• Phonetics and Speech Sciences
• /
• v.15 no.3
• /
• pp.75-82
• /
• 2023

Automatic speech recognition (ASR) has been revolutionized with deep learning-based approaches, among which self-supervised learning methods have proven to be particularly effective. In this study, we aim to enhance the performance of OpenAI's Whisper model, a multilingual ASR system on the Korean language. Whisper was pretrained on a large corpus (around 680,000 hours) of web speech data and has demonstrated strong recognition performance for major languages. However, it faces challenges in recognizing languages such as Korean, which is not major language while training. We address this issue by fine-tuning the Whisper model with an additional dataset comprising about 1,000 hours of Korean speech. We also compare its performance against a Transformer model that was trained from scratch using the same dataset. Our results indicate that fine-tuning the Whisper model significantly improved its Korean speech recognition capabilities in terms of character error rate (CER). Specifically, the performance improved with increasing model size. However, the Whisper model's performance on English deteriorated post fine-tuning, emphasizing the need for further research to develop robust multilingual models. Our study demonstrates the potential of utilizing a fine-tuned Whisper model for Korean ASR applications. Future work will focus on multilingual recognition and optimization for real-time inference.

• Journal of Digital Policy
• /
• v.2 no.4
• /
• pp.1-10
• /
• 2023

This study conducted preliminary study to identify effective ways to use ChatGPT in traffic policing by analyzing ChatGPT's responses to the driver's license test and the road traffic accident appraiser test. I collected ChatGPT responses for the driver's license test item pool and the road traffic accident appraiser test using the OpenAI API with Python code for 30 iterative experiments, and analyzed the percentage of correct answers by test, year, section, and consistency. First, the average correct answer rate for the driver's license test and the for road traffic accident appraisers test was 44.60% and 35.45%, respectively, which was lower than the pass criteria, and the correct answer rate after 2022 was lower than the average correct answer rate. Second, the percentage of correct answers by section ranged from 29.69% to 56.80%, showing a significant difference. Third, it consistently produced the same response more than 95% of the time when the answer was correct. To effectively utilize ChatGPT, it is necessary to have user expertise, evaluation data and analysis methods, design a quality traffic law corpus and periodic learning.

• Journal of Intelligence and Information Systems
• /
• v.29 no.4
• /
• pp.165-184
• /
• 2023

Recently, Deep learning analysis of unstructured text data using language models, such as Google's BERT and OpenAI's GPT has shown remarkable results in various applications. Most language models are used to learn generalized linguistic information from pre-training data and then update their weights for downstream tasks through a fine-tuning process. However, some concerns have been raised that privacy may be violated in the process of using these language models, i.e., data privacy may be violated when data owner provides large amounts of data to the model owner to perform fine-tuning of the language model. Conversely, when the model owner discloses the entire model to the data owner, the structure and weights of the model are disclosed, which may violate the privacy of the model. The concept of offsite tuning has been recently proposed to perform fine-tuning of language models while protecting privacy in such situations. But the study has a limitation that it does not provide a concrete way to apply the proposed methodology to text classification models. In this study, we propose a concrete method to apply offsite tuning with an additional classifier to protect the privacy of the model and data when performing multi-classification fine-tuning on Korean documents. To evaluate the performance of the proposed methodology, we conducted experiments on about 200,000 Korean documents from five major fields, ICT, electrical, electronic, mechanical, and medical, provided by AIHub, and found that the proposed plug-in model outperforms the zero-shot model and the offsite model in terms of classification accuracy.

