• The Journal of the Convergence on Culture Technology
• /
• v.7 no.1
• /
• pp.640-645
• /
• 2021

The deep learning based end-to-end TTS system consists of Text2Mel module that generates spectrogram from text, and vocoder module that synthesizes speech signals from spectrogram. Recently, by applying deep learning technology to the TTS system the intelligibility and naturalness of the synthesized speech is as improved as human vocalization. However, it has the disadvantage that the inference speed for synthesizing speech is very slow compared to the conventional method. The inference speed can be improved by applying the non-autoregressive method which can generate speech samples in parallel independent of previously generated samples. In this paper, we introduce FastSpeech, FastSpeech 2, and FastPitch as Text2Mel technology, and Parallel WaveGAN, Multi-band MelGAN, and WaveGlow as vocoder technology applying non-autoregressive method. And we implement them to verify whether it can be processed in real time. Experimental results show that by the obtained RTF all the presented methods are sufficiently capable of real-time processing. And it can be seen that the size of the learned model is about tens to hundreds of megabytes except WaveGlow, and it can be applied to the embedded environment where the memory is limited.

• 한국사회복지학회:학술대회논문집
• /
• 2003.05a
• /
• pp.359-376
• /
• 2003

본 연구의 목적은 저소득층 편부모의 자아통제가 심리적 안녕감에 미치는 요인들을 살펴봄으로써 이들 대상의 심리적 안녕감을 향상시킬 수 있는 정색적 시사점을 얻고자 하였다. 본 연구는 편의표본추출방법에 따라 저소득층이 밀집된 48개 지역에 거주하는 635명의 편부모를 대상으로 구조화된 자기 기입식 설문지를 이용만 우편설문조사를 실시하였다. 본 연구에서 종속 변수인 심리적 안녕감은 우울증, 자존감, 삶의 만족도 등이며, 독립변수로 인구사회적 배경 요인(모델 1), 사회적 지원 요인(모델 2), 자아통제 요인(모델 3) 등이 사용되었다. 본 연구의 주요 초점인 자아통제 요인은 내외통제성, 적극적 대처, 소극적 대처 등으로 구성되었다. 자료분석은 위계적 다중회귀분석(Hierarchical Multiple Regression Analysis)이 활용되었다. 본 연구에서 나타난 주요 결과는 다음과 같다. 1) 우울증 모델에서는 자아통제 요인이 모두 유의한 영향을 미치는 것으로 판명되었다. 내적 통제감이 높고 소극적 대처력이 낮고 적극적 대처력이 높을수록 우울증이 낮게 나타났다. 다른 유의한 변인은 건강상태로 건강할수록 우울증이 낮았다. 2) 자존감 모델에서는 자아통제 요인 중 내외통제청과 적극적 대처력이 유의한 변인으로 판명되어, 내적 통제감이 놓고 적극적 대처를 많이 하는 편부모일수록 자존감이 높았다. 또한 고연령이고, 건강상태가 좋고, 종교를 갖고 있으며 자녀와의 관계가 가까울수록 자존감이 높았다. 3) 삶의 만족도 모델에서는 자아통제, 소극적 대처, 적극적 대처가 유의한 변인으로 분석되어, 내적 통제감이 높고, 소극적 대처는 낮으며 적극적 대처력이 높은 편부모가 삶의 민족도가 높은 것으로 나타났다. 수입, 건강, 종교, 자녀와의 관계는 두 모델에서는 유의하였으나, 자아통제 요인을 첨가하자 의미성이 없어졌다. 예상외로 부모, 형제, 친구, 공공기관, 종교기관으로부터 받는 사회적 지지는 거의 모든 모델에서 유의한 영향을 주지 않았다. 이러한 결과는 저소득층에게 제공되어지는 사회적 지지가 미미한 편으로 변이가 없기 때문으로 해결될 수 있다. 따라서, 저소득층 편부모에게 사회적 지지는 물론 자아통제를 제고하는 것이 그들의 심리적 안녕감을 높게 하는 주요 방안이라 팔 수 있다. 앞으로 저소득층 편부모의 자아통제를 제고할 수 있는 다양한 프로그램의 개발 및 이의 적극적 실행방안이 모색되어야 할 것이다.

