• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1252-1257
• /
• 2017

In dialogues, speakers' intentions can be represented by sets of an emotion, a speech act, and a predicator. Therefore, dialogue systems should capture and process these implied characteristics of utterances. Many previous studies have considered such determination as independent classification problems, but others have showed them to be associated with each other. In this paper, we propose an integrated model that simultaneously determines emotions, speech acts, and predicators using a convolution neural network. The proposed model consists of a particular abstraction layer, mutually independent informations of these characteristics are abstracted. In the shared abstraction layer, combinations of the independent information is abstracted. During training, errors of emotions, errors of speech acts, and errors of predicators are partially back-propagated through the layers. In the experiments, the proposed integrated model showed better performances (2%p in emotion determination, 11%p in speech act determination, and 3%p in predicator determination) than independent determination models.

• Journal of Digital Convergence
• /
• v.20 no.3
• /
• pp.53-61
• /
• 2022

The purpose of this study was to test whether or not the AI chatbot is effective in acquiring target grammar for Korean EFL college students: prepositions and articles. A quasi-experiment was conducted with 46 first-year students taking part in a required English course. They were randomly divided into two groups: the experimental and control groups (23 students for each, respectively). The experimental group was engaged in six chat sessions with a chatbot over 6 weeks. A pretest and a posttest were used to examine the effectiveness of the chatbot by comparing any changes made in error frequencies of the target grammar in participants' English compositions. The results show that after a conversation with the chatbot, the experimental group significantly reduced the mean of omission errors in both prepositions and articles. To have a great effect in other error categories, chatbot feedback needs to be improved to reduce short responses or inaccurate utterances of students and induce them to actively participate in the conversation.

• The Journal of Korean Association of Computer Education
• /
• v.5 no.3
• /
• pp.127-133
• /
• 2002

In this paper a web-based real-time education system, which is able to support education through multimedia, is suggested for the expansion of learner's creative ability in the school. This system is designed so that it can support three things: 1) a real time interaction between instructors and learners, 2) individual learning through such an interaction, and 3) a coercive distribution of display by instructions for preventing the deviation of learners from learning. Also, this system, which UML is applied to, makes efficient interaction possible through the module for the real-time exchange and management of messages even in the multi-user environment. Through this system, not only the simulation by learners can be made for experiments and practices, but also Questions and respondence can be supported on the procedure of experiments and the analysis of their results. This system is built on constructivism, and aimed at helping the learning progress and knowledge formation of learners.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.518-523
• /
• 2021

챗봇과 같은 대화형 에이전트 사용이 증가하면서 채팅에서의 혐오 표현 사용도 더불어 증가하고 있다. 혐오 표현을 자동으로 탐지하려는 노력은 다양하게 시도되어 왔으나, 챗봇 데이터를 대상으로 한 혐오 표현 탐지 연구는 여전히 부족한 실정이다. 이 연구는 혐오 표현을 포함한 챗봇-사용자 대화 데이터 35만 개에 한국어 말뭉치로 학습된 KoELETRA 기반 혐오 탐지 모델을 적용하여, 챗봇-사람 데이터셋에서의 혐오 표현 탐지의 성능과 한계점을 검토하였다. KoELECTRA 혐오 표현 분류 모델은 챗봇 데이터셋에 대해 가중 평균 F1-score 0.66의 성능을 보였으며, 오탈자에 대한 취약성, 맥락 미반영으로 인한 편향 강화, 가용한 데이터의 정확도 문제가 주요한 한계로 포착되었다. 이 연구에서는 실험 결과에 기반해 성능 향상을 위한 방향성을 제시한다.

• 한국지구과학회:학술대회논문집
• /
• 2010.04a
• /
• pp.25-25
• /
• 2010

본 연구는 에너지와 물질 순환을 주제로 한 대학생들의 집단적 과제 수행 상황에서 구성원들의 상호작용 기능과 요소에 따른 역할 분배의 양상을 밝히고자 하였다. 집단 내 구성원에게 개별적인 역할이 주어져 있지 않은 상태에서 과제가 부여되었을 때, 집단 내에서 자연스럽게 형성되는 역할의 종류를 상호작용 기능과 양상에 따라 유형화해보았다. 연구 참여자는 9명의 환경 교육 전공 대학생으로, 2008년 11월에 열린 한중일 환경교육 캠프에 참여한 한국 학생들이다. 캠프 활동 중 하나로 충남 홍성 문당리 환경농업마을의 에너지와 물질 순환에 대해 토의하고 그 결과를 그림으로 표현하여 발표하는 활동이 있었는데, 본 연구에서는 발표를 위해 준비하는 과정을 연구 맥락으로 하였다. 그리고 약 2시간에 걸친 과제 수행 과정을 촬영 후 전사하여 연구 자료로 삼았다. 관련 선행 연구 고찰과 연구 자료에의 적용 및 수정을 통해 상호작용 기능과 요소를 코딩하기 위한 분석틀을 마련하였으며, 문제가 제기되고 그 문제에 대한 합의가 이루어지는 대화를 분석의 단위로 정하였다. 대화 단위별로 구성원 간 말차례가 오고간 모습을 화살표로 도식화하여 대화 패턴을 분류하였으며, 대화를 말차례별로 상호작용 기능과 요소에 따라 코딩화하였다. 화살표로 도식화된 대화 패턴과 상호작용 코딩 결과에 따라 집단 내에서 자발적으로 형성되는 구성원의 역할을 분류하고 명명하였다. 이렇게 집단적 과제 수행 상황에서 학생들에 의해 자생되는 역할을 상호작용의 관점에서 유형화해봄으로써 협동 학습이나 토의 활동 등 집단적 상호작용을 통한 과제 수행에 있어서 학생이나 교사의 역할에 대해 시사점을 줄 수 있을 것이다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.12 no.2
• /
• pp.119-128
• /
• 2006

As dialogue systems are widely required, the research on natural language generation in dialogue has raised attention. Contrary to conventional dialogue systems that reply to the user with a set of predefined answers, a newly developed dialogue system generates them dynamically and trains the answers to support more flexible and customized dialogues with humans. This paper proposes an evolutionary method for generating sentences using interactive genetic programming. Sentence plan trees, which stand for the sentence structures, are adopted as the representation of genetic programming. With interactive evolution process with the user, a set of customized sentence structures is obtained. The proposed method applies to a dialogue-based train booking agent and the usability test demonstrates the usefulness of the proposed method.

• Journal of the Korean earth science society
• /
• v.37 no.4
• /
• pp.243-268
• /
• 2016

The purpose of this study was to explore the effects of Cogenerative Dialogues embedded in a modeling-centered science learning and instruction on 7th grade female $students{\acute{i}}$ understanding of scientific models and modelling A total of 49 7th grade female students in two classrooms participated in a series of five modeling-centered science lessons, and 17 students volunteered to participate in this study. Participating students were divided into four groups, and two groups were randomly assigned to a treatment group who were asked to participate in Cogenerative Dialogues after each lesson, while the others, a control group, who did not. For data analysis, Upmeier and $Kr{\ddot{u}ger^{\prime}s$ framework was used to explore $participants{\acute{i}}$ understanding of model, and a revised $Baek{\acute{i}}s$ framework was used to examine $participants{\acute{i}}$ modeling process. Data analysis indicated that students who participated in Cogenerative Dialogues generally showed richer understanding of scientific models, as well as modeling, than the others who did not. This study suggests that Cogenerative Dialogues can be used as an educationally meaningful method for science educators to encourage students actively participate in a whole process of science instruction and learning, which assists them to increase their understanding not only of scientific models and modeling specifically but also of the nature and processes of scientific practice in general.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04b
• /
• pp.711-713
• /
• 2000

본 논문은 멀티미디어의 특성을 활용하여 초등학교 5학년의 음악과 전 영역을 멀티미디어 CAI로 설계하고 구현함으로써 음악에 대한 기초개념과 악곡에 대한 이해와 관심을높이고 학습자에 대한 개별학습과 교수-학습의 효율성을 높이고자 하였다. 이에 따라서 멀티미디어 CAI에 관한 이론을 탐색하고 CAI의 설계원리 및 교과의 특성을 고려하여 Hannafin와 Peck이 제안한 코스웨어 설계모형을 모델로 하였다. 설계모델에 따라 교육과정을 분석하여 멀티미디어 적용요소를 추출하였으며 저작환경에 적합한 스토리보드 형식을 작성하였다. CAI 코스웨어 설계모형에 준거하여 객체지향적이고 상호대화적인 접근을 가능하도록 멀티미디어 디렉터를 도구로 사용하였다. 본 CAI 프로그램은 멀티미디어(Text, Image, Graphic, animation, sound)를 활용하여 주의집중과 동기유발을 높혔고 특히 가창, 기악, 창작, 감상, 이론적 내용, 형성평가의 모든 음악적 영역을 교육과정의 내용에 일치시켜 충실한 교수-학습이 이루어지게 하여 모든 교사의 현장수업에 대한 부담감을 감소시켰다.

• Journal of Korean language education
• /
• v.28 no.2
• /
• pp.235-264
• /
• 2017

The purpose of this study is to explore and categorize the rhetorical strategies of text construction in research articles and to provide data for academic writing education for foreign graduate students. This study analyzes 30 research articles by Korean writers from Korean language and Korean language education fields, and categorizes the rhetorical strategies according to the roles of the writer as a RA form composer, a manager of research content, and a communicator. On the basis of the strategies, this study analyzes 18 term papers of foreign graduate students and inspects their weaknesses in using the rhetorical strategies. Based on the results of analysis, this study suggests rhetorical strategy education for KAP learners that emphasizes validity and clarifies argument along with attracting readers.

• Journal of Korean Elementary Science Education
• /
• v.41 no.1
• /
• pp.93-105
• /
• 2022

This study investigated the aspects of how three preservice elementary school teachers conducted the data-text transforming discourses in their science simulation teaching and how their epistemological conversations worked for learners' construction of scientific knowledge. Three preservice teachers, who had presented simulation teaching on the seasonal change of constellations, participated in the study. The results revealed that one preservice teacher, who had implemented the transforming discourses of data-to-evidence and model-to-explanation, appeared to facilitate learners' knowledge construction. The other two preservice teachers had difficulty helping learners construct science knowledge due to their lack of transforming discourses. What we should consider for improving preservice elementary school teachers' teaching competencies was discussed based on a detailed comparison of three cases of preservice teachers' data-text transforming.