• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.184-187
• /
• 2020

다중 지문 기계독해는 질문과 여러 개의 지문을 입력받고 입력된 지문들에서 추출된 정답 중에 하나의 정답을 출력하는 문제이다. 다중 지문 기계독해에서는 정답이 있을 단락을 선택하는 순위화 방법에 따라서 성능이 크게 달라질 수 있다. 본 논문에서는 단락 안에 정답이 있을 확률을 예측하는 단락 재순위화 모델과 선택된 단락에서 서술형 정답을 위한 세부적인 정답의 경계를 예측하는 세부 단락 선별 기법을 제안한다. 단락 순위화 모델 학습의 경우 모델 학습을 위해 각 단락의 출력에 softmax와 cross-entroy를 이용한 손실 값과 sigmoid와 평균 제곱 오차의 손실 값을 함께 학습하고 키워드 매칭을 함께 적용했을 때 KorQuAD 2.0의 개발셋에서 상위 1개 단락, 3개 단락, 5개 단락에서 각각 82.3%, 94.5%, 97.0%의 재현율을 보였다. 세부 단락 선별 모델의 경우 입력된 두 단락을 비교하는 duoBERT를 이용했을 때 KorQuAD 2.0의 개발셋에서 F1 83.0%의 성능을 보였다.

• Phonetics and Speech Sciences
• /
• v.6 no.1
• /
• pp.39-45
• /
• 2014

This study investigates twenty four Korean females' production of English front vowels focusing on the distinction in /i/ vs /ɪ/ and /ɛ/ vs /${\ae}$/ and formant values of stressed and unstressed vowels compared with those of native English speakers. The Korean learners were asked to read a textbook passage which includes ten sentences including target vowels. The major results indicate that: (1) Korean learners have trouble producing a distinct version (tense and lax) of front vowels in the paragraph reading; (2) The vowel space of the stressed vowels in a paragraph is smaller than that of embedded sentences; and (3) The vowel quality of the unstressed vowels produced by the Korean learners is similar to that of the native English speakers. The findings from this study can be applied to the pronunciation teaching for the Korean learners of English vowels and realization of English stress.

• Phonetics and Speech Sciences
• /
• v.7 no.2
• /
• pp.111-117
• /
• 2015

The purpose of this study is to investigate how Chinese learners of Korean produce Korean plain plosives differently in a reading passage and isolated sentences. There are several studies on Korean plosives produced by Chinese speakers, but the study comparing the production of reading passage and isolated sentences are rare. For these purposes, ten Chinese speakers' VOT values of Korean plain plosives were measured using Speech Analyzer. The results show that there is no significant difference between the plain plosive production of a reading passage and that of isolated sentences. In the further studies, the measurement of pitch with VOT is needed.

• Journal of the Korean Academy of Child and Adolescent Psychiatry
• /
• v.27 no.2
• /
• pp.109-118
• /
• 2016

Objectives: The aim of this study was to develop the computerized Comprehensive Learning Test-Reading (CLT-R) to evaluate the cognitive processes and achievements related to their basic reading ability and identify dyslexia in children and adolescents in South Korea. We also obtained the normative data and evaluated the reliability and validity of the test. Methods: We developed the CLT-R, including the word attack/nonword decoding, paragraph reading, sound blending, nonword repetition, rapid automatized naming, letter-sound matching, visual attention, orthography awareness, and digit span tests, for the purpose of diagnosing dyslexia. We investigated the reliability and validity of the tests and gathered the normative data from 399 subjects (male 48.9%), aged 5-14 years, from the last grade in kindergarten to middle school, dwelling in Seoul and Gyeonggi Province, South Korea. Results: No statistical differences were observed between the means of the tests and retests of the CAT. The mean of the correlation coefficient of the test-retest scores was 0.85. According to the construct validity test calculated by principal constant analysis using the oblique rotation method, 4 factors explained 70.0% of the cumulative variances. In addition, the normative data were obtained for all of the CLT-R subtests. Conclusion: The computerized CLT-R can be used as a reliable and valid tool to evaluate the reading achievement and reading related cognitive process in Korean children and adolescents in schools, clinics, and research institutes.

• Journal of the Korean Society of Laryngology, Phoniatrics and Logopedics
• /
• v.28 no.2
• /
• pp.84-88
• /
• 2017

Background and Objectives : The purpose of this study was to compare the usefulness of Cepstral peak prominence (CPP) with parameter of Multiple Dimensional Voice Program (MDVP) in evaluating unilateral vocal fold paraylsis patients with subjective voice impairment. Materials and Methods : From July 2014 to August 2016, 37 patients with unilateral vocal fold paralysis who had been diagnosed with unilateral vocal fold paralysis and had received two or more voice tests before and after the diagnosis were evaluated for maximum phonation time (MPT), MDVP and CPP. Respectively. Voice tests were performed with short vowel /a/ and paragraph reading. Results : The CPP-a (CPP with vowel /a/) and CPP-s (CPP with paragraph reading) of the Cepstrum were statistically negatively correlated with G, R, B, and A before the voice therapy. Jitter, Shimmer, and NHR of MDVP were positively correlated with G, R, B. Jitter, Shimmer, and NHR of the MDVP were significantly correlated with the Cepstrum index. G, B, A and CPP-a and CPP-s showed a statistically significant negative correlation and a somewhat higher correlation coefficient between 0.5 and 0.78. On the other hand, in MDVP index, there was a positive correlation with G and B only with Jitter of 0.4. Conclusion : CPP can be an important evaluation tool in the evaluation of speech in the unilateral vocal cord paralysis when speech energy changes or the cycle is not constant during speech.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.11
• /
• pp.2133-2138
• /
• 2017

