• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.597-599
• /
• 2022

영아는 울음이라는 비언어적 의사 소통 방식을 사용하여 모든 욕구를 표현한다. 하지만 영아의 울음소리를 파악하는 것에는 어려움이 따른다. 영아의 울음소리를 해석하기 위해 많은 연구가 진행되었다. 이에 본 논문에서는 3D 특징 벡터를 이용한 영아의 울음소리 분류를 제안한다. Donate-a-corpus-cry 데이터 세트는 복통, 트림, 불편, 배고픔, 피곤으로 총 5 개의 클래스로 분류된 데이터를 사용한다. 데이터들은 원래 속도의 90%와 110%로 수정하는 방법인 템포조절을 통해 증강한다. Spectrogram, Mel-Spectrogram, MFCC 로 특징 벡터화를 시켜준 후, 각각의 2 차원 특징벡터를 묶어 3차원 특징벡터로 구성한다. 이후 3 차원 특징 벡터를 ResNet 과 EfficientNet 모델로 학습을 진행한다. 그 결과 2 차원 특징 벡터는 0.89(F1) 3 차원 특징 벡터의 경우 0.98(F1)으로 0.09 의 성능 향상을 보여주었다.

• Korean Journal of Childcare and Education
• /
• v.11 no.6
• /
• pp.97-114
• /
• 2015

This study is aimed to determine whether there is any meaningful correlation between predictors and is intended to examine the relative influences that the pedagogical content knowledge of science teaching, scientific attitudes, science teaching intentions have on their science teaching efficacy. The results, showed that the childcare center teachers' pedagogical content knowledge of science teaching, scientific attitudes, science teaching intentions, and science teaching efficacy are all correlated with each other. The teachers' science teaching efficacy are predicted by their scientific attitudes, science teaching intentions, and pedagogical content knowledge of science teaching-learning methods. Specifically, for teachers of 0 to 2-year-olds, their science teaching efficacy is predicted by their science teaching intentions and scientific attitudes. For teachers of 3 to 5-year-olds, their science teaching is affected by their scientific attitudes and their pedagogical content knowledge.

• Korean Journal of Childcare and Education
• /
• v.7 no.3
• /
• pp.43-65
• /
• 2011

The purpose of this study was to investigate the current status of English education for toddlers at home and institutes, and influential variables upon mothers' demand on English education for their toddlers. Four hundred forty-one mothers who have toddlers under 36 months-old were surveyed using a questionnaire. Data were analyzed using frequency and logistic regression. Results of this study were as follows: It was found that 30.9% of mothers were providing English education to their toddlers. They started English education when their toddlers were between 24 months and 36 months because of parents' decision or English education in child care centers. English education was mostly given to their toddlers once or twice a week, less than 30 minutes a time with Korean teachers who majored in English. English lessons were focused on simple dialogue using multi media, worksheets, and workbooks. In other sense, 66.0% of mothers showed a positive perspective toward English education for toddlers. The higher salary, years of education, and degree of agreement on English education based on physiological, linguistic, and social perspectives were, the stronger the demand on English education was. Results of this study suggested that appropriate guidance for English education for toddlers needs to be provided.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.11a
• /
• pp.465-468
• /
• 2005

강화 학습이란 개체가 동적인 환경에서 시행착오를 통해 자신의 최적 행동을 찾아내는 기법이다. 특히 Q-learning과 같은 비(非)모델 기반의 강화학습은 사전에 환경에 대한 모델을 필요로 하지 않으며, 다양한 상태와 행동들을 충분히 경험한다면 최적의 행동 전략에 도달할 수 있으므로 여러 분야에 적용되고 있다. 본 논문에서는 로봇의 행동을 효율적으로 제어하기 위하여 Q-learning을 이용하였다. 로봇 축구 시스템은 공과 여러 대의 로봇이 시시각각 움직이는 시변 환경이므로 모델링이 상당히 복잡하다. 공을 골대 가까이 보내는 것이 로봇 축구의 목표지만 때로는 공을 무조건 골대 방향으로 보내는 것보다 더 효율적인 전략이 있을 수도 있다. 어떤 상황에서 어떤 행동을 하여야 장기적으로 보았을 때 더 우수한지 학습을 통해 로봇 스스로가 판단해가도록 시스템을 구현하고, 학습된 결과를 분석한다.

• Journal of KIISE:Software and Applications
• /
• v.31 no.11
• /
• pp.1431-1438
• /
• 2004

Reinforcement learning learns policies for accomplishing a task's goal by experience through interaction between agent and environment. Q-learning, basis algorithm of reinforcement learning, has the problem of curse of dimensionality and slow learning speed in the incipient stage of learning. In order to solve the problems of Q-learning, new function approximation methods suitable for reinforcement learning should be studied. In this paper, to improve these problems, we suggest Fuzzy Q-Map algorithm that is based on online fuzzy clustering. Fuzzy Q-Map is a function approximation method suitable to reinforcement learning that can do on-line teaming and express uncertainty of environment. We made an experiment on the mountain car problem with fuzzy Q-Map, and its results show that learning speed is accelerated in the incipient stage of learning.

