• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 추계학술발표대회
• /
• pp.61-63
• /
• 2021

Machine learning model gained the popularity in recent years as multi-national companies have incorporated machine learning in their services. Such service is called machine learning as a service (MLaSS). Such services are provided to users based on charge-per-query which triggers the motivations for adversaries to steal the trained victim model to reduce the cost of using the service. Therefore, it is important for companies that provide MLaSS to protect their intellectual property (IP) against adversaries. It has been arms race between the attack and defence in a context of the privacy of machine learning models. In this paper, we provide a comprehensive study of recent development in protecting privacy of machine learning models.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2022년도 학술발표회
• /
• pp.143-143
• /
• 2022

Precipitation plays an essential role in water resources management and disaster prevention. Therefore, the understanding related to spatiotemporal characteristics of rainfall is necessary. Nowadays, highly accurate precipitation is mainly obtained from gauge observation systems. However, the density of gauge stations is a sparse and uneven distribution in mountainous areas. With the proliferation of technology, satellite-based precipitation sources are becoming increasingly common and can provide rainfall information in regions with complex topography. Nevertheless, satellite-based data is that it still remains uncertain. To overcome the above limitation, this study aims to take the strengthens of machine learning to generate a new reanalysis of precipitation data by fusion of multiple satellite precipitation products (SPPs) with gauge observation data. Several machine learning algorithms (i.e., Random Forest, Support Vector Regression, and Artificial Neural Network) have been adopted. To investigate the robustness of the new reanalysis product, observed data were collected to evaluate the accuracy of the products through Kling-Gupta efficiency (KGE), probability of detection (POD), false alarm rate (FAR), and critical success index (CSI). As a result, the new precipitation generated through the machine learning model showed higher accuracy than original satellite rainfall products, and its spatiotemporal variability was better reflected than others. Thus, reanalysis of satellite precipitation product based on machine learning can be useful source input data for hydrological simulations in ungauged river basins.

• 한국공간구조학회논문집
• /
• 제24권2호
• /
• pp.83-90
• /
• 2024

Machine learning is widely applied to various engineering fields. In structural engineering area, machine learning is generally used to predict structural responses of building structures. The aging deterioration of reinforced concrete structure affects its structural behavior. Therefore, the aging deterioration of R.C. structure should be consider to exactly predict seismic responses of the structure. In this study, the machine learning based seismic response prediction model was developed. To this end, four machine learning algorithms were employed and prediction performance of each algorithm was compared. A 3-story coupled shear wall structure was selected as an example structure for numerical simulation. Artificial ground motions were generated based on domestic site characteristics. Elastic modulus, damping ratio and density were changed to considering concrete degradation due to chloride penetration and carbonation, etc. Various intensity measures were used input parameters of the training database. Performance evaluation was performed using metrics like root mean square error, mean square error, mean absolute error, and coefficient of determination. The optimization of hyperparameters was achieved through k-fold cross-validation and grid search techniques. The analysis results show that neural networks and extreme gradient boosting algorithms present good prediction performance.

• 한국멀티미디어학회논문지
• /
• 제24권4호
• /
• pp.528-537
• /
• 2021

Face recognition has gained significant notice because of its application in many businesses: security, healthcare, and marketing. In this paper, we will present the recognition method using the combination of correlation filters (CF) and Support Vector Machine (SVM). Firstly, we evaluate the performance and compared four different correlation filters: minimum average correlation energy (MACE), maximum average correlation height (MACH), unconstrained minimum average correlation energy (UMACE), and optimal-tradeoff (OT). Secondly, we propose the machine learning approach by using the OT correlation filter for features extraction and SVM for classification. The numerical results on National Cheng Kung University (NCKU) and Pointing'04 face database show that the proposed method OT-SVM gets higher accuracy in face recognition compared to other machine learning methods. Our approach doesn't require graphics card to train the image. As a result, it could run well on a low hardware system like an embedded system.

• East Asian mathematical journal
• /
• 제37권3호
• /
• pp.333-354
• /
• 2021

A support vector machine (SVM) is a state-of-the-art machine learning model rooted in structural risk minimization. SVM is underestimated with regards to its application to real world problems because of the difficulties associated with its use. We aim at showing that the performance of SVM highly depends on which kernel function to use. To achieve these, after providing a summary of support vector machines and kernel function, we constructed experiments with various benchmark datasets to compare the performance of various kernel functions. For evaluating the performance of SVM, the F1-score and its Standard Deviation with 10-cross validation was used. Furthermore, we used taylor diagrams to reveal the difference between kernels. Finally, we provided Python codes for all our experiments to enable re-implementation of the experiments.