• Phonetics and Speech Sciences
• /
• v.13 no.4
• /
• pp.47-53
• /
• 2021

Autoregressive Text-to-Speech (TTS) models suffer from inference instability and slow inference speed. Inference instability occurs when a poorly predicted sample at time step t affects all the subsequent predictions. Slow inference speed arises from a model structure that forces the predicted samples from time steps 1 to t-1 to predict the sample at time step t. In this study, an end-to-end non-autoregressive fast text-to-speech model is suggested as a solution to these problems. The results of this study show that this model's Mean Opinion Score (MOS) is close to that of Tacotron 2 - WaveNet, while this model's inference speed and stability are higher than those of Tacotron 2 - WaveNet. Further, this study aims to offer insight into the improvement of non-autoregressive models.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.8
• /
• pp.1551-1559
• /
• 2015

Recently, the performance of smart devices is almost similar to that of the existing PCs, thus the users of smart devices can perform similar works such as messengers, SNSs(Social Network Services), smart banking, etc. originally performed in PC environment using smart devices. Although the development of smart devices has led to positive impacts, it has caused negative changes such as an increase in security threat aimed at mobile environment. Specifically, the threats of mobile devices, such as leaking private information, generating unfair billing and performing DDoS(Distributed Denial of Service) attacks has continuously increased. Over 80% of the mobile devices use android platform, thus, the number of damage caused by mobile malware in android platform is also increasing. In this paper, we propose android based malware detection mechanism using time-series analysis, which is one of statistical-based detection methods.We use auto-regressive moving-average model which is extracting accurate predictive values based on existing data among time-series model. We also use fast and exact malware detection method by extracting possible malware data through Z-Score. We validate the proposed methods through the experiment results.

• Water for future
• /
• v.9 no.2
• /
• pp.87-100
• /
• 1976

These studies are aimed at the analysis of systematic variation pattern of water resources in Korean river catchments and the development of their simulation models from the stochastic analysis of monthly and annual hydrologic data as main elements of water resources, i.e. rainfall and streamflow. In the analysis, monthly & annual rainfall records in Soul, Taegu, Pusan and Kwangju and streamflow records at the main gauging stations in Han, Nakdong and Geum river were used. Firstly, the systematic variation pattern of annual streamflow was found by the exponential function relationship between their standard deviations and mean values of log-annual runoff. Secondly, stochastic characteristics of annual rainfall & streamflow series were studied by the correlogram Monte Carlo method and a single season model of 1st-order Markov type were applied and compared in the simulation of annual hydrologic series. In the simulation, single season model of Markov type showed better results than LN-model and the simulated data were fit well with historical data. But it was noticed that LN-model gave quite better results in the simulation of annual rainfall. Thirdly, stochastic characteristics of monthly rainfall & streamflow series were also studied by the correlogram and spectrum analysis, and then the Model-C, which was developed and applied for the synthesis of monthly perennial streamflow by lst author and is a Markov type model with transformed skewed random number, was used in the simulation of monthly hydrologic series. In the simulation, it was proved that Model-C was fit well for extended area in Korea and also applicable for menthly rainfall as well as monthly streamflow.

• Journal of Biomedical Engineering Research
• /
• v.9 no.1
• /
• pp.93-100
• /
• 1988

The Auto Regressive Parameter Estimation and Pattern Classification of EKG Signal for Automatic Diagnosis. This paper presents the results from pattern discriminant analysis of an AR (auto regressive) model parameter group, which represents the HRV (heart rate variability) that is being considered as time series data. HRV data was extracted using the correct R-point of the EKG wave that was A/D converted from the I/O port both by hardware and software functions. Data number (N) and optimal (P), which were used for analysis, were determined by using Burg's maximum entropy method and Akaike's Information Criteria test. The representative values were extracted from the distribution of the results. In turn, these values were used as the index for determining the range o( pattern discriminant analysis. By carrying out pattern discriminant analysis, the performance of clustering was checked, creating the text pattern, where the clustering was optimum. The analysis results showed first that the HRV data were considered sufficient to ensure the stationarity of the data; next, that the patern discrimimant analysis was able to discriminate even though the optimal order of each syndrome was dissimilar.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2016.05a
• /
• pp.280-280
• /
• 2016

