• 말소리와 음성과학
• /
• 제2권3호
• /
• pp.115-124
• /
• 2010

The current study investigated the role of acoustic correlates of domain-initial strengthening in lexical segmentation of a non-native language. In a series of cross-modal identity-priming experiments, native Korean listeners heard English auditory stimuli and made lexical decision to visual targets (i.e., written words). The auditory stimuli contained critical two word sequences which created temporal lexical ambiguity (e.g., 'mill#company', with the competitor 'milk'). There was either an IP boundary or a word boundary between the two words in the critical sequences. The initial CV of the second word (e.g., [$k_{\Lambda}$] in 'company') was spliced from another token of the sequence in IP- or Wd-initial positions. The prime words were postboundary words (e.g., company) in Experiment 1, and preboundary words (e.g., mill) in Experiment 2. In both experiments, Korean listeners showed priming effects only in IP contexts, indicating that they can make use of IP boundary cues of English in lexical segmentation of English. The acoustic correlates of domain-initial strengthening were also exploited by Korean listeners, but significant effects were found only for the segmentation of postboundary words. The results therefore indicate that L2 listeners can make use of prosodically driven phonetic detail in lexical segmentation of L2, as long as the direction of those cues are similar in their L1 and L2. The exact use of the cues by Korean listeners was, however, different from that found with native English listeners in Cho, McQueen, and Cox (2007). The differential use of the prosodically driven phonetic cues by the native and non-native listeners are thus discussed.

• 음성과학
• /
• 제11권3호
• /
• pp.111-128
• /
• 2004

In English, a focused word with new information receives a pitch accent. This paper examines how English native speakers and Korean speakers produce and perceive focus in English yes-no questions. The production experiments show that native speakers realize an appropriate intonation of yes-no questions, in which a focused word has a low pitch accent followed by a high phrasal accent and a high boundary tone. However, Korean speakers usually give a high tone to a focused word. In a like manner, the perception experiments show that English native speakers judge a word with a low tone to be focused, while Korean speakers have difficulty in comprehending a focused word realized as a low tone. And it is found that Korean speakers tend to perceive low tones on sentence initial and final focused words better than those on sentence medial focused words, and they often perceive a word with a relatively high fundamental frequency or a sharp rise of fundamental frequency as a focused word. This paper shows that Korean speakers have trouble to produce and perceive an appropriate tonal pattern of a focused yes-no question, and that can cause confusion in a conversation with native speakers.

• 음성과학
• /
• 제13권1호
• /
• pp.29-41
• /
• 2006

This study investigates the degree to which various prosodic cues at the boundaries of prosodic phrases in Korean contribute to word segmentation. Since most phonological words in Korean are produced as one Accentual Phrase (AP), it was hypothesized that the detection of acoustic cues at AP boundaries would facilitate word segmentation. The prosodic characteristics of Korean APs include initial strengthening at the beginning of the phrase and pitch rise and final lengthening at the end. A perception experiment utilizing an artificial language learning paradigm revealed that cues conforming to the aforementioned prosodic characteristics of Korean facilitated listeners' word segmentation. Results also indicated that duration and amplitude cues were more helpful in segmentation than pitch. Nevertheless, results did show that a pitch cue that did not conform to the Korean AP interfered with segmentation.

• 말소리와 음성과학
• /
• 제3권3호
• /
• pp.63-68
• /
• 2011

This study investigates how Korean students employ acoustic-phonetic cues in perceiving word boundaries of near-homophonous English phrases. For this study, 60 Korean college students participated in the experiment of discriminating word boundaries for 42 pairs of stimuli comprising the allophonic cues of aspiration and glottal stop. Results were analysed in terms of the correctness of responses and the correlation between correctness and confidence. Results showed that stimuli pairs of the glottal stop cue give a higher correctness but those of aspiration a relatively lower correctness. Comparison of the results of this study with those of the previous studies of English and Japanese speakers showed that Korean and Japanese speakers of English give a substantially lower correctness than native speakers of English, while Korean learners of English as a foreign language provide a lower correctness than Japanese speakers of English as a second language.

• 정보관리학회지
• /
• 제13권1호
• /
• pp.47-63
• /
• 1996

기존의 한글 자동 색인 방법들은 어절 단위 색인법과 형태소 단위 색인법으로 분류될 수 있다. 전자는 문서내의 어절에서 비색인 분절을 절단함으로써 색인어를 추출하는 방법으로, 문서들이 많은 복합 명사들을 포함할 경우 검색 효과가 저하된다. 후자는 형태소 해석이나 구문 해석을 이용하여 중요한 의미를 갖는 명사나 명사구를 추출하는 방법으로 단일 명사를 추출함으로써 복합 명사의 띄어쓰기 문제를 극복할 수 있다. 그러나 색인 과정에서 요구되는 많은 언어 정보를 개발하고 유지 보수해야 하는 부담을 지니고 있다. 본 논문에서는 기존의 색인 방법들의 문제점들을 완화할 수 있는 새로운 색인 방법을 제안한다. 그리고 실험을 통하여 제안하는 방법의 성능을 평가한다.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1998년도 제15회 음성통신 및 신호처리 워크샵(KSCSP 98 15권1호)
• /
• pp.87-92
• /
• 1998

In general, the synthesis unit database was constructed by recording isolated word. In that case, each boundary of word has typical prosodic pattern like a falling intonation or preboundary lengthening. To get natural synthetic speech using these kinds of database, we must artificially distort original speech. However, that artificial process rather resulted in unnatural, unintelligible synthetic speech due to the excessive prosodic modification on speech signal. To overcome these problems, we gathered thousands of sentences for synthesis database. To make a phone level synthesis unit, we trained speech recognizer with the recorded speech, and then segmented phone boundaries automatically. In addition, we used laryngo graph for the epoch detection. From the automatically generated synthesis database, we chose the best phone and directly concatenated it without any prosody processing. To select the best phone among multiple phone candidates, we used prosodic information such as break strength of word boundaries, phonetic contexts, cepstrum, pitch, energy, and phone duration. From the pilot test, we obtained some positive results.

