• MALSORI
• /
• no.35_36
• /
• pp.13-24
• /
• 1998

This study aims at examining the relationship between tonal events and their related vowel and syllable duration in Korean. Two things were investigated: one is to see if there is a hierarchical relationship in prosodic unit-final-lengthening and the other is to see if accentual phrase initial high tone syllable gets lengthened. Generally, higher prosodic units show larger degree of lengthening of the final vowel and also final syllable duration than the lower ones except for accentual phrase: Mean duration of utterance-final or intonational-phrase-final syllable(and its vowels) was longer than that of accentual-phrase-final or word-final syllable(and its vowels). However, mean duration of accentual phrase final syllable was shorter than that of word final syllable. Mean vowel duration of accentual phrase initial high tone syllable was shorter than that of any other prosodic unit. Its mean syllable duration, however, was longer than that of accentual-phrase-final or word-final syllable, indicating that strong consonants(fortis and aspirated) frequently appear in the accentual phrase initial position and this position is a prosodically strong position showing longer duration as well as high tone.

• 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.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.3
• /
• pp.62-67
• /
• 1999

In this paper, we investigates a reliable model of the Predictive Recurrent Neural Network for the speech recognition. Predictive Neural Networks are modeled by syllable units. For the given input syllable, then a model which gives the minimum prediction error is taken as the recognition result. The Predictive Neural Network which has the structure of recurrent network was composed to give the dynamic feature of the speech pattern into the network. We have compared with the recognition ability of the Recurrent Network proposed by Elman and Jordan. ETRI's SAMDORI has been used for the speech DB. In order to find a reliable model of neural networks, the changes of two recognition rates were compared one another in conditions of: (1) changing prediction order and the number of hidden units: and (2) accumulating previous values with self-loop coefficient in its context. The result shows that the optimum prediction order, the number of hidden units, and self-loop coefficient have differently responded according to the structure of neural network used. However, in general, the Jordan's recurrent network shows relatively higher recognition rate than Elman's. The effects of recognition rate on the self-loop coefficient were variable according to the structures of neural network and their values.

• 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.

• Phonetics and Speech Sciences
• /
• v.1 no.2
• /
• pp.129-140
• /
• 2009

The purpose of the present study is to investigate frequency related information and syllable structure constraints on Sino-Korean. Previous studies on Sino-Korean have mostly investigated the historical change of sounds and reviewed archaic features of Chinese language in Sino-Korean. Unfortunately, there is little study on the sounds of contemporary Sino-Korean in terms of syllable structure constraints. For the purpose of the present study, sounds of 7,742 Chinese characters used in Sino-Korean (7,795 syllables) were investigated and syllable matrices made based on the results of frequency related information. As a result, 483 syllable types were observed and the most frequently observed syllables were as follows: /ku/ (103) > /ki/ (100) > /ju/ (87) > /pi/ (86). Only 16 out of 19 consonants are used for Sino-Korean. /$t^{\ast}$/ and /$p^{\ast}$/ are never used in Sino-Korean and /kh, $s^{\ast}$, $k^{\ast}$/ occur only a few times (3, 2, 1 respectively). /k/ (17.5%) shows the highest frequency and /n, ${\eta}$, 1, tc, m/ occupied the next rankings. Among 20 vowel types, /a/ showed the highest frequency and /o, u, i, $j{\Lambda}$, ${\Lambda}$/ occupied the next rankings. Based on the syllable matrices, gaps were observed and classified into accidental or systematic ones. Onset and nucleus, nucleus and coda, onset and coda, and other syllable structure constraints of Sino-Korean were listed.

• Speech Sciences
• /
• v.12 no.3
• /
• pp.175-183
• /
• 2005

In this study, realization of strong and weak syllables in English by Korean leaners with low English proficiency was examined through experiment. The aspects of three acoustic characteristics-duration, pitch, amplitude-were measured and compared with native speakers of English. It was assumed that production of duration, pitch and amplitude of strong and weak syllable by Korean learners would be different from that of English native speakers. According to the production experiments, English native speakers produced strong syllable longer, higher and louder than weak syllable. However, Korean leaners produced strong syllable higher and louder than weak syllable, but not longer enough. Specifically, weak syllable by Korean leaners was longer and strong syllable shorter than native speakers. Furthermore, the difference in duration of syllables between Korean leaners and English native speakers is more significant than pitch and amplitude. As a result, the duration was more important cue for the realization of stress than pitch and amplitude. However, Korean leaners did not produce duration of stressed syllables as English native speakers did, even though they produce the pitch and amplitude of stressed syllable in a similar way to native speakers. The reasons for those were considered, too.

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.268-272
• /
• 2016

영어권 언어가 어절 단위로 품사를 부여하는 반면, 한국어는 굴절이 많이 일어나는 교착어로서 데이터부족 문제를 피하기 위해 형태소 단위로 품사를 부여한다. 이러한 구조적 차이 안에서 한국어에 적합한 품사 태깅 단위는 지속적으로 논의되어 왔으며 지금까지 음절, 형태소, 어절, 구가 제안되었다. 본 연구는 어절 단위로 태깅함으로써 야기되는 복잡한 품사 태그와 데이터부족 문제를 해소하기 위해 어절에서 주요 실질 형태소와 주요 형식 형태소만을 뽑아 새로운 어절을 생성하고, 생성된 단순한 어절에 대해 CRF 태깅을 수행하였다. 실험결과 평가 말뭉치에서 미등록 어절 등장 비율은 9.22%에서 5.63%로 38.95% 감소시키고, 어절단위 정확도를 85.04%에서 90.81%로 6.79% 향상시켰다.

• Phonetics and Speech Sciences
• /
• v.6 no.3
• /
• pp.121-130
• /
• 2014

This study investigates the acoustic property of aspiration realized in English voiceless stops of /Cr/ and /sCr/ clusters. VOT is measured from stops in these clusters produced by two groups; one from native speakers of English and the other from Korean native speakers. Aspiration of stops in different types of clusters is compared to various phonological factors such as location of stress, syllable type, and position in word. Pursuing the idea that phonetic realization is correlated with phonological representation, attempts are made to account for the gradient nature of aspiration of stops on the basis of syllable structure at the phonetic level, which may vary in the wake of resyllabification. Voiceless stops in /Cr/ and /sCr/ clusters are further compared to results obtained in the previous study on /sC/ cluster. Variations in aspiration are also characterized in terms of segmental precedence relation of stops in the clusters, namely, post-[s], pre-[r], or both.

• Phonetics and Speech Sciences
• /
• v.8 no.3
• /
• pp.1-9
• /
• 2016

This paper investigated the phoneme distribution and phonological processes of orthographic and pronounced phrasal words in light of syllable structure in the Seoul Corpus in order to provide linguists and phoneticians with a clearer understanding of the Korean language system. To achieve the goal, the phrasal words were extracted from the transcribed label scripts of the Seoul Corpus using Praat. Following this, the onsets, peaks, codas and syllable types of the phrasal words were analyzed using an R script. Results revealed that k0 was most frequently used as an onset in both orthographic and pronounced phrasal words. Also, aa was the most favored vowel in the Korean syllable peak with fewer phonological processes in its pronounced form. The total proportion of all diphthongs according to the frequency of the peaks in the orthographic phrasal words was 8.8%, which was almost double those found in the pronounced phrasal words. For the codas, nn accounted for 34.4% of the total pronounced phrasal words and was the varied form. From syllable type classification of the Corpus, CV appeared to be the most frequent type followed by CVC, V, and VC from the orthographic forms. Overall, the onsets were more prevalent in the pronunciation more than the codas. From the results, this paper concluded that an analysis of phoneme distribution and phonological processes in light of syllable structure can contribute greatly to the understanding of the phonology of spoken Korean.

• Koreanishche Zeitschrift fur Deutsche Sprachwissenschaft
• /
• v.8
• /
• pp.1-23
• /
• 2003

Im Deutschen wird generell einem Suffix zwei verschiedene prosodische Strukturen zugewiesen: ein konsonanteninitiales Suffix bildet sein eigenes phonologisches Wort, ein vokalinitiales Suffix dagegen nicht. $F\"{u}r$ diese Unterscheidung zwischen zwei Suffixklassen nennt man innerhalb einer regelbasierten Theorie zwei Kriterien: (i) Phonologisches Wort fungiert als $Dom\"{a}ne{\;}f\"{u}r$ die Silbifizierung und die Koordinationsreduktion, (ii) Konsonanteninitiale Suffixe $tr\"{a}gen$ einen Nebenakzent, $w\"{a}hrend$ einem vokalinitialen Suffix kein Akzent zugewiesen wird. Diese Arbeit zeigt, dass das erste Kriterium kein Argument $f\"{u}r$ die Annahme eines phonologischen Wortes ist. Die Existenz eines phonologischen Wortes und die Unterscheidung zwischen zwei Suffixklassen hinsichtlich des prosodischen Status werden durch einen Zirkelschluss $begr\"{u}ndet$, indem die Annahme eines phonologischen Wortes $Ph\"{a}nomene$ wie Silbifizierung und Koordinationsreduktion rechtfertigt, diese $Ph\'{a}nomene$ wiederum ihrerseits die Annahme eines phonologischen Wortes. Die hier vorgeschlagene Constraintsinteraktion erfasst dagegen den wesentlichen Charakter der beiden $Ph\"{a}nomene$, ohne dabei einen direkten Bezug auf das phonologische Wort zu nehrnen. Auch im Zusammenhang mit dem zweiten Kriterium stellt sich heraus, dass es keinen Grund gibt, Akzentunterschiede bei Suffixen anzunehmen. Die Allomorphie der Superlativ-Suffixe +est/+st ergibt sich aus dem Constraintranking OCP$\gg$ $fu{\ss}bezogene$ Constraints, nicht aus dem Akzentunterschied zwischen Suffixen.