• Journal of The Korean Association For Science Education
• /
• v.34 no.3
• /
• pp.273-283
• /
• 2014

The purpose of this study was to examine the potential of the use of the Classroom Response System (CRS), a kind of new ICT medium, in a quiz problem-solving oriented high school life science class. To find the usefulness of CRS as a teaching and learning strategy, the CRS group (n=34) sent prompt individual answers to the teachers' questions using the CRS terminal (Clicker), and the teacher then asked additional reasons of the individuals and gave personalized feedback. In the control group (n=35), the CRS was not used while the teacher asked overall questions and gave feedback in an undifferentiated way. As a result, the CRS increased students' interest and concentration during class, but there were no significant differences in study achievement between the two groups. However, there were significant differences between the medium-level groups when the two groups were divided into smaller ones based on their pre-scores. We suggest that, for effective use of the CRS for diagnostic and formative assessment, teachers should develop a teaching and learning strategy that can produce appropriate questions of various levels in advance, investigate the exact reasons for students' answers, and give customized feedback by individual as much as possible.

• Journal of The Korean Association For Science Education
• /
• v.44 no.1
• /
• pp.119-130
• /
• 2024

This study explores the possibility of using a generative artificial intelligence, ChatGPT, for student assessment in science subjects. In order to achieve our goal, we developed assessment items, collected students' responses, and input them into ChatGPT to implement the assessment procedures. Subsequently, we shared the assessment results from ChatGPT with science teachers and compared them to the teachers' assessment process to investigate the use of ChatGPT in student assessment. Regarding the results, in terms of setting the scoring rubric, we found the rubric generated by ChatGPT to be generally appropriate. However, the consistency between the scoring results obtained from ChatGPT and those determined by the teachers was relatively low. This inconsistency was more pronounced in items with additional assessment components and a more intricate rubric. In regard to feedback on student responses, there were some instances where the feedback generated was scientifically incorrect or beyond the scope of the curriculum, but there were also some positives, such as the provision of exemplary answers to questions and additional examples that helped students learn further. From these results, the teachers perceived limitations in using ChatGPT to conduct assessment in terms of reliability, which is considered crucial in student assessment, but suggested that it could be used to support assessment. Finally, synthesizing these findings, implications for utilizing ChatGPT in student assessment were suggested.

• Journal of Korean Association for Spatial Structures
• /
• v.9 no.4
• /
• pp.49-59
• /
• 2009

The seismic responses of small-scale spatial frames such as school gymnasiums are usually evaluated using static analysis, although time-history analysis should be carried out to fully incorporate the dynamic responses of the structures against seismic motions. In this study, advanced static analysis procedures arc presented for school gymnasiums that will improve the performance evaluation against seismic motions. The seismic loads are approximated by equivalent static loads corresponding to the two performance levels; i.e., Levels 1 and 2 defined by the Japanese building standard. The importance of utilizing the eigenmode in the load pattern is discussed. Simple static analysis procedures are presented for evaluation of maximum vertical acceleration. It is shown that the static analysis for Level 2 input significantly underestimates the responses by dynamic analysis; however, the inelastic responses for Level 2 are shown to be successfully evaluated using the equivalent linearization that is similar to the $^{\circ}$Dmethod based on calculation of limit strength$^{\circ}{\pm}$ for building frames in Japan.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2006.05a
• /
• pp.268-271
• /
• 2006

Combustion response function of a solid propellant is measured and calculated to study and model the feedback process between acoustic waves and combustion field. Standard Pulsed DB/AB method and related one-dimensional approximate analysis of T-burner are used to obtain the response function at a driving natural frequency. The problems related with simultaneous ignition of propellant samples are also mentioned and treated.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.14 no.1
• /
• pp.1-9
• /
• 2010

The actual hysteretic behavior of structural elements and systems is closer to smooth hysteretic behavior than piece-wise linear behavior. This paper presents a methodology for computing the constant-ductility inelastic response spectra for smooth hysteretic behaviors. The effect of the hysteretic smoothness on the inelastic response spectra for acceleration, displacement, and input energy is evaluated. The results indicate that increasing smoothness in the hysteretic behavior decreases the inelastic response spectra.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2010.04a
• /
• pp.145-148
• /
• 2010

항만 구조물의 건전성 평가 기술의 개발을 위한 기초 연구로서, 강제진동해석을 통하여 케이슨 구조-지반 경계부의 손상에 대한 진동응답을 분석하고자 한다. 이를 위해 세 단계의 연구를 수행하였다. 첫째, 케이슨 구조물의 진동특성 분석을 위해 시간영역기반의 AR(auto-regressive)모델을 선정하였다. 둘째, 모형 케이슨 구조물을 대상으로 진동응답 계측실험을 수행하였으며, AR-모델을 통해 진동특징을 실험적으로 분석하였다. 셋째, 대상 케이슨 시스템의 유한요소모델을 구성하고, 구조-지반 경계부의 손상에 따른 동적응답 특성의 변화를 수치적으로 분석하였다. 이를 위해 강제진동을 모사 하였으며, 구조-지반 경계부의 강성변화에 따른 케이슨 구조물의 진동응답의 변화를 분석하였다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.269-274
• /
• 2023

질의응답 (Question Answering)은 주어진 질문을 이해하여 그에 맞는 답변을 생성하는 자연어 처리 분야의 핵심적인 기계 독해 작업이다. 현재 대다수의 자연어 이해 작업은 사전학습 언어 모델에 미세 조정 (finetuning)하는 방식으로 학습되고, 질의응답 역시 이러한 방법으로 진행된다. 하지만 미세 조정을 통한 전이학습은 사전학습 모델의 크기가 커질수록 전이학습이 잘 이루어지지 않는다는 단점이 있다. 게다가 많은 양의 파라미터를 갱신한 후 새로운 가중치들을 저장하여야 한다는 용량의 부담이 존재한다. 본 연구는 최근 대두되는 deep prompt tuning 방법론을 한국어 추출형 질의응답에 적용하여, 미세 조정에 비해 학습시간을 단축시키고 적은 양의 파라미터를 활용하여 성능을 개선했다. 또한 한국어 추출형 질의응답에 최적의 prompt 길이를 최적화하였으며 오류 분석을 통한 정성적인 평가로 deep prompt tuning이 모델 예측에 미치는 영향을 조사하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.3 no.4
• /
• pp.24-31
• /
• 1993

원전 증기발생기의 수위제어기를 퍼지제어기로 구성하는 경우 증기발생기의 단독운전이 아닌 전체발전소의 가동중에 소속함수 조정이 이루어져야 하는 제약환경을 고려할 때 많은 반복실행횟수를 필요로 하는 임의적인 시행착오방법이나 일반화되어 있지 않은 복잡한 방법보다는 보다 간단명료하고 예측가능한 조정경로를 통해 소속함수를 조정할 필요가 있다. 여기에서는 그 방안으로서 프로세스 제어응답의 최대초과량 상승시간의 합으로 평가함수를 도입하고 descent method를 이용하여 제어응답의 결과로 얻는 평가함수의 최소점을 따라 소속함수를 조정해가는 방법을 제시하였다.

• Proceedings of the KIEE Conference
• /
• 2000.07d
• /
• pp.2953-2955
• /
• 2000

본 논문에서는 전문가의 지식이 없는 상황에서 자동적으로 최적의 퍼지 제어기를 설계하는 방법에 대해 연구한다. 먼저 퍼지 제어기의 규칙 설정을 위해 기존의 PID 제어기의 입출력 데이터를 클러스터링한다. 군집된 데이터들로부터 클러스터의 수를 파악하고 이를 바탕으로 퍼지 제어를 위한 규칙의 수를 결정한다. 둘째로 퍼지 제어기의 여러 파라미터들은 유전자 알고리즘을 적용하여 최적화한다. GA를 이용한 최적화 과정에서는 성능평가 기준으로 기준입력에 대한 시스템 응답간의 오차와 오버슈트의 크기를 사용하여 응답이 빠르고 안정적인 제어기를 설계하도록 진화방향을 설정한다. 이렇게 만들어진 퍼지 제어기의 성능을 기존의 PID 제어기와 비교 평가한다

• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.8
• /
• pp.861-866
• /
• 2014

This study evaluates the modeling accuracy of the existing three phase shift models on which the time domain oscillator phase noise theories are based. For the evaluation, this study investigates how accurately the three models can model the phase shift impulse train response of a simple planar oscillator. Evaluation result reveals that Kaertner model most accurately reflects the oscillator's phase shift impulse train responses for five different impulse train inputs, whereas PP model exhibited the worst performance in modeling the phase shift impulse train responses.