• Korean Journal of Cognitive Science
• /
• v.14 no.3
• /
• pp.13-21
• /
• 2003

This study presents a design of Teachable Agent(TA) and its theoretical background. TA is an intelligent agent to which students as tutors teach, pose questions, and provide feedbacks using a concept map. TA consists of four independent Modules, Teach Module, Q&A Module, Test Module, and Resource Module. In Teach Module, students teach TA by constructing concept map. In Q&A Module, both students and TA ask questions and answer questions each other through an interactive window. To assess TA's knowledge and provide feedback to students, Test Module consists of a set of predetermined questions which TA should pass. From Resource Module, students can search and look up important information to teach, ask questions, and provide feedbacks whenever they want. It is expected that TA should provide student tutors with an active role in learning and positive attitude toward the subject matter by enhancing their cognition as well as motivation.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.8
• /
• pp.393-400
• /
• 2016

Recently, there have been various researches on efficient and automatic analysis on urban environment methods that utilize the computer vision and machine learning technology. Among many new analyses, urban safety analysis has received a major attention. In order to predict more accurately on safety score and reflect the human visual perception, it is necessary to consider the generic and local information that are most important to human perception. In this paper, we use Double-column Convolutional Neural network consisting of generic and local columns for the prediction of urban safety. The input of generic and local column used re-sized and random cropped images from original images, respectively. In addition, a new learning method is proposed to solve the problem of over-fitting in a particular column in the learning process. For the performance comparison of our Double-column Convolutional Neural Network, we compare two Support Vector Regression and three Convolutional Neural Network models using Root Mean Square Error and correlation analysis. Our experimental results demonstrate that our Double-column Convolutional Neural Network model show the best performance with Root Mean Square Error of 0.7432 and Pearson/Spearman correlation coefficient of 0.853/0.840.

• Korean journal of communication and information
• /
• v.65
• /
• pp.76-92
• /
• 2014

The purpose of this paper is to disclose the characteristics of SNS service platform in the perspective of 'dispositif'. To achieve this I extended Foucault's concept of 'dispositif' to the Internet and service platform of Google, Facebook. The platform dispositif is a complex of the interface services for the users, collection of data produced by users and the control of them. These three aspects are combined in the service platform modifying their modes in terms of time and space. This study would provide a critical approach toward the appropriation and control of user activities based upon platform.

• Korean journal of communication and information
• /
• v.77
• /
• pp.164-199
• /
• 2016

This study analyzes the socio-economical contexts of Japan's Karaoke inflow in Busan of the 1980s, and examines the relationship between Karaoke culture of the 1980s and 'Noraebang' culture of the early 1990s in Busan from the perspective of the SCOT(social construction of technology) theory. By the end of 1970s, Japan's Karaoke was introduced under the contexts of structural transformations of a geisha tourism in the East Asian regions. Karaoke culture in Busan of the 1970s and 1980s has formulated social recognitions of the novelty of Noraebang culture in the 1990s, although it has done so through the ways of misunderstandings and Nationalism.

• Proceedings of KOSOMES biannual meeting
• /
• 2005.05a
• /
• pp.7-12
• /
• 2005

A ship can be considered as a large human-machine system and the interaction between worker and system affects the work performed and its efficiency. Inside the bridge of a ship, there exist many auditory signals as well as visual signals. However, only a few studies have been performed related to human recognition to alarm systems in bridge. In this study, auditory icons and abstract sounds are compared to find more effective means of alarm systems. the study result shows tint auditory icons are recognized faster than abstract sounds. This result is expected to be used as a basic data for developing performance criteria of auditory display inside bridge and for designing integrated ship bridge alarm system.

• Journal of Korea Society of Industrial Information Systems
• /
• v.26 no.4
• /
• pp.1-9
• /
• 2021

This study was to investigate the types of emergencies transported by the Southwestern Coast Guard, the need for telemedicine guidance, and the perception and attitude of smart glasses as a communication method targeting 31 coast guards. A relatively high frequency and training requirement were confirmed for bleeding, abrasion, and abdominal pain. The demand for telemedicine guidance on medication and triage was higher, and the perceived usefulness and attitude scores for the use of smart glasses were 3.76±0.61 and 3.64±0.45, respectively. A moderate correlation between perceived usefulness and attitude toward smart glasses was