• ETRI Journal
• /
• 제41권3호
• /
• 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.

• 전기전자학회논문지
• /
• 제26권3호
• /
• pp.475-482
• /
• 2022

Signal pre-processing and post-processing are some areas of study around electrical resistance tomography due to the low spatial resolution of pixel-based reconstructed images. In addition, methods that improve integrity and noise reduction are candidates for application in ERT. Lately, formulations of image processing methods provide new implementations and studies to improve the response against noise. For example, compact variational mode decomposition has recently shown good performance in image decomposition and segmentation. The results from this first approach of C-VMD to ERT show an improvement due to image segmentation, providing filtering of noise in the background and location of the target.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제26권3호
• /
• pp.156-184
• /
• 2022

In this article, we propose a novel variational model for restoring color images corrupted by mixed multiplicative Gamma noise and additive Gaussian noise. The model involves a data-fidelity term that characterizes the mixed noise as an infimal convolution of two noise distributions and the saturation-value total variation (SVTV) regularization. The data-fidelity term facilitates suitable separation of the multiplicative Gamma and Gaussian noise components, promoting simultaneous elimination of the mixed noise. Furthermore, the SVTV regularization enables adequate denoising of homogeneous regions, while maintaining edges and details and diminishing the color artifacts induced by noise. To solve the proposed nonconvex model, we exploit an alternating minimization approach, and then the alternating direction method of multipliers is adopted for solving subproblems. This contributes to an efficient iterative algorithm. The experimental results demonstrate the superior performance of the proposed model compared to other existing or related models, with regard to visual inspection and image quality measurements.

• ETRI Journal
• /
• 제45권4호
• /
• pp.678-689
• /
• 2023

This paper describes a novel end-to-end deep generative model-based speaker recognition system using prosodic features. The usefulness of variational autoencoders (VAE) in learning the speaker-specific prosody representations for the speaker recognition task is examined herein for the first time. The speech signal is first automatically segmented into syllable-like units using vowel onset points (VOP) and energy valleys. Prosodic features, such as the dynamics of duration, energy, and fundamental frequency (F₀), are then extracted at the syllable level and used to train/adapt a speaker-dependent VAE from a universal VAE. The initial comparative studies on VAEs and traditional autoencoders (AE) suggest that the former can efficiently learn speaker representations. Investigations on the impact of gender information in speaker recognition also point out that gender-dependent impostor banks lead to higher accuracies. Finally, the evaluation on the NIST SRE 2010 dataset demonstrates the usefulness of the proposed approach for speaker recognition.

• Structural Engineering and Mechanics
• /
• 제59권4호
• /
• pp.671-682
• /
• 2016

In this paper, a new analytical approach has been presented for solving nonlinear conservative oscillators. Variational approach leads us to high accurate solution with only one iteration. Two different high nonlinear examples are also presented to show the application and accuracy of the presented approach. The results are compared with numerical solution using runge-kutta algorithm in different figures and tables. It has been shown that the variatioanl approach doesn't need any small perturbation and is accurate for nonlinear conservative equations.

• 한국광학회지
• /
• 제13권1호
• /
• pp.51-57
• /
• 2002

Buried 채널 도파로의 필드 분포에 대한 analytic 표현식을 effective index method(EIM)와, variational method(VM) 사용하여 구한 뒤 angular spectrum방법을 적용하여 도파로 폭과 두께에 따른 최소의 반사율을 주는 코팅층의 최적 굴절율과 정규화된 최적 코팅 두께를 구하였고 이를 비교 검토하였다. 도파로 폭이 큰 영역에서는 두 방법으로 구한 코팅층의 파라메타가 비슷한 결과를 보였으나 도파로 폭이 작아질수록 VM을 사용하여 구한 코팅층의 최적 굴절율과 정규화된 최적 두께는 클래딩층만으로 구성된 물질과 공기사이에 존재하는 코팅층의 최적 굴절율과 정규화된 최적 두께로 접근한 반면에 EIM을 사용하여 구한 경우는 차이가 많이 발생하였다. Buried채널 도파로의 도파로 폭과 두께가 같을 때, 활성층과 클래딩층의 굴절율 차이의 비에 관계없이 VM을 사용하여 구한 quasi-TE모드와 quasi-TM모드의 공차지도는 거의 동일한 영역에 존재하였다. 반면에 free space radiation mode(FSRM) 방법의 경우는 활성층과 클래딩층의 굴절율 차이의 비가 10%일 때, quasi-TE 모드와 quasi-TM모드의 공차지도는 서로 다른 영역에 존재하였다. 따라서 VM과 angular spectrum방법을 사용하여 구한 공차지도가 FSRM 방법을 사용하여 구한 공차지도보다 정확함을 알 수 있었다.

• 대한수학회보
• /
• 제50권1호
• /
• pp.57-71
• /
• 2013

In this paper, we establish the existence of at least three solutions to a Navier boundary problem involving the ($p_1$, ${\cdots}$, $p_n$)-biharmonic systems. We use a variational approach based on a three critical points theorem due to Ricceri [B. Ricceri, A three critical points theorem revisited, Nonlinear Anal. 70 (2009), 3084-3089].

• 대한수학회논문집
• /
• 제25권2호
• /
• pp.313-325
• /
• 2010

General models for the relaxed proximal point algorithm using the notion of relative maximal accretiveness (RMA) are developed, and then the convergence analysis for these models in the context of solving a general class of nonlinear inclusion problems differs significantly than that of Rockafellar (1976), where the local Lipschitz continuity at zero is adopted instead. Moreover, our approach not only generalizes convergence results to real Banach space settings, but also provides a suitable alternative to other problems arising from other related fields.

• Journal of applied mathematics & informatics
• /
• 제19권1_2호
• /
• pp.19-32
• /
• 2005

It is shown that the problem of minimizing (maximizing) a quadratic cost functional (quadratic gain functional) given the dynamics dx = (fx + gu)dt + hdb(t, a) where b(t, a) is a fractional Brownian motion of order a, 0 < 2a < 1, can be solved completely (and meaningfully!) by using the dynamical equations of the moments of x(t). The key is to use fractional Taylor's series to obtain a relation between differential and differential of fractional order.

• Journal of applied mathematics & informatics
• /
• 제36권5_6호
• /
• pp.491-504
• /
• 2018

In this article, we consider a fourth order nonlinear difference system. By making use of the critical point theory, we obtain some new existence theorems of at least one periodic solution with minimal period. Our main approach used in this article is the variational technique and the Saddle Point Theorem.