특정 자료의 시간의 흐름에 따른 예측치를 추정하는 방법으로 AR Model 즉, 자기회귀모형이 많이 사용되고 있다. AR Model은 변수의 현재 값을 과거 값의 함수로 나타내게 되는데, 이런 시계열 분석 모델을 사용할 때 매개변수의 추정 과정이 필수적으로 요구된다. 일반적으로 매개변수를 추정하는 방법에는 확률적근사법(stochastic approximation), 최소제곱법(method of least square), 자기상관법(method of autocorrelation method), 최우도법(method of maximum likelihood) 등이 있다. AR Model에서 가장 많이 사용되는 최우도법은 표본크기가 충분히 클 때 가장 효율적인 방법으로 평가되지만 수치적으로 해를 구하는 과정이 복잡한 경우가 많으며, 해를 구하지 못하는 어려움이 따르기도 한다. 또한 표본 크기가 작을 때 일반적으로 잘 일치하지 않은 결과를 얻게 된다. 우리나라의 강우, 유량 등의 자료는 자료의 수가 적은 경우가 많기 때문에 최우도법을 통한 매개변수 추정 시 불확실성이 내재되어있지만 그것을 정량적으로 제시하는데 한계가 있다. 본 연구에서는 AR Model의 매개변수 추정 시 Bayesian 기법으로 매개변수의 사후분포(posterior distribution)를 제공하여 매개변수의 불확실성 구간을 정량적으로 표현하게 됨으로써, 시계열 분석을 통해 보다 신뢰성 있는 예측치를 얻을 수 있으리라 판단된다.

• The Korean Journal of Applied Statistics
• /
• v.21 no.4
• /
• pp.705-716
• /
• 2008

The objectives of this study are the detection and forecast of climate change signal in the annual mean of surface temperature data, which are generated by MRI/JMA CGCM over the Korean Peninsula. MRI/JMA CGCM outputs consist of control run data(experiment with no change of $CO_2$ concentration) and scenario run data($CO_2$ 1%/year increase experiment to quadrupling) during 142 years for surface temperature and precipitation. And ECMWF reanalysis data during 43 years are used as observations. All data have the same spatial structure which consists of 42 grid points. Two statistical models, the Bayesian fingerprint method and the regression model with autoregressive error(AUTOREG model), are separately applied to detect the climate change signal. The forecasts up to 2100 are generated by the estimated AUTOREG model only for detected grid points.

• The Journal of Society for e-Business Studies
• /
• v.22 no.4
• /
• pp.1-19
• /
• 2017

Wikipedia is one of the world's most visited sites for content collaboration. Its success is due to thousands of volunteers' motivation and commitment to contribute their knowledge to Wikipedia. In this paper, we use the Cox regression model to assess the effect of self-loop editing on the promotion of Wikipedia featured articles. We collected 2978 Wikipedia featured article editing history from start of Wikipedia until 2011. We use self-loops as a proxy measure for Wikipedia editors' commitment, and find that self-loop editing has a positive effect on the promotion of featured articles. We further distinguish the self-loop into a short-term self-loop and a long-term self-loop. We find that long-term self-loop editing is more helpful than short-term self-loop editing. This research has been conducted with both theoretical and practical application methods.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.4
• /
• pp.76-84
• /
• 1993

In this thesis, general method using autoregressive model in the functional separation of the myoelectric signal of human arm movements are suggested. Covariance method and sequential least squares algorithm were used to determine the model parameters and the order of signal model to describe six arm movement patterns` the forearm flexion and extension, the wrist pronation and supination, rotation-in and rotation out. The confidence interval to classify the functions of arm movement was defined by the mean and standard deviation of total squares error. With the error signals of autoregressive(AR) model, the result showed that the highest success, rate was abtained in the case of 4th order, and success rate was decreased with increase of order. This technique might be applied to biomedical-and rehabilitation-engi-neering.