• Journal of Embryo Transfer
• /
• v.33 no.3
• /
• pp.149-157
• /
• 2018

The objective of this study was to evaluate the effect of treating dairy cattle with exogenous human chorionic gonadotrophin (hCG), five (5) days post artificial insemination (AI) on serum progesterone (P4) concentration and pregnancy rate. In this experiment, five days after AI, cows were assigned randomly to two groups namely: a) treated group (67) which were administrered with 1500 IU hCG (Chorulon) and b) control group (61), which received no treatment. On day 5, 10, 15 and 20 after the artificial insemination, blood samples from a total of 8 cows (4 from each group) were collected and were analyzed for serum P4 concentration. Cows were detected for estrus according to standing heat by visual observation. Cows that were detected still in estrus after days 18-24 were re-inseminated and recorded as not pregnant (open). Pregnancy diagnosis was conducted by ultrasonographic examination and transrectal palpation of the uterus on approximately 60 days in cows that observed to be not in estrus. The conception rate in hCG treated and control groups were 52.5 and 36.1%, respectively. The results proved that there were no significant differences in conception rate between two groups (p=0.0568). However, pregnancy rates were reduced by hCG treatment. Average serum P4 concentrations did not differ between Hcg-treated and control groups on day 5 (0.377 versus 0.375 ng/ml). On day 20 serum P4 concentrations were greater in the treated group compared with the control group (3.085 versus 2.010 ng/ml). The treatment with hCG seemed to increase P4 level compared with the control. In conclusion, the results of this study showed that 1500 IU of hCG administered on 5 day post AI increased conception rate in dairy cows. This was supported by the results on serum P4 concentration which was greater in hCG treated group.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.36 no.4
• /
• pp.601-606
• /
• 2009

Amelogenesis Imperfecta (AI) is a genetic disorder which retards the development of enamel and it can be classified into three types: hypoplastic, hypomaturation, hypocalcified type. This can occur both in deciduous and permanent dentition. A 8 year 8 month old patient with a chief complaints of delayed eruption on upper anteriors, calculus deposit on lower anteriors and anterior openbite visited the clinic. Anteriors had thin layer of enamel and were very narrow. Especially lower anteriors had rough surface and were in bad shape. Teeth were very hypersensitive to thermal changes. Upper and lower first molars showed severe attrition on the occlusal surface. Radiographs also verified hypoplastic enamel in the whole dentition including the teeth in the tooth bud. The patient was diagnosed as hypoplastic AI, and is being treated at the pediathc and prosthodontic department of the Kyunghee dental university hospital. To improve the function, esthetics, hypersensitivity of the AI patients, restorations on the posteriors and the anteriors with oral hygiene instruction are necessary, Constant follow-up check is needed until full growth and after full growth, cooperative care with the other department is needed.

• Journal of Convergence for Information Technology
• /
• v.11 no.3
• /
• pp.229-235
• /
• 2021

Artificial intelligence technology is influencing across all areas as technology advances and social needs change. Therefore, Korean universities have actively developed and operated classes related to artificial intelligence, and have emphasized the importance of artificial intelligence not only in major education but also in liberal arts education. However, there is a lack of research on the development of educational methods and educational programs because artificial intelligence education in liberal arts is in its early stages. SSI is an education that can apply social and ethical problems related to science to open problems that can creatively and reasonably present solutions. SSI can be applied to make AI education more effective. In this study, an artificial intelligence liberal arts curriculum applied SSI was developed with three purposes: First, it is designed is designed so that students subject to education can access it by considering its characteristics as actors of the intelligent information society. Second, it is designed so that students can experience artificial intelligence programs themselves and deal with science technology and social relevance in depth, focusing on various examples of real life. Third, it is designed and approached so that students can participate and cooperate for the purpose of solving common problems to develop cooperative problem-solving skills.

• Journal of Industrial Convergence
• /
• v.20 no.4
• /
• pp.23-30
• /
• 2022

Recently, with the development of artificial intelligence (AI) and deep learning, the importance of conversational artificial intelligence chatbots is being highlighted. In addition, chatbot research is being conducted in various fields. To build a chatbot, it is developed using an open source platform or a commercial platform for ease of development. These chatbot platforms mainly use RNN and application algorithms. The RNN algorithm has the advantages of fast learning speed, ease of monitoring and verification, and good inference performance. In this paper, a method for improving the inference performance of RNNs and applied algorithms was studied. The proposed method used the word group expansion learning technique of key words for each sentence when RNN and applied algorithm were applied. As a result of this study, the RNN, GRU, and LSTM three algorithms with a cyclic structure achieved a minimum of 0.37% and a maximum of 1.25% inference performance improvement. The research results obtained through this study can accelerate the adoption of artificial intelligence chatbots in related industries. In addition, it can contribute to utilizing various RNN application algorithms. In future research, it will be necessary to study the effect of various activation functions on the performance improvement of artificial neural network algorithms.

