• 말소리와 음성과학
• /
• 제7권3호
• /
• pp.131-138
• /
• 2015

In this paper, we propose L1-norm regularization for state vector adaptation of subspace Gaussian mixture model (SGMM). When you design a speaker adaptation system with GMM-HMM acoustic model, MAP is the most typical technique to be considered. However, in MAP adaptation procedure, large number of parameters should be updated simultaneously. We can adopt sparse adaptation such as L1-norm regularization or sparse MAP to cope with that, but the performance of sparse adaptation is not good as MAP adaptation. However, SGMM does not suffer a lot from sparse adaptation as GMM-HMM because each Gaussian mean vector in SGMM is defined as a weighted sum of basis vectors, which is much robust to the fluctuation of parameters. Since there are only a few adaptation techniques appropriate for SGMM, our proposed method could be powerful especially when the number of adaptation data is limited. Experimental results show that error reduction rate of the proposed method is better than the result of MAP adaptation of SGMM, even with small adaptation data.

• Journal of Information Processing Systems
• /
• 제11권3호
• /
• pp.465-482
• /
• 2015

IEEE 802.11p is a standard MAC protocol for wireless access in vehicular environments (WAVEs). If a packet collision happens when a safety message is sent out, IEEE 802.11p chooses a random back-off counter value in a fixed-size contention window. However, depending on the random choice of back-off counter value, it is still possible that less important messages are sent out first while more important messages are delayed longer until sent out. In this paper, we present a new scheme for safety message scheduling, called the enhanced message priority mechanism (EMPM). It consists of the following two components: the benefit-value algorithm, which calculates the priority of the messages depending on the speed, deceleration, and message lifetime; and the back-off counter selection algorithm, which chooses the non-uniform back-off counter value in order to reduce the collision probability and to enhance the throughput of the highly beneficial messages. Numerical results show that the EMPM can significantly improve the throughput and delay of messages with high benefits when compared with existing MAC protocols. Consequently, the EMPM can provide better QoS support for the more important and urgent messages.

• Journal of Magnetics
• /
• 제21권4호
• /
• pp.660-665
• /
• 2016

The electromagnetic apparatus plays an important role in high power electrical systems. It is of great importance to provide an effective approach for the optimization of the high power electromagnetic apparatus. However, premature convergence and few Pareto solution set of the optimization for electromagnetic apparatus always happen. This paper proposed a modified multi-objective particle swarm optimization algorithm based on the niche sorting strategy. Applying to the modified algorithm, this paper guarantee the better Pareto optimal front with an enhanced distribution. Aiming at shortcomings in the closing bounce and slow breaking velocity of electromagnetic apparatus, the multi-objective optimization model was established on the basis of the traditional optimization. Besides, by means of the improved multi-objective particle swarm optimization algorithm, this paper processed the model and obtained a series of optimized parameters (decision variables). Compared with other different classical algorithms, the modified algorithm has a satisfactory performance in the multi-objective optimization problems in the electromagnetic apparatus.

• 한국표면공학회지
• /
• 제52권2호
• /
• pp.53-57
• /
• 2019

Stability and activity of boron doped diamond (BDD) electrode are key factors for water treatment. In this study, BDD electrodes were prepared on various substrates such as Nb, Si, Ti, and $TiN_x/Ti$ by hot filament chemical vapor deposition (HFCVD) method. BDD/Ti film showed the delamination between BDD and Ti substrate due to the formation of TiC layer caused by diffusion of carbon. On the other hand, $BDD/TiN_x/Ti$ showed remarkably improved stability, compared to BDD/Ti. It was confirmed that $TiN_x$ intermediate layer act as barrier layer for diffusion of carbon. High potential window of 2.8 eV was maintained on the $BDD/TiN_x/Ti$ electrode and, better wastewater treatment capability and longer electrode working life than BDD/Nb, BDD/Si and BDD/Ti were obtained.

