• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.619-621
• /
• 2005

은닉 마코프 모델은 이산 동역학을 표현할 수 있는 확률 모형이다. 우도 함수 최적화를 수행하는 전통적인 Baum-Welch 학습 알고리즘은 국소해로 수령하기 쉬우며, 우도함수의 특성상 복잡한 모델을 선호하는 바이어스가 존재한다. 베이지안 프레임워크에서는 파라미터를 랜덤 변수로 보고 이에 대한 사후 확률 분포를 추정하여 이 문제를 해결할 수 있다. 본 논문에서는 베이지안 추정을 위한 결정론적 근사화 기법인 변분 베이지안 방법을 이용, 출력 노드에 가우시안 혼합 노드를 지니는 일반화된 HMM의 추론 방법을 유도한다. 인공 데이터에 대한 실험을 통해, 본 방법이 효과적인 HMM 학습을 수행할 수 있음을 보인다.

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.928-940
• /
• 2017

State-of-the-art speaker recognition systems may work better for the English language. However, if the same system is used for recognizing those who speak different languages, the systems may yield a poor performance. In this work, the decisions of a Gaussian mixture model-universal background model (GMM-UBM) and a learning vector quantization (LVQ) are combined to improve the recognition performance of a multilingual speaker identification system. The difference between these classifiers is in their modeling techniques. The former one is based on probabilistic approach and the latter one is based on the fine-tuning of neurons. Since the approaches are different, each modeling technique identifies different sets of speakers for the same database set. Therefore, the decisions of the classifiers may be used to improve the performance. In this study, multitaper mel-frequency cepstral coefficients (MFCCs) are used as the features and the monolingual and cross-lingual speaker identification studies are conducted using NIST-2003 and our own database. The experimental results show that the combined system improves the performance by nearly 10% compared with that of the individual classifier.

• ETRI Journal
• /
• v.41 no.3
• /
• pp.298-307
• /
• 2019

This paper investigates the use of the inverse-free sparse Bayesian learning (SBL) approach for peak-to-average power ratio (PAPR) reduction in orthogonal frequency-division multiplexing (OFDM)-based multiuser massive multiple-input multiple-output (MIMO) systems. The Bayesian inference method employs a truncated Gaussian mixture prior for the sought-after low-PAPR signal. To learn the prior signal, associated hyperparameters and underlying statistical parameters, we use the variational expectation-maximization (EM) iterative algorithm. The matrix inversion involved in the expectation step (E-step) is averted by invoking a relaxed evidence lower bound (relaxed-ELBO). The resulting inverse-free SBL algorithm has a much lower complexity than the standard SBL algorithm. Numerical experiments confirm the substantial improvement over existing methods in terms of PAPR reduction for different MIMO configurations.

• International Journal of Contents
• /
• v.14 no.4
• /
• pp.24-29
• /
• 2018

EEG-based brain-computer interfaces has focused on explicitly expressed intentions to assist physically impaired patients. For EEG-based-computer interfaces to function effectively, it should be able to understand users' implicit information. Since it is hard to gather EEG signals of human brains, we do not have enough training data which are essential for proper classification performance of implicit intention. In this paper, we improve the subject independent classification of implicit intention through the generation of additional training data. In the first stage, we perform the PCA (principal component analysis) of training data in a bid to remove redundant components in the components within the input data. After the dimension reduction by PCA, we train ICA (independent component analysis) network whose outputs are statistically independent. We can get additional training data by adding Gaussian noises to ICA outputs and projecting them to input data domain. Through simulations with EEG data provided by CNSL, KAIST, we improve the classification performance from 65.05% to 66.69% with Gamma components. The proposed sample generation method can be applied to any machine learning problem with fewer samples.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.330-333
• /
• 2019

