequently, a set of keywords that have similar weights presents similar accidents. We extracted ten accident processes containing related keywords and used them to understand the risk factors determining how an accident proceeds. This information helps identify how a checklist for an accident report should be structured.

• International Journal of Computer Science & Network Security
• /
• 제22권11호
• /
• pp.1-10
• /
• 2022

This study aims to build a model is capable of classifying the categories of hadith, according to the reliability of hadith' narrators (sahih, hassan, da'if, maudu) and according to what was attributed to the Prophet Muhammad (saying, doing, describing, reporting ) using the supervised learning algorithms, with a view to discover a relationship between these classifications, based on the outputs of this model, which might be useful to avoid the controversy and useless debate on automatic classifications of hadith, using some of the statistical methods such as chi-square, information gain and association rules. The experimental results showed that there is a relation between these classifications, most of Sahih hadiths are belong to saying class, and most of maudu hadiths are belong to reporting class. Also the best classifier had given high accuracy was MultinomialNB, it achieved higher accuracy reached up to 0.9708 %, for his ability to process high dimensional problems and identifying the most important features that are relevant to target data in training stage. Followed by LinearSVC classifier, reached up to 0.9655, and finally, KNeighborsClassifier reached up to 0.9644.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권8호
• /
• pp.2787-2800
• /
• 2022

The aging society increases emergency situations of the elderly living alone and a variety of social crimes. In order to prevent them, techniques to detect emergency situations through voice are actively researched. This study proposes CutPaste-based anomaly detection model using multi-scale feature extraction in time series streaming data. In the proposed method, an audio file is converted into a spectrogram. In this way, it is possible to use an algorithm for image data, such as CNN. After that, mutli-scale feature extraction is applied. Three images drawn from Adaptive Pooling layer that has different-sized kernels are merged. In consideration of various types of anomaly, including point anomaly, contextual anomaly, and collective anomaly, the limitations of a conventional anomaly model are improved. Finally, CutPaste-based anomaly detection is conducted. Since the model is trained through self-supervised learning, it is possible to detect a diversity of emergency situations as anomaly without labeling. Therefore, the proposed model overcomes the limitations of a conventional model that classifies only labelled emergency situations. Also, the proposed model is evaluated to have better performance than a conventional anomaly detection model.