• The Journal of Korean Academic Society of Nursing Education
• /
• v.19 no.3
• /
• pp.371-383
• /
• 2013

Purpose: The purpose of this study was to develop a module for infants in pediatric nursing based on problem-based learning (PBL) and to analyze its effect. Methods: The data used in the study was collected from 79 junior nursing students. For the analysis of the module effect, one group post-test design was applied for measurement of academic achievement. Results: The module for infants is comprised of a 4-hour PBL module. The academic achievement tests are composed of formative (tutor evaluation, self-evaluation) and summative evaluations. The self-evaluation revealed the strong need for PBL and the importance of PBL in the problem solving process and integrative thinking. The mean score of the subjects' learning satisfaction, problem solving process and concept map assessments were all above 50 percent. There was no significant difference in the mean score between the infant PBL and other teaching methods. Conclusion: The PBL proved to be a useful learning method to promote the critical and integrative thinking process.

• Journal of the Korean Society of Child Welfare
• /
• no.54
• /
• pp.79-111
• /
• 2016

The purpose of this study was to examine the structural models in which early cumulative risk factors affect children's language(indicated by expressive vocabularies) and social development(indicated by peer competence) at age 3 thorough their effects on the home learning environment. To examine the hypothesized models, the data of 1,725 families from the second and the fourth waves of the Panel Study of Korean Children was used. Correlation analysis and structural equation modeling were conducted to test the models. First, the cumulative risk factors at age 1 and 3 were highly correlated, implying the stability of the risk factors over time. The more cumulative risk factors at age 1 predicted the lower level of the home learning environment at age 3, which, in turn, was significantly related to both language and social development at age 3. However, the early cumulative risk factors did not directly influence later developmental outcomes. Moreover, the cumulative risk factors at age 3 were directly related to the child's language development, but neither social development northe home learning environment. In addition, the mediational role of the home learning environment (i.e., cumulative risk factors at age 1${\rightarrow}$home learning environment${\rightarrow}$language and social development) was statistically supported. In conclusion, the early cumulative risk factors in infancy indirectly predicted children's development at age 3 through the home learning environment. The practical implications for the early intervention and support for the families with infants who are experiencing multiple risk factors were discussed.

• Korean Journal of Childcare and Education
• /
• v.10 no.6
• /
• pp.223-241
• /
• 2014

The objective of this study was to plan a direction for the connection of the area of inquiry in daily life between infants and young children in Korea based on the analysis of the early learning standards for infants aged 0~2, and young children aged 3~4/5 in 17 different states in America. The problem of the study is as follows: What is the content area of cognitive development, science and mathematics of early learning standards regarding age? The data used in this study was collected from a total of 17 states in America in which 12 states classify infants and young children as aged 0~2 and 3~4/5 respectively and 5 states do not classify the age groups. The results obtained from the analysis of the given issues are as follows. First, in the case of the five states that showed the group aged 0~4/5, the contents related to the area of inquiry in daily life of the early learning standards were included in the mathematics and science and a part of these contents were also included in cognitive developments. In the case of the 12 states that classified the groups aged 0~2 and 3~4/5 respectively, the group aged 0~2 included cognitive developments and the group aged 3~4/5 includes mathematics and science. Second, the contents related to the area of inquiry in daily life of the early learning standards in the five states that showed the single group aged 0~4/5 include the contents in order of scientific idea, biology, physics, and the earth and space. In the 12 states that classify different age groups. Third, in the case of the states that separate age groups, the group aged 0~2 include the contents of number and operation, geometry and space while the group aged 3~4/5 include the contents of number and operations, geometry and space, and measurement. The implications of this study was that it is necessary to take into account the linking between development and characteristics of the subjects.

• Korean Journal of Child Studies
• /
• v.27 no.6
• /
• pp.199-215
• /
• 2006

In this study of the documentation process, a toddler's behavior was interpreted through use of text and picture both of which are subjective representations of reality. Two teachers deconstructed and reconstructed their knowledge of a photographed toddler by first referring to scientific theories from a text. They used these theories to analyze markings drawn on paper by the toddler in the photograph. But, the theories from the text were transformed into a new theory in the documentation process of reflecting on the history of markings in different contexts. Thus, the semiotics of pictures and texts changed in addition to change in meanings. This demonstrates the variability of signs, as in the concept of "rhizome" (Deleuze and Guattari, 1999).

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.2
• /
• pp.76-90
• /
• 2024

Although artificial intelligence (AI) can be utilized in various domains such as smart city, healthcare, it is limited due to concerns about the exposure of personal and sensitive information. In response, the concept of distributed machine learning has emerged, wherein learning occurs locally before training a global model, mitigating the concentration of data on a central server. However, overall learning phase in a collaborative way among multiple participants poses threats to data privacy. In this paper, we systematically analyzes recent trends in privacy protection within the realm of distributed machine learning, considering factors such as the presence of a central server, distribution environment of the training datasets, and performance variations among participants. In particular, we focus on key distributed machine learning techniques, including horizontal federated learning, vertical federated learning, and swarm learning. We examine privacy protection mechanisms within these techniques and explores potential directions for future research.