• Journal of Medicine and Life Science
• /
• 제15권2호
• /
• pp.56-61
• /
• 2018

NPWT는 90년대 이후 많은 연구가 이루어지며 발전해왔다. 이 기술은 사용하기에 간단하고 다양한 상처에 적용할 수 있고 효용성과 비용-효과가 뛰어나며 부작용이 적다. 따라서 NPWT를 일반적인 드레싱 대신에 사용하면 육아조직의 빠른 성장과 상처 수축을 통해 피부 이식이나 피판이 필요한 부위를 줄일 수 있고 봉합이 불가했던 부위를 가능하게 하므로 창상 환자에 대한 의료 비용을 감소시키고 만성 창상 환자가 입원하지 않고 움직이면서 창상 관리가 가능하도록 할 수 있는 유용한 방법이다.

• Smart Structures and Systems
• /
• 제30권3호
• /
• pp.303-315
• /
• 2022

The degradation of wheel tread may result in serious hazards in the railway operation system. Therefore, timely wheel defect diagnosis of in-service trains to avoid tragic events is of particular importance. The focus of this study is to develop a novel wheel defect detection approach based on the relevance vector machine (RVM) which enables online detection of potentially defective wheels with trackside monitoring data acquired under different running-speed conditions. With the dynamic strain responses collected by a trackside monitoring system, the cumulative Fourier amplitudes (CFA) characterizing the effect of individual wheels are extracted to formulate multiple probabilistic regression models (MPRMs) in terms of multi-kernel RVM, which accommodate both variables of vibration frequency and running speed. Compared with the general single-kernel RVM-based model, the proposed multi-kernel MPRM approach bears better local and global representation ability and generalization performance, which are prerequisite for reliable wheel defect detection by means of data acquired under different running-speed conditions. After formulating the MPRMs, we adopt a Bayesian null hypothesis indicator for wheel defect identification and quantification, and the proposed method is demonstrated by utilizing real-world monitoring data acquired by an FBG-based trackside monitoring system deployed on a high-speed trial railway. The results testify the validity of the proposed method for wheel defect detection under different running-speed conditions.

• Journal of Microbiology and Biotechnology
• /
• 제32권9호
• /
• pp.1209-1216
• /
• 2022

To better understand the effects of PEGylation and biotinylation on the delivery efficiency of proteins, the cationic protein lysozyme (LZ) and anionic protein bovine serum albumin (BSA) were chemically conjugated with poly(ethylene glycol) (PEG) and biotin-PEG to primary amine groups of proteins using N-hydroxysuccinimide reactions. Four types of protein conjugates were successfully prepared: PEGylated LZ (PEG-LZ), PEGylated BSA (PEG-BSA), biotin-PEG-conjugated LZ (Bio-PEG-LZ), and biotin-PEG-conjugated BSA (Bio-PEG-BSA). PEG-LZ and Bio-PEG-LZ exhibited a lower intracellular uptake than that of LZ in A549 human lung cancer cells (in a two-dimensional culture). However, Bio-PEG-BSA showed significantly improved intracellular delivery as compared to that of PEG-BSA and BSA, probably because of favorable interactions with cells via biotin receptors. For A549/fibroblast coculture spheroids, PEG-LZ and PEG-BSA exhibited significantly decreased tissue penetration as compared with that of unmodified proteins. However, Bio-PEG-BSA showed tissue penetration comparable to that of unmodified BSA. In addition, citraconlyated LZ (Cit-LZ) showed reduced spheroid penetration as compared to that of LZ, probably owing to a decrease in protein charge. Taken together, chemical conjugation of targeting ligands-PEG to anionic proteins could be a promising strategy to improve intracellular delivery and in vivo activity, whereas modifications of cationic proteins should be more delicately designed.