본 연구는 기존 관중 수 예측에 주로 사용되는 ARIMA 모형과 다른 GKFN(Network with Gaussian kernel functions) 모델을 시계열 모델로 제안하고 여러 변수 간의 상관관계를 분석한 MLP(Multilayer Perceptron) 모델을 각각 따로 만들어 두 가지 RMSE값의 가중치를 결합한 새로운 모델을 최종적으로 제안한다. GKFN 모델은 phase space 분석을 위해 smoothness measure를 측정하고 커널 개수를 늘려가며 학습시키는 방법이다. 또한, MLP 모델은 관중 수에 영향을 주는 여러 변수(날짜, 날씨 등 팀과 관련된 특징들)의 상관관계를 correlation coefficient 값을 이용해 분석하고 높은 상관관계를 가지는 변수들을 이용해 MLP 모델을 만들어 학습하는 것이다. 이를 통해 프로야구팀 기아 타이거즈의 일일 단위 관중 수를 예측하고자 하였다. 관중 수 예측을 통해 구단과 관객 모두 긍정적인 활용이 가능할 것이다. 훈련 자료는 2010년부터 2018년까지 9년 동안 기아 타이거즈의 일별 관중 수를 자료로 하였다.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.25 no.1
• /
• pp.72-81
• /
• 2022

Maritime patrol aircraft is an efficient solution for detecting submarines at sea. The aircraft can only detect submarines by sonobuoy, but the number of buoy is limited. In this paper, we present the online sonobuoy deployment method for estimating the location of submarines. We use Gaussian process regression to estimate the submarine existence probability map, and Bayesian optimization to decide the next best position of sonobuoy. Further, we show the performance of the proposed method by simulation.

• Nuclear Engineering and Technology
• /
• v.55 no.10
• /
• pp.3716-3724
• /
• 2023

Loose parts monitoring and detecting alarm type in real Nuclear Power Plant have challenges such as background noise, insufficient alarm data, and difficulty of distinction between alarm data that occur during start and stop. Although many signal processing methods and alarm determination algorithms have been developed, it is not easy to determine valid alarm and extract the meaning data from alarm signal including background noise. To address these issues, this paper proposes a denoising autoencoder-based majority vote classification. Training and test data are prepared by acquiring alarm data from real NPP and simulation facility for data augmentation, and noisy data is reproduced by adding Gaussian noise. Using DAEs with 3, 5, 7, and 9 layers, features are extracted for each model and classified into neural networks. Finally, the results obtained from each DAE are classified by majority voting. Also, through comparison with other methods, the accuracy and the false alarm rate are compared, and the excellence of the proposed method is confirmed.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.1
• /
• pp.9-18
• /
• 2023

Predicting the compressive strength of high-performance concrete (HPC) is challenging because of the use of additional cementitious materials; thus, the development of improved predictive models is essential. The purpose of this study was to develop an HPC compressive-strength prediction model using an ensemble machine-learning method of combined bagging and stacking techniques. The result is a new ensemble technique that integrates the existing ensemble methods of bagging and stacking to solve the problems of a single machine-learning model and improve the prediction performance of the model. The nonlinear regression, support vector machine, artificial neural network, and Gaussian process regression approaches were used as single machine-learning methods and bagging and stacking techniques as ensemble machine-learning methods. As a result, the model of the proposed method showed improved accuracy results compared with single machine-learning models, an individual bagging technique model, and a stacking technique model. This was confirmed through a comparison of four representative performance indicators, verifying the effectiveness of the method.

• Journal of Satellite, Information and Communications
• /
• v.12 no.3
• /
• pp.74-79
• /
• 2017

In this paper, we propose a system model that predicts the original signal transmitted from the transmitter using the received signal in the power line communication system based on the multi - layer perceptron which is one of the machine learning algorithms. Power line communication system using communication system using power network has more noise than communication system using general communication line. It causes a problem that the performance of the power line communication system is degraded. In order to solve this problem, the communication system model proposed in this paper minimizes the influence of noise through original signal prediction and mitigates the performance degradation of the power line communication system. In this paper, we prove that the original signal is predicted by applying the proposed communication system model to the white noise environment.

• Journal of the Korea Convergence Society
• /
• v.8 no.10
• /
• pp.45-52
• /
• 2017

Drifter is an equipment for observing the characteristics of seawater in the ocean, and it can be used to predict effluent oil diffusion and to observe ocean currents. In this paper, we design models or the prediction of drifter trajectory using machine learning. We propose methods for estimating the trajectory of drifter using support vector regression, radial basis function network, Gaussian process, multilayer perceptron, and recurrent neural network. When the propose mothods were compared with the existing MOHID numerical model, performance was improve on three of the four cases. In particular, LSTM, the best performed method, showed the imporvement by 47.59% Future work will improve the accuracy by weighting using bagging and boosting.