• KIISE Transactions on Computing Practices
• /
• v.22 no.7
• /
• pp.315-319
• /
• 2016

Recurrent neural network based language models (RNN LM) have shown improved results in language model researches. The RNN LMs are limited to post processing sessions, such as the N-best rescoring step of the wFST based speech recognition. However, it has considerable vocabulary problems that require large computing powers for the LM training. In this paper, we try to find the 1st pass N-gram model using word embedding, which is the simplified deep neural network. The class based language model (LM) can be a way to approach to this issue. We have built class based vocabulary through word embedding, by combining the class LM with word N-gram LM to evaluate the performance of LMs. In addition, we propose that part-of-speech (POS) tagging based LM shows an improvement of perplexity in all types of the LM tests.

• Journal of Information Technology Services
• /
• v.22 no.4
• /
• pp.141-157
• /
• 2023

The present work aimed to present the development of 'Funt', the augmented reality-based digital app as an educational tool for foreign language learning. Our work further evaluated the learning efficacy of the tool by the assessment of the three dependent measures including learning motivation, learning satisfaction, and learning involvement. With a learning app of 'Funt', students can use AR app to access recognition-based or location-based experiences such that any objects, artifacts, or media appear to be in the app. Students are then able to interact with the digital content by manipulating it to learn more about it. Students's engagement should also increase when they create their own experience in AR to demonstrate their understanding of a particular concept or words. Learning effects were evaluated on survey data collected from a hundred respondents aging six to nine years. One-group design for pre-test and post-test was utilized to examine the differences of learning efficacy by comparing the non-'Funt' group and the Funt group scores. A pairwise t-Test was performed for pairwise comparisons between two learning groups. The results indicate that the 'Funt' group scored significantly higher than the non-'Funt' group in the measures of learning motivation, learning satisfaction, and learning involvement. Overall, our results suggest that 'Funt' attracted the students' attention, provided them with a fun context to learn English vocabulary, and develop positive motivation and satisfaction towards vocabulary learning through AR technology.

• Hwankyungkyoyuk
• /
• v.16 no.2
• /
• pp.15-28
• /
• 2003

In effort to help vitalize environmental education which is the most efficient way to preserve environment and solve environmental problems and also to provide necessary basic data, this research was conducted on the primary and secondary school teachers in Daegu for their awareness of the elements of environmental education, for their interests in environment and environmental issues, for their sensitivity on the seriousness of the environmental issues and for their knowledge of environmental vocabulary. Followings are the results: 1. 96% of the teachers supported the necessity of school education on environment, but only 51% went for adopting environment as an independent subject. 2. The majority of 57% said that they came to recognize environment and environmental issues 'through media such as TV and radio'. For the desirable form of environmental education, 64% supported 'field study or experience activity'. As for the undermining factors, the majority of 50% cited 'excessive focus of school education on college entrance' and 29% 'limitations of class hours'. 3. With regard to their interests in environment and environmental issues, they were between 3.43～4.08 point range out of 5 points. For their sensitivity about the seriousness of environment and environmental problems, the survey showed the range of 3.49～4.28 points out of 5 points. 4. There was no remarkable difference in the level of recognition between male and female teacher. But, according to disparity of age, teachers who are in their forties and fifties recognized better than teachers in twenties and thirties. Also, there was a striking difference among primary school, middle school, and high school teachers. High school teachers had the highest recognition level, while, middle school teachers had the lowest recognition level.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.7
• /
• pp.312-318
• /
• 2006

The conventional lip-synch system has a two-step process, speech segmentation and recognition. However, the difficulty of speech segmentation procedure and the inaccuracy of training data set due to the segmentation lead to a significant Performance degradation in the system. To cope with that, the connected vowel recognition method using Head-Body-Tail (HBT) model is proposed. The HBT model which is appropriate for handling relatively small sized vocabulary tasks reflects co-articulation effect efficiently. Moreover the 7 vowels are merged into 3 classes having similar lip shape while the system is optimized by employing a class dependent SCHMM structure. Additionally in both end sides of each word which has large variations, 8 components Gaussian mixture model is directly used to improve the ability of representation. Though the proposed method reveals similar performance with respect to the CHMM based on the HBT structure. the number of parameters is reduced by 33.92%. This reduction makes it a computationally efficient method enabling real time operation.

• The Journal of the Convergence on Culture Technology
• /
• v.6 no.3
• /
• pp.185-191
• /
• 2020

This study aims to understand the differences in cultural and emotional perceptions about the color image of public healthcare facilities in Romania, an Eastern European country that is relatively lacking in recognition but is gradually expanding trade. For this, color images were selected through a review of previous studies, and a questionnaire survey was constructed based on the colorimetric data by visiting 8 public healthcare facilities such as medical facilities, 4 social sports facilities, and 8 nursing facilities. An online survey was conducted on the color image of public facilities with 89 Koreans and 86 Romanians, and frequency and cross-analysis was conducted using the SPSS statistical analysis program to examine the color images of public healthcare facilities of Koreans and Romanians. The difference in perception was identified. As a result, it was found that there was a statistically significant difference in the perception of color images of public healthcare facilities between countries in vocabulary evaluation and image evaluation, and this was interpreted as different meanings for groups residing in different cultures. Therefore, it implies that cultural differences in perception should be considered when establishing an environment related to this in the future.