• 한국컴퓨터정보학회논문지
• /
• 제26권4호
• /
• pp.93-103
• /
• 2021

금융당국의 채권추심 가이드라인, 추심업자에 대한 직접적인 관리 감독 수행 등의 노력에도 불구하고 채무자에 대한 불법, 부당한 채권 추심은 지속되고 있다. 이러한 불법, 부당한 채권추심행위를 효과적으로 예방하기 위해서는 비정형데이터 기계학습 등 기술을 활용하여 적은 인력으로도 불법 추심행위에 대한 점검 등에 대한 모니터링을 강화 할 수 있는 방법이 필요하다. 본 연구에서는 대부업체의 추심 녹취 파일을 입수하여 이를 텍스트 데이터로 변환하고 위법, 위규 행위를 판별하는 규칙기반 검출과 SVM(Support Vector Machine) 등 기계학습을 결합한 불법채권추심 분류 모델을 제안하고 기계학습 알고리즘에 따라 얼마나 정확한 식별을 하였는지를 비교해 보았다. 본 연구는 규칙기반 불법 검출과 기계학습을 결합하여 분류에 활용할 경우 기존에 연구된 기계학습만을 적용한 분류모델 보다 정확도가 우수하다는 것을 보여 주었다. 본 연구는 규칙기반 불법검출과 기계학습을 결합하여 불법여부를 분류한 최초의 시도이며 후행연구를 진행하여 모델의 완성도를 높인다면 불법채권 추심행위에 대한 소비자 피해 예방에 크게 기여할 수 있을 것이다.

• Structural Engineering and Mechanics
• /
• 제66권3호
• /
• pp.399-409
• /
• 2018

This paper demonstrates the potential application of machine learning algorithms for approximate prediction of the load and deflection capacities of the novel type of Laced Steel Concrete-Composite (LSCC) beams proposed by Anandavalli et al. (Engineering Structures 2012). Initially, global and local responses measured on LSCC beam specimen in an experiment are used to validate nonlinear FE model of the LSCC beams. The data for the machine learning algorithms is then generated using validated FE model for a range of values of the identified sensitive parameters. The performance of four well-known machine learning algorithms, viz., Support Vector Regression (SVR), Minimax Probability Machine Regression (MPMR), Relevance Vector Machine (RVM) and Multigene Genetic Programing (MGGP) for the approximate estimation of the load and deflection capacities are compared in terms of well-defined error indices. Through relative comparison of the estimated values, it is demonstrated that the algorithms explored in the present study provide a good alternative to expensive experimental testing and sophisticated numerical simulation of the response of LSCC beams. The load carrying and displacement capacity of the LSCC was predicted well by MGGP and MPMR, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권1호
• /
• pp.451-465
• /
• 2017

This paper presents a novel discriminative visual tracking algorithm with an adaptive incremental extreme learning machine. The parameters for an adaptive incremental extreme learning machine are initialized at the first frame with a target that is manually assigned. At each frame, the training samples are collected and random Haar-like features are extracted. The proposed tracker updates the overall output weights for each frame, and the updated tracker is used to estimate the new location of the target in the next frame. The adaptive learning rate for the update of the overall output weights is estimated by using the confidence of the predicted target location at the current frame. Our experimental results indicate that the proposed tracker can manage various difficulties and can achieve better performance than other state-of-the-art trackers.

• 반도체디스플레이기술학회지
• /
• 제20권3호
• /
• pp.131-135
• /
• 2021

Numbers of studies related to optimization of design of organic light emitting diodes(OLED) through machine learning are increasing. We propose the generative method of the image to assess the performance of the device combining with machine learning technique. Principle parameter regarding dark spot growth mechanism of the OLED can be the key factor to determine the long-time performance. Captured images from actual device and randomly generated images at specific time and initial pinhole state are fed into the deep neural network system. The simulation reinforced by the machine learning technique can predict the device parameters accurately and faster. Similarly, the inverse design using multiple layer perceptron(MLP) system can infer the initial degradation factors at manufacturing with given device parameter to feedback the design of manufacturing process.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 추계학술대회
• /
• pp.326-327
• /
• 2022

본 연구는 머신러닝을 이용한 임상적 의사결정을 위한 알고리즘으로 환자를 간호하는 간호인력이 장기간 누워있는 환자를 보살힐 경우 욕창예방간호 수행에 도움을 주기 위한 시스템 개발에 활용될 욕창 분류 알고리즘이다. 머신러닝을 실시한 결과 알고리즘의 learning accuracy는 82.14%, test accuracy는 82.58%로 나타났다.