• Journal of Advanced Marine Engineering and Technology
• /
• v.38 no.2
• /
• pp.182-187
• /
• 2014

We propose nonlinear GMM-based transformation functions in an attempt to deal with the over-smoothing effects of linear transformation for voice processing. The proposed methods adopt RBF networks as a local transformation function to overcome the drawbacks of global nonlinear transformation functions. In order to obtain high-quality modifications of speech signals, our voice conversion is implemented using the Harmonic plus Noise Model analysis/synthesis framework. Experimental results are reported on the English corpus, MOCHA-TIMIT.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.04a
• /
• pp.403-413
• /
• 1996

Two-phase self-organizing neuro-modeling (SONM). the global SONM and local SONM, is designed for tracking non-stationary manufacturing processes. Radial basis function (RBF) neural network is employed, and self-tuning estimator is also developed for the determination of RBF network parameters on-line. A pattern recognition approach is presented for identifying a correct RBF neural network, which is used for identifying current manufacturing processes. Experimental results showed that the proposed approach is suitable for tracking non-stationary processes.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.1037-1040
• /
• 1999

This paper investigates a RBF(Radial Basis Function) equalizer for channel equalization. RBF network has an identical structure to the optimal Bayesian symbol-decision equalizer solution. Therefore RBF can be employed to implement the Bayesian equalizer. Proposed algorithm of this paper makes channel states estimation to be unncessary, also makes center number which is needed indivisual channel to be minimum. Bayesian Equalizer has the theorical optimum performance. Proposed Equalizer performance is compared with this Baysian equalizer performance.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04b
• /
• pp.478-480
• /
• 2000

본 논문에서는 데이터를 능동적으로 선택하고, 그 데이터에 맞추어 RBF 은닉 뉴런을 증가시키는 신경망을 제안한다. 현재의 신경망에 대해서 가장 학습이 어려운 데이터를 선택해서 신경망을 학습하고, 학습한 신경망에 대해서 다시 에러가 가장 큰 데이터를 뽑아서 학습시키는 과정을 반복한다. 5개의 실세계 데이터에 대해 실험을 해보고, Platt이 제안한 RAN과 성능을 비교한다. 점진적으로 임계 데이터를 선택해서 학습을 함으로써, 전체 데이터를 다 사용하지도 않고도, 전체 데이터를 다 사용한 경우와 비슷한 성능을 보임을 실험을 통해서 알 수 있다.

• Journal of the Korean Professional Engineers Association
• /
• v.42 no.3
• /
• pp.47-53
• /
• 2009

Riverbed filtration(RBF) system is used to develop ground water and infiltrated water supplies from permeable sand and gravel deposits. RBF plants are constructed with a reinforced concrete caisson that serves as a wet well pumping station. The lateral well screens are projected horizontally into waterbearing deposits from inside the caisson. Riverbed filtration(RBF) is a low-cost and efficient alternative water treatment for drinking-water applications.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2576-2578
• /
• 2003

RBF 네트워크에서 기저함수의 위치는 네트워크의 성능에 매우 큰 영향을 미친다. 몇몇 응용들에서 교사학습을 이용한 기저함수의 위치 선정이 비교사학습에 비해 우수함을 보인다. 그러나 교사학습에 의한 네트워크는 시그모이드 네트워크와 같은 긴 학습시간을 필요로 한다. 본 논문에서는 오차함수의 gradient와 Hessian을 이용해 교사학습에서 요구하는 학습시간을 단축시키면서 기저함수의 최적위치를 추정하였다.

• Interaction and multiscale mechanics
• /
• v.6 no.4
• /
• pp.395-409
• /
• 2013

Molecular dynamics (MD) systems are highly nonlinear and nonlocal, and the conventional model order reduction methods are ineffective for MD systems. The RBF-POD method (Lee and Chen, 2013) employed a radial basis function (RBF) approximated potential energies and inter-atomic forces of MD systems under the framework of the proper orthogonal decomposition (POD) method for the reduced-order modeling of MD systems. In this work, we focus on the numerical procedures of the RBF-POD method and demonstrate how to apply this approach to the modeling of ds-DNA molecules under stretching and bending conditions.

• Proceedings of the IEEK Conference
• /
• 2002.07a
• /
• pp.26-29
• /
• 2002

Radial Basis Function (RBF) networks is known as efficient method in classification problems and function approximation. The basis function of RBF networks is usual adopted normal distribution like the Gaussian function. The output of the Gaussian function has the maximum at the center and decrease as increase the distance from the center. For learning of neural network, the method treating the limited area of input space is sometimes more useful than the method treating the whole of input space. The q-normal distribution is the set of probability density function include the Gaussian function. In this paper, we introduce the RBF networks with the basis function of q-normal distribution and actually approximate a function using the RBF networks.

• Korean Journal of Computational Design and Engineering
• /
• v.21 no.4
• /
• pp.443-452
• /
• 2016

This paper proposes a method for expensive black box optimization using radial basis functions (RBFs). The proposed algorithm is a computational strategy that uses a RBF model approximating the expensive black box function to predict an optimum. First, a RBF-based approximation technique is introduced and a sampling plan for estimation of the black box function is described. Then the proposed algorithm is explained, which presents the pseudo-codes for implementation and the detailed description of each step performed in the optimization process. In addition, numerical experiments will be given to analyze the performance of the proposed algorithm, by investigating computation accuracy, number of function evaluations, and convergence history. Finally, geometric distance problem as application example will be also presented for showing the algorithm applicability to different engineering problems.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1903_1904
• /
• 2009

본 논문은 퍼지 RBF 뉴럴네트워크를 이용한 트랙킹 검출 방법을 제시한다. IEC 60112에서 규정한 실험 장치와 방법에 따라 실험을 수행하였다. NI 장비를 사용하여 전류 파형을 측정하고, 측정된 전류 파형으로부터 FFT, 웨이블렛등의 신호처리 기법을 사용하여 12개의 특징점을 추출한다. 추출된 특징점들을 퍼지 RBF 뉴럴네트워크의 입력으로 사용하여 트랙킹 발생 유무를 검출한다. 퍼지 RBF 뉴럴네트워크는 WLSE를 사용하여 학습하고, HFC-PGA를 이용하여 특징점들의 선택, 퍼지 규칙의 수, 후반부 다항식 차수, 퍼지화 계수등을 최적화 하였다.