• 말소리와 음성과학
• /
• 제6권4호
• /
• pp.3-15
• /
• 2014

An articulatory study (using an Electromagnetic Articulography, EMA) was conducted to explore effects of prosodic boundary strength (Intonational Phrase/IP versus Word/Wd), and focus (Focused/accented, Neutral, Unfocused/unaccented) on the kinematic realization of /m/ in the coda (${\ldots}$am#i${\ldots}$) and the onset (${\ldots}$a#mi${\ldots}$) conditions in Korean. (Here # refers to a prosodic boundary such as an IP or a Wd boundary). Several important points have emerged. First, the boundary effect on /m/s was most robustly observed in the temporal dimension in both the coda (IP-final) and the onset (IP-initial) conditions, generally in line with cross-linguistically observable boundary-related lengthening patterns. Crucially, however, in contrast with boundary-related slowing-down effects that have been observed in English, both the IP-final and IP-initial temporal expansions of Korean /m/s were not accompanied by an articulatory slowing down. They were, if anything, associated with a faster movement in the lip opening (release) phase (into the vowel). This suggests that the mechanisms underlying boundary-related temporal expansions may differ between languages. Second, observed boundary-induced strengthening effects (both spatial and temporal expansions, especially on the IP-initial /m/s) were remarkably similar to prominence (focus)-induced strengthening effects, which is again counter to phrase-initial strengthening patterns observed in English in which boundary effects are dissociated from prominent effects. This suggests that initial syllables in Korean may be a common focus for both boundary and prominence marking. These results, taken together, imply that the boundary-induced strengthening in Korean is different in nature from that in English, each being modulated by the individual language's prosodic system. Third, the coda and the onset /m/s were found to be produced in a subtly but significantly different way even in a Wd boundary condition, a potentially neutralizing (resyllabification) context. This suggests that although the coda may be phonologically 'resyllabified' into the following syllable in a phrase-medial position, its underlying syllable affiliation is kinematically distinguished from the onset.

• Journal of Information Processing Systems
• /
• 제13권5호
• /
• pp.1103-1125
• /
• 2017

This paper aims to extract an ObjectProperty-UsageMethod relation, in particular the HerbalMedicinalProperty-UsageMethod relation of the herb-plant object, as a semantic relation between two related sets, a herbal-medicinal-property concept set and a usage-method concept set from several web documents. This HerbalMedicinalProperty-UsageMethod relation benefits people by providing an alternative treatment/solution knowledge to health problems. The research includes three main problems: how to determine EDU (where EDU is an elementary discourse unit or a simple sentence/clause) with a medicinal-property/usage-method concept; how to determine the usage-method boundary; and how to determine the HerbalMedicinalProperty-UsageMethod relation between the two related sets. We propose using N-Word-Co on the verb phrase with the medicinal-property/usage-method concept to solve the first and second problems where the N-Word-Co size is determined by the learning of maximum entropy, support vector machine, and naïve Bayes. We also apply naïve Bayes to solve the third problem of determining the HerbalMedicinalProperty-UsageMethod relation with N-Word-Co elements as features. The research results can provide high precision in the HerbalMedicinalProperty-UsageMethod relation extraction.

• 말소리와 음성과학
• /
• 제8권1호
• /
• pp.53-62
• /
• 2016

In this study, we examined whether Chinese learners of Korean can correctly produce the phonological process on the lexical and post-lexical level. For this purpose 4 Korean native speakers and 10 advanced and 10 intermediate Chinese learners of Korean participated in the production test. The materials analyzed constituted 10 Korean sentences in which nasalization can be applied on the syllable boundary, word boundary(w-boundary) as well as accentual phrase boundary(AP-boundary). The results show that for Korean speakers, nasalization was applied 100% at all level whereas for Chinese speakers, the rate of application of nasalization is different according to prosodic constituents and Korean proficiency. Nasalization was more frequently applied at the lexical level than the post-lexical level, and it is more frequent in the w-boundary conditions than in the AP-boundary conditions. However, the rate of nasalization in the w-boundary is close to the lexical level. The pronunciation errors were committed either as non application of nasalization or coda obstruent ommission. In the case of non application of nasalization, Chinese learners of Korean produced the target syllables as underling forms, which were not transformed as surface forms. In addition, we can observe the ommission of coda obstruents in 'lenis obstruents+nasal sound' sequences. As a result, nasalization is blocked by this omission.

• 한국산학기술학회논문지
• /
• 제7권2호
• /
• pp.182-187
• /
• 2006

본 논문은 통계 기반 접근 방식인 HMM(Hidden Markov model)과 생물학의 개체명에 관한 온톨로지 정보를 이용한 생물학 문서에서의 개체명(named entity) 경계 인식 방법을 제안한다. 제안하는 방법은 31개의 자질 정보를 이용한 평탄화 기법을 사용하며 생물학 개체명의 계층 정보를 이용하여 HMM의 자료 부족 문제를 완화시킬 수 있도록 하였다. 개체명 경계 인식의 학습과 실험을 위하여 GENIA 코퍼스 ver 2.1을 사용하였으며 개체명 경계 인식 실험을 수행한 결과 모든 부류를 사용한 경우보다 정확도 및 실행 속도가 개선됨을 확인하였다.