• Journal of KIISE
• /
• v.44 no.5
• /
• pp.503-509
• /
• 2017

In this paper, the copy mechanism and input feeding are applied to recurrent neural network(RNN)-search model in a Korean-document summarization in an end-to-end manner. In addition, the performances of the document summarizations are compared according to the model and the tokenization format; accordingly, the syllable-unit, morpheme-unit, and hybrid-unit tokenization formats are compared. For the experiments, Internet newspaper articles were collected to construct a Korean-document summary data set (train set: 30291 documents; development set: 3786 documents; test set: 3705 documents). When the format was tokenized as the morpheme-unit, the models with the input feeding and the copy mechanism showed the highest performances of ROUGE-1 35.92, ROUGE-2 15.37, and ROUGE-L 29.45.

• Spatial Information Research
• /
• v.23 no.3
• /
• pp.1-10
• /
• 2015

Due to recent improvement of various sensor technologies, indoor positioning becomes available. However, Indoor positioning technologies by Wi-Fi radio map and acceleration sensor and digital campus still have a certain level of errors and a number of researches have been done to increase the positioning accuracy of the indoor positioning. If we could provide a room level accuracy, indoor location based services with current indoor positioning methods such as Wi-Fi radio map and acceleration sensors would be possible. In this paper, we propose an indoor map matching method to provide a room level accuracy based on hidden markov model.

• Journal of the Korea Computer Industry Society
• /
• v.2 no.6
• /
• pp.867-874
• /
• 2001

In this paper, considering the pronunciation characteristic of Korean, recognition units which can shorten the recognition time and reflect the coarticulation effect simultaneously are suggested. These units are composed of monophone and hipbone ones. Monophone units are applied to the vowels which represent stable characteristic. Biphones are used to the consonant which vary according to adjacent vowel. In the experiment of word recognition of PBW445 database, the compound units result in comparable recognition accuracy with 57％ speed up compared with triphone units and better recognition accuracy with similar speed. In addition, we can reduce the memory size because of fewer units.

• Proceedings of the Korean Institute Of Construction Engineering and Management
• /
• 2004.11a
• /
• pp.565-570
• /
• 2004

The construction Progress is widely used 3s a critical index for effective project management. However, the methods, structure, data, and accuracy of progress measurement may vary depending on specific characteristics of the project, organization, or location. Even in an organization, different projects may utilize different measurement methods to effectively achieve their own management purpose. The excessive effort required to manipulate very detailed progress data is also an issue Therefore the purpose of this study is to dove]op an automated progress measurement model utilizing standard progress measurement package (SPMP).

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.5
• /
• pp.24-31
• /
• 1994

In this paper, we study continuous speech recognition of Korean language using acoustic models of word juncture coarticulation. To alleviate the performance degradation due to coarticulation problems, we use context-dependent units that model inter-word transitions in addition to intra-word transitions. In all cases the initial phone of each word has to be specified for each possible final phone of the previous word similarly for the final phone of each word. To improve the robustness of the HMM parameters, the covariance matrix is smoothed. We also use position-dependent units to improve the discriminative power between units. Simulation results show that when the improved models of word juncture coarticulation are used. the recognition performance is considerably improved compared to the baseline system using only intra-word units.

• The Korean Journal of Applied Statistics
• /
• v.24 no.2
• /
• pp.217-225
• /
• 2011

In this paper, a seasonal variable selection method using the spectral analysis accompanied with seasonal linear model is suggested. The suggested method is applied to the prediction of intra-day call arrivals at a large North American commercial bank call center and a signi cant intra-month seasonal variable I detected. This newly detected seasonal factor is included in the seasonal linear model and is compared with the seasonal linear models without this variable to see whether the new variable helps to improve the forecasting performance. The seasonal linear model with the new variable outperformed the models without it in one-day-ahead forecasting.