• Journal of Korea Game Society
• /
• v.9 no.4
• /
• pp.67-80
• /
• 2009

In this paper, we described the development of a Korean Manual Alphabet (KMA) learning game with avatar. KMA letters correspond to the vocabulary of Korean Sign Language (KSL) when spelling a word. Each KMA letter corresponds to a letter of the Korean Alphabet (KA) and KA is represented as hand shapes by sign language user. We developed a KMA learning game for a beginner to learn KMA letters from sign language avatar and practice KMA presentation easily. The system composed of sign language teacher avatar GUI popup window based on OpenGL, KMA letter recognition module, KA letter raining game module and USB camera. A user learns a KMA letter with expressing KA syllabic from avatar and inputs a KMA letter to the system using USB camera. We evaluated the efficiency of the developed system through the verification of users.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.15-23
• /
• 2022

In the field of patents, as NLP(Natural Language Processing) is a challenging task due to the linguistic specificity of patent literature, there is an urgent need to research a language model optimized for Korean patent literature. Recently, in the field of NLP, there have been continuous attempts to establish a pre-trained language model for specific domains to improve performance in various tasks of related fields. Among them, ELECTRA is a pre-trained language model by Google using a new method called RTD(Replaced Token Detection), after BERT, for increasing training efficiency. The purpose of this paper is to propose KorPatELECTRA pre-trained on a large amount of Korean patent literature data. In addition, optimal pre-training was conducted by preprocessing the training corpus according to the characteristics of the patent literature and applying patent vocabulary and tokenizer. In order to confirm the performance, KorPatELECTRA was tested for NER(Named Entity Recognition), MRC(Machine Reading Comprehension), and patent classification tasks using actual patent data, and the most excellent performance was verified in all the three tasks compared to comparative general-purpose language models.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.81-84
• /
• 1999

본 논문에서는 중규모급 단어인식기의 실시간 구현을 위한 무감독 단어집단화 알고리듬을 제안한다. 무감독 단어집단화는 인식대상 어휘 수가 많은 대용량 음성인식 시스템에서 대상 어휘 수를 줄여주는 역할을 하는 전처리기의 성격을 갖는다. 무감독 집단화를 위해 각 단어의 유$\cdot$무성음 고유의 특성을 잘 반영할 수 있는 특징 파라미터 5개를 사용하여 패턴 인식과 회귀분석에서 널리 사용되고 있는 분류$\cdot$회귀트리(Classification And Regression Tree)에 적용시키는 방법으로 접근하였고, 각 단어의 frame 수를 일정하게 n개로 분할(segment)하여 1개의 tree를 생성시키는 방법과 각 segment에 해당하는 tree를 생성시켜 segment들 사이의 교집합 성분으로 단어들을 집단화 하였다 실험결과 탐색 대상단어 22개에서 평균2.21개로 줄어 전체 대상 단어의 $10\%$만을 탐색하여 인식할 수 있는 방법을 제시할 수 있었다.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.3
• /
• pp.248-252
• /
• 2004

Cultural, social, economic and other various environmental factors affect our language and different words and terminology are used and coined for different contexts, resulting in quantitative change of vocabulary. This paper presents an automatic pronunciation generator using selection procedure for exceptional pronunciation words from added text corpus, which reflects this dynamic nature of language. For our experiment, we used the text corpus released by ETRI for speech recognition. consisting or 53,750 sentences (740.497 Eojols), and obtained a 100% performance level of the proposed automatic pronunciation generator.

• English Language & Literature Teaching
• /
• v.10 no.3
• /
• pp.41-66
• /
• 2004

Under the textbook-driven approach to English education in the Korean selling, the importance of the English textbook can not be overemphasized as the main source of learning materials. Recently, with the development of computer-based language corpora, the recognition of the importance of collocations and the availability of computerized databases of words have caused a resurgence and facilitation in the instruction of collocation. The primary purpose of the present study is to identify the characteristics of lexical collocation and the extent of its use in high school 10th-grade textbooks. From all the analyses, it is revealed that the language materials reflect various constructed collocation in the case of adjective+noun and noun+noun collocations in a natural context. However, verb+noun and adverb+verb collocations are not fully reflected. This is true for delexicalized verbs, and verb and adjective intensifiers. Also the language materials do not provide sufficient support for the lexical syllabus, even though all textbooks may be somewhat adequate in terms of vocabulary size. Finally, based on the analyses of the texts, the suggestions for English collocation instruction are made in the lexical approach.