• 산경연구논집
• /
• 제14권5호
• /
• pp.1-9
• /
• 2023

Purpose: Visual design has many prospective applications, including attaining business, social, educational, and political goals. The purpose of this study is to gain an understanding of how visual design influences consumer-facing brand values and marketing messages. In order to be successful in marketing, one must promptly attract the attention of the clients one is attempting to sell to. Research design, data and methodology: In order for the research to provide useful information, it will be necessary to sort the findings in descending order of importance and construct an integrated whole out of the knowledge gleaned from various books, papers, and other sources. This method is essential for establishing the degree of objectivity present in each piece of research and combining the results of those studies. Results: Visual marketing promotes a product or service through moving or still images, illustrations, logos, and other forms of digital content. The audience may comprehend your message better if you include visuals. Even if people only scan your content, they may be more likely to notice it if it is presented appealingly. Conclusions: In conclusion, this study found that visual design substantially influences both the demand for products and the awareness of brand values within their target audiences

• 지능정보연구
• /
• 제21권3호
• /
• pp.19-36
• /
• 2015

대표적 소셜미디어인 유튜브는 기존 폐쇄형 콘텐츠 서비스와는 다르게 개방형 콘텐츠 서비스로 이용자들의 참여와 공유를 통하여 많은 인기를 유지하고 있다. 콘텐츠 산업에서 중요한 위치를 차지하고 있는 유투브 상의 콘텐츠 확산 요인에 관한 기존의 연구들은 댓글 수 등과 같은 일반적 정보 특성 요인과 조회 수 간에 상관관계 등을 분석하는 것이 대부분이었다. 최근 네트워크 구조를 기반으로 한 연구들도 진행되었으나 대부분 콘텐츠를 이용하는 대상인 구독자나 지인 등을 중심으로 한 인적 관계 네트워크 구조 연구가 대부분이었다. 이에 본 연구에서는 실질적인 콘텐츠를 중심으로 한 네트워크 구조와 일반요인을 통합한 모델을 제시하고 확산요인을 분석하고자 한다. 이를 위해 통합 모델 인과관계 분석과 함께 21,307개의 유튜브 콘텐츠를 콘텐츠 기반 네트워크 구조로 분석하였다. 본 연구를 통해 기존에 알려진 일반적 요인과 네트워크 요인들이 모두 조회수에 영향을 주는 인과관계를 통계적으로 재검증하였으며 통합적으로는 등록자의 구독자 수, 경과시간, 매개 중심성, 댓글 수, 근접 중심성, 클러스터링 계수, 평균 평점 순으로 조회 수에 긍정적인 영향을 미치는 것으로 분석되었다. 하지만 네트워크 요인중 연결정도 중심성과 고유벡터 중심성은 부정적 영향을 주는 것으로 분석되었다. 본 연구를 통하여 유튜브 콘텐츠 확산에 대한 일반영향요인과 구조적인 현상을 함께 규명하였다. 본 연구는 기업들이 유튜브와 같은 콘텐츠 서비스를 통한 온라인 마케팅 활동 시 콘텐츠들의 구조적인 면을 고려할 수 있는 근거를 제공하였으며 음반산업의 수요예측이나 콘텐츠 제작 업체들의 원활한 서비스 제공을 위한 설명력있는 영향요인 및 모델이 될 수 있을 것이다.

• Corrosion Science and Technology
• /
• 제4권3호
• /
• pp.89-94
• /
• 2005

This paper presents the results of a study that was undertaken to optimize the ratio of the components of a new multi-component inhibitor blend composed of orthophosphate/ phosphonates/ acrylate copolymer/ isothiazolone. The effects of newly developed inhibitor on carbon steel dissolution in synthetic cooling water were studied through weight loss tests, electrochemical tests, scale tests, and micro-organism tests. The obtained results were compared to blank (uninhibited specimen) and showed that developed inhibitor revealed very good corrosion, scale, and micro-organism inhibition simultaneously. All measurements indicated that the efficiency of the blended mixture exceeded 90 %. The inhibitive effects arose from formation of protective films which might contain calcium phosphate, calcium phosphonate, and iron oxide. The nature of protective films formed on the carbon steel was studied by scanning electron microscopy (SEM) and auger electron spe ctroscopy (AES). Inhibitor used in this study appeared to have better performance for scale inhibition due to their superior crystal modification effect and excellent calcium carbonate scale inhibition properties. The effect of inhibitor on microorganisms was evaluated through minimum inhibitory concentration (MIC) test. All kinds of micro-organisms used in this study were inhibited under 78ppm concentration of inhibitor.