• Journal of KIISE
• /
• v.41 no.9
• /
• pp.642-651
• /
• 2014

This paper proposes three probabilistic models for syllable-based Korean morphological analysis, and presents the performance of proposed probabilistic models. Probabilities for the models are acquired from POS-tagged corpus. The result of 10-fold cross-validation experiments shows that 98.3% answer inclusion rate is achieved when trained with Sejong POS-tagged corpus of 10 million eojeols. In our models, POS tags are assigned to each syllable before spelling recovery and morpheme generation, which enables more efficient morphological analysis than the previous probabilistic models where spelling recovery is performed at the first stage. This efficiency gains the speed-up of morphological analysis. Experiments show that morphological analysis is performed at the rate of 147K eojeols per second, which is almost 174 times faster than the previous probabilistic models for Korean morphology.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.453-458
• /
• 2018

본 논문은 한국어 자소를 인식 단위로 사용한 hybrid CTC-Attention 모델 기반 end-to-end speech recognition을 제안한다. End-to-end speech recognition은 기존에 사용된 DNN-HMM 기반 음향 모델과 N-gram 기반 언어 모델, WFST를 이용한 decoding network라는 여러 개의 모듈로 이루어진 과정을 하나의 DNN network를 통해 처리하는 방법을 말한다. 본 논문에서는 end-to-end 모델의 출력을 추정하기 위해 자소 단위의 출력구조를 사용한다. 자소 기반으로 네트워크를 구성하는 경우, 추정해야 하는 출력 파라미터의 개수가 11,172개에서 49개로 줄어들어 보다 효율적인 학습이 가능하다. 이를 구현하기 위해, end-to-end 학습에 주로 사용되는 DNN 네트워크 구조인 CTC와 Attention network 모델을 조합하여 end-to-end 모델을 구성하였다. 실험 결과, 음절 오류율 기준 10.05%의 성능을 보였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2012.07a
• /
• pp.358-360
• /
• 2012

본 논문에서는 새로운 율모델을 기반으로한 프레임 단위 HEVC 율제어 기법을 제안한다. 기존의 비디오 압축표준과는 달리 HEVC 는 계층 구조를 지닌 쿼드트리 기반 움직임 예측 및 변환 부호화를 수행한다. 본 논문에서는 쿼드트리 계층의 깊이에 따라 신호의 통계적 특성이 매우 달라지는 것은 이용하여 라플라시안 확률 모델을 각 쿼드트리 계층에 독립적으로 이용한 새로운 율모델을 이용한 율제어 기법을 제안한다. 제안방법에서는 계층적 부호화 단위인 CU 를 계층 깊이에 따라 세 가지 카테고리로 분류하고 각 카테고리에 따라 변환 계수에 대한 라플라시안 확률 분포 함수를 율-양자화 모델을 만든다. 제안된 율모델은 특성이 매우 다른 각 CU 깊이에 따라 독립적인 라플라이안 확률 분포 함수를 이용하기 때문에 매우 정확하고 적응적인 비트율 예측이 가능하므로 보다 안정적이고 정확한 율제어가 가능하다. 실험결과는 제안된 율제어 기법이 단일 확률 분포 함수를 사용했을 경우보다 평균 0.16dB 의 PSNR 향상이 있었음을 보여주었으며 제안된 방법은 각 프레임에 대한 목표 비트에 보다 안정적으로 부호화하는 것을 보여주었다.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.15 no.2
• /
• pp.38-48
• /
• 2001

Kalman Filter model of demand for residental water and consumption pattern wore tested for their ability to explain the hourly residental demand for water in metro-politan distribution system. The daily residental demand can be obtained from Kalman Filter model which is optimized by statistical analysis of input variables. The hourly residental demand for water is calculated from the daily residental demand and consumption pattern. The consumption pattern which has 24 time rates is characterized by data granulization in accordance with season kind, weather and holiday. The proposed approach is applied to water distribution system of metropolitan areas in Korea and its effectiveness is checked.