The analysis of existing XML documents and other documents was centered on words. It can be implemented using a morpheme analyzer, but it can classify many words in the document and cannot grasp the core contents of the document. In order for a user to efficiently understand a document, a paragraph containing a main word must be extracted and presented to the user. The proposed system retrieves keyword in the normalized XML document. Then, the user extracts the paragraphs containing the keyword inputted for searching and displays them to the user. In addition, the frequency and weight of the keyword used in the search are informed to the user, and the order of the extracted paragraphs and the redundancy elimination function are minimized so that the user can understand the document. The proposed system can minimize the time and effort required to understand the document by allowing the user to understand the document without reading the whole document.

• Phonetics and Speech Sciences
• /
• v.6 no.4
• /
• pp.67-77
• /
• 2014

This study investigated (1) whether Korean learners of English use discourse-level prosody in L2 production as native speakers of English do, and (2) whether discourse-level prosody is also found in the Korean language, as is evident in the prosody of native speakers of English. The study compared the production of the same 15 sentences in two types of reading materials, sentence-level and discourse-level. This study analyzed the onset pitch, sentence mean pitch and pause length to examine the paratone (intonational paragraph) realization in discourse-level speech. The results showed that in L2 discourse-level prosody, the Korean speakers were limited in displaying paratone and did not made significant difference between sentence-level and discourse-level prosody. On the other hand, in L1 discourse-level text, both English and Korean participants demonstrated paratone using pitch. However, there were differences in using prosodic cues between two groups. In using pauses, the ES group paused longer before both the orthographically marked and not marked topic sentences. The KS group paused longer only before the orthographically marked topic sentence in both L1 and L2 text reading. In the comparison of sentence-level and discourse-level prosody, the topic sentences were marked by different prosodic cues. English participants used higher sentence mean pitch, and the Korean participants used higher onset pitch.

• The Journal of Korean Physical Therapy
• /
• v.24 no.1
• /
• pp.29-34
• /
• 2012

Purpose: The purpose of this study was to evaluate the difference between respiratory function and speech production, according to the seat surface inclination while in the sitting position. Methods: Respiratory function (FVC, FEV1) and speech production (inspiratory frequency, unit reading time, paragraph reading time) were measured in 3 sitting conditions: horizontal seat surface, seat surface tilted forward 15 degrees, and seat surface tilted backward 15 degrees. Results: We found that the mean values of FVC and FEV1 were statistically significant different according to three types of sitting positions (p<0.05). The following result was observed: forward tilted sitting > horizontal sitting > backward tilted sitting. There was no significant difference in speech production between the different positions. Respiratory function and speech production had a significantly negative correlation in the forward tilted condition and the backward tilted condition. Conclusion: This finding suggests that the seat surface inclination have an effect on respiratory function. Especially, forward tilted sitting may be an effective posture that may help increases the respiratory function.

• Journal of information and communication convergence engineering
• /
• v.19 no.1
• /
• pp.48-53
• /
• 2021

Herein, we propose a document analysis system that analyzes papers or reports transformed into XML(Extensible Markup Language) format. It reads the document specified by the user, extracts keywords from the document, and compares the frequency of keywords to extract the top-three keywords. It maintains the order of the paragraphs containing the keywords and removes duplicated paragraphs. The frequency of the top-three keywords in the extracted paragraphs is re-verified, and the paragraphs are partitioned into 10 sections. Subsequently, the importance of the relevant areas is calculated and compared. By notifying the user of areas with the highest frequency and areas with higher importance than the average frequency, the user can read only the main content without reading all the contents. In addition, the number of paragraphs extracted through the deep learning model and the number of paragraphs in a section of high importance are predicted.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.326-329
• /
• 2019

기계독해를 실용화하기 위해 단락을 검색하는 검색 모델은 최근 기계독해 모델이 우수한 성능을 보임에 따라 그 필요성이 더 부각되고 있다. 그러나 기존 검색 모델은 질의와 단락의 어휘 일치도나 유사도만을 계산하므로, 기계독해에 필요한 질의 어휘의 문맥에 해당하는 단락 검색을 하지 못하는 문제가 있다. 본 논문에서는 이러한 문제를 해결하기 위해 Word2vec의 입력 단어열의 벡터에 해당하는 IN Weight Matrix와 출력 단어열의 벡터에 해당하는 OUT Weight Matrix를 사용한 단락 검색 모델을 제안한다. 제안 방법은 기존 검색 모델에 비해 정확도를 측정하는 Precision@k에서 좋은 성능을 보였다.