• Korean Journal of Culture and Social Issue
• /
• v.28 no.3
• /
• pp.513-527
• /
• 2022

This study was conducted to examining the socio-cultural impact of the COVID-19 pandemic that swept the world around 2020, and the transformation of norms and social problems due to COVID-19. For this, the characteristics of changes in the socio-cultural norms of the 14th century European Black Death, a representative example of the pandemic, were derived, and based on this, the COVID-19 pandemic was analyzed. The Black Death served as an opportunity to change social norms based on the existing religious authority and the power of the feudal system to the Enlightenment. The population declination and labor shortage also promoted commercialization and mechanization. Printing, which spread during this period, led to the popularization of knowledge, which raised the level of thinking and led to epochal scientific development. This became the foundation of the Industrial Revolution. Like the recent Black Death, COVID-19 has triggered changes in social norms. The technological environment of metaverse, a mixture of virtual and reality, has changed the norm of a consistent identity into free and open identities exerting various potentials through alternate characters. In addition, meme, which are about people being friendly to those with the same worldview as him on the metaverse, weakened the sense of isolation in non-face-to-face situations. Artificial intelligence (AI), which developed during the COVID-19 pandemic, has entered the stage of being used for creative activities beyond the function of assisting humans. Discussions were held on what new social problems would be created by the social norms changed due to the COVID-19 pandemic.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.1
• /
• pp.71-79
• /
• 2024

ChatGPT, an interactive artificial intelligence (AI) chatbot developed by Open AI in the U.S., gaining popularity with great repercussions around the world. Some academia are concerned that ChatGPT can be used by students for plagiarism, but ChatGPT is also widely used in a positive direction, such as being used to write marketing phrases or website phrases. There is also an opinion that ChatGPT could be a new future for "search," and some analysts say that the focus should be on fostering rather than excessive regulation. This study analyzed consciousness about ChatGPT for college students through a survey of their perception of ChatGPT. And, plagiarism inspection systems were prepared to establish an education support model using ChatGPT and ChatGPT. Based on this, a university education support model using ChatGPT was constructed. The education model using ChatGPT established an education model based on text, digital, and art, and then composed of detailed strategies necessary for the era of the 4th industrial revolution below it. In addition, it was configured to guide students to use ChatGPT within the permitted range by using the ChatGPT detection function provided by the plagiarism inspection system, after the instructor of the class determined the allowable range of content generated by ChatGPT according to the learning goal. By linking and utilizing ChatGPT and the plagiarism inspection system in this way, it is expected to prevent situations in which ChatGPT's excellent ability is abused in education.

• Journal of the Korean Home Economics Association
• /
• v.47 no.6
• /
• pp.91-108
• /
• 2009

The aim of this study was to assist the development of parenting-aid policies by analyzing information from life perspectives, family system theory, and integrated family policy perspectives. For this purpose, we conducted a literature using a variety of sources including internet sites, laws, published material related to current parenting-aid policies. Based on analyses we formulated four agendas: empowering parent competency, supporting diverse family parenting, guaranteeing parent's rights, and supporting parenting of dual-income families parenting. The results were as follow: Firstly, for empowering parent competency, we proposed the development of family relationship empowerment programs including family life cycle and children-raising support services. Secondly, for diverse family parenting, we proposed access to integrated parenting-aid programs dependent on family types. Thirdly, for guaranteeing parent’s rights, we proposed the supporting of child raising costs and tax returns to families with children, and provide child allowances. Finally, for dual-income family parenting, we proposed that schools and communities must recognize that dual-income families are a universal family type, establish family friendly business culture, and to develop a more democratic domestic family-relationship. This policy proposed a new paradigm where parents must be recognized as partners and stakeholder in development of family related policies. Ultimately, such policies will contribute to increased birth rate and development of a more respectful society.