• Journal of Korean Elementary Science Education
• /
• v.41 no.1
• /
• pp.79-92
• /
• 2022

This study examines the propositional connections and markers for connecting propositions in the texts of the 3^rd- and 6^th-grade science textbooks developed under the 2015 Revised Science Curriculum. A selection of texts from Korean and social science textbooks were analyzed and compared to those from science textbooks as well. The propositional connections were classified into emphasis, elaboration, exemplification, listing, addition, order, correspondence, causal relation, condition, and purpose types. The markers for the relationship of propositions were classified as demonstrative, using conjunctive, using a comma, using distinctive linguistic elements, and no marker types. The results showed that the frequency of propositional connections in the texts of the 6^th-grade textbooks was lower than that of Korean and/or social science textbooks. However, the frequency of the propositional connections in the texts of the 3^rd-grade textbooks was found to be lower than that of the social science textbook but higher than that of the Korean textbook. The types of order, listing, condition, and causal relation were dominant in science as well as Korean and social science textbooks. Over 40% of the markers for the relationship of propositions were found to be the no marker type, with the ratio of the no marker type being especially higher in the categories of order and causal relation.

• Journal of the Korean Society for Library and Information Science
• /
• v.57 no.1
• /
• pp.93-114
• /
• 2023

Information extraction can facilitate the intensive analysis of documents by providing semantic triples which consist of named entities and their relations recognized in the texts. However, most of the research so far has been carried out separately for named entity recognition and relation extraction as individual studies, and as a result, the effective performance evaluation of the entire information extraction systems was not performed properly. This paper introduces two models of end-to-end information extraction that can extract various entity names in clinical records and their relationships in the form of semantic triples, namely pipeline and joint models and compares their performances in depth. The pipeline model consists of an entity recognition sub-system based on bidirectional GRU-CRFs and a relation extraction module using multiple encoding scheme, whereas the joint model was implemented with a single bidirectional GRU-CRFs equipped with multi-head labeling method. In the experiments using i2b2/VA 2010, the performance of the pipeline model was 5.5% (F-measure) higher. In addition, through a comparative experiment with existing state-of-the-art systems using large-scale neural language models and manually constructed features, the objective performance level of the end-to-end models implemented in this paper could be identified properly.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.10
• /
• pp.123-132
• /
• 2023

Git's commit messages manage the history of source changes during project progress or operation. By utilizing this historical data, project risks and project status can be identified, thereby reducing costs and improving time efficiency. A lot of research related to this is in progress, and among these research areas, there is research that classifies commit messages as a type of software maintenance. Among published studies, the maximum classification accuracy is reported to be 95%. In this paper, we began research with the purpose of utilizing solutions using the commit classification model, and conducted research to remove the limitation that the model with the highest accuracy among existing studies can only be applied to programs written in the JAVA language. To this end, we designed and implemented an additional step to standardize source change data into natural language using GPT. This text explains the process of extracting commit messages and source change data from Git, standardizing the source change data with GPT, and the learning process using the DistilBERT model. As a result of verification, an accuracy of 91% was measured. The proposed model was implemented and verified to ensure accuracy and to be able to classify without being dependent on a specific program. In the future, we plan to study a classification model using Bard and a management tool model helpful to the project using the proposed classification model.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.1
• /
• pp.42-50
• /
• 2023

The development of technology has achieved digital innovation throughout the media industry, including production techniques and editing technologies, and has brought diversity in the form of consumer viewing through the OTT service and streaming era. The convergence of big data and deep learning networks automatically generated text in format such as news articles, novels, and scripts, but there were insufficient studies that reflected the author's intention and generated story with contextually smooth. In this paper, we describe the flow of pictures in the storyboard with image caption generation techniques, and the automatic generation of story-tailored scenarios through language models. Image caption using CNN and Attention Mechanism, we generate sentences describing pictures on the storyboard, and input the generated sentences into the artificial intelligence natural language processing model KoGPT-2 in order to automatically generate scenarios that meet the planning intention. Through this paper, the author's intention and story customized scenarios are created in large quantities to alleviate the pain of content creation, and artificial intelligence participates in the overall process of digital content production to activate media intelligence.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.55-58
• /
• 2024

최근 각 기업의 AI 면접시스템 도입이 증가하고 있으며, AI 면접에 대한 실효성 논란 또한 많은 상황이다. 본 논문에서는 AI 면접 과정에서 지원자를 평가하는 방식을 시각, 음성, 자연어처리 3영역에서 구현함으로써, 면접 지원자를 다방면으로 분석 방법론의 적절성에 대해 평가하고자 한다. 첫째, 시각적 측면에서, 면접 지원자의 감정을 인식하기 위해, 합성곱 신경망(CNN) 기법을 활용해, 지원자 얼굴에서 6가지 감정을 인식했으며, 지원자가 카메라를 응시하고 있는지를 시계열로 도출하였다. 이를 통해 지원자가 면접에 임하는 태도와 특히 얼굴에서 드러나는 감정을 분석하는 데 주력했다. 둘째, 시각적 효과만으로 면접자의 태도를 파악하는 데 한계가 있기 때문에, 지원자 음성을 주파수로 환산해 특성을 추출하고, Bidirectional LSTM을 활용해 훈련해 지원자 음성에 따른 6가지 감정을 추출했다. 셋째, 지원자의 발언 내용과 관련해 맥락적 의미를 파악해 지원자의 상태를 파악하기 위해, 음성을 STT(Speech-to-Text) 기법을 이용하여 텍스트로 변환하고, 사용 단어의 빈도를 분석하여 지원자의 언어 습관을 파악했다. 이와 함께, 지원자의 발언 내용에 대한 감정 분석을 위해 KoBERT 모델을 적용했으며, 지원자의 성격, 태도, 직무에 대한 이해도를 파악하기 위해 객관적인 평가지표를 제작하여 적용했다. 논문의 분석 결과 AI 면접의 다면적 평가시스템의 적절성과 관련해, 시각화 부분에서는 상당 부분 정확도가 객관적으로 입증되었다고 판단된다. 음성에서 감정분석 분야는 면접자가 제한된 시간에 모든 유형의 감정을 드러내지 않고, 또 유사한 톤의 말이 진행되다 보니 특정 감정을 나타내는 주파수가 다소 집중되는 현상이 나타났다. 마지막으로 자연어처리 영역은 면접자의 발언에서 나오는 말투, 특정 단어의 빈도수를 넘어, 전체적인 맥락과 느낌을 이해할 수 있는 자연어처리 분석모델의 필요성이 더욱 커졌음을 판단했다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.3
• /
• pp.439-449
• /
• 2024

Recent rapid advancements in voice generation technology have enabled the natural synthesis of voices using text alone. However, this progress has led to an increase in malicious activities, such as voice phishing (voishing), where generated voices are exploited for criminal purposes. Numerous models have been developed to detect the presence of synthesized voices, typically by extracting features from the voice and using these features to determine the likelihood of voice generation.This paper proposes a new model for extracting voice features to address misuse cases arising from generated voices. It utilizes a deep learning-based audio codec model and the pre-trained natural language processing model BERT to extract novel voice features. To assess the suitability of the proposed voice feature extraction model for voice detection, four generated voice detection models were created using the extracted features, and performance evaluations were conducted. For performance comparison, three voice detection models based on Deepfeature proposed in previous studies were evaluated against other models in terms of accuracy and EER. The model proposed in this paper achieved an accuracy of 88.08%and a low EER of 11.79%, outperforming the existing models. These results confirm that the voice feature extraction method introduced in this paper can be an effective tool for distinguishing between generated and real voices.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.3
• /
• pp.148-155
• /
• 2024

Despite the increasing importance of irony and sarcasm detection in the field of natural language processing, research on the Korean language is relatively scarce compared to other languages. This study aims to experiment with various models for irony detection in Korean text. The study conducted irony detection experiments using KoBERT, a BERT-based model, and ChatGPT. For KoBERT, two methods of additional training on sentiment data were applied (Transfer Learning and MultiTask Learning). Additionally, for ChatGPT, the Few-Shot Learning technique was applied by increasing the number of example sentences entered as prompts. The results of the experiments showed that the Transfer Learning and MultiTask Learning models, which were trained with additional sentiment data, outperformed the baseline model without additional sentiment data. On the other hand, ChatGPT exhibited significantly lower performance compared to KoBERT, and increasing the number of example sentences did not lead to a noticeable improvement in performance. In conclusion, this study suggests that a model based on KoBERT is more suitable for irony detection than ChatGPT, and it highlights the potential contribution of additional training on sentiment data to improve irony detection performance.

• Journal of the Korean Society for information Management
• /
• v.41 no.1
• /
• pp.261-282
• /
• 2024

This study proposed and evaluated electroencephalography (EEG)-based and eye-tracking-based methods to determine relevance by utilizing users' implicit relevance feedback while navigating content in a digital library. For this, EEG/eye-tracking experiments were conducted on 32 participants using video, image, and text data. To assess the usefulness of the proposed methods, deep learning-based artificial intelligence (AI) techniques were used as a competitive benchmark. The evaluation results showed that EEG component-based methods (av_P600 and f_P3b components) demonstrated high classification accuracy in selecting relevant videos and images (faces/emotions). In contrast, AI-based methods, specifically object recognition and natural language processing, showed high classification accuracy for selecting images (objects) and texts (newspaper articles). Finally, guidelines for implementing a digital library interface based on EEG, eye-tracking, and artificial intelligence technologies have been proposed. Specifically, a system model based on implicit relevance feedback has been presented. Moreover, to enhance classification accuracy, methods suitable for each media type have been suggested, including EEG-based, eye-tracking-based, and AI-based approaches.

• Phonetics and Speech Sciences
• /
• v.16 no.3
• /
• pp.39-48
• /
• 2024

Reading fluency refers to the ability to read sentences or paragraphs accurately, quickly, and with appropriate prosodic expression. Most reading fluency assessments exclude expressive ability because it is difficult to objectively measure. Therefore, in this study, we examined all elements of reading fluency by analyzing prosodic characteristics of reading scenario texts to maximize expressive reading. The subjects were 30 male students in the first and second grades of middle school (15 normal and 15 poor readers). To analyze the accuracy aspect, error types at the syllable level were analyzed for each group, and related acoustic variables were measured and examined in terms of prosodic aspects. The reading accuracy analysis showed that the poor reading group had a higher error rate than the normal. In terms of error types, the normal group showed the order of 'substitution>omission>correction>insertion>repetition', whereas the poor reading group was in the order of 'correction>substitution>repetition/insertion>omission'. For the speech tempo, the dyslexic students were slower than the typical students for all sentence types. The prosodic variables also showed a high frequency of accentual phrases (AP) and intonation phrases (IP) in sentences along with a wide intensity range.

• Korean Journal of Cognitive Science
• /
• v.17 no.2
• /
• pp.75-102
• /
• 2006

Three experiments were conducted using a verification task to examine good and poor readers' generation of causal inferences(with because sentences) and contrastive inferences(with although sentences). The unfamiliar, critical verification statement was either explicitly mentioned or was implied. In Experiment 1, both good and poor readers responded accurately to the critical statement, suggesting that both groups had the linguistic knowledge necessary to the required inferences. Differences were found, however, in the groups' verification latencies. Poor, but not good, readers responded faster to explicit than to implicit verification statements for both because and although sentences. In Experiment 2, poor readers were induced to generate causal inferences for the because experimental sentences by including fillers that were apparently counterfactual unless a causal inference was made. In Experiment 3, poor readers were induced to generate contrastive inferences for the although sentences by including fillers that could only be resolved by making a contrastive inference. Verification latencies for the critical statements showed that poor readers made causal inferences in Experiment 2 and contrastive inferences in Experiment 3 doting comprehension. These results were discussed in terms of context effect: Specific encoding operations performed on anomaly backgrounded in another passage would form part of the context that guides the ongoing activity in processing potentially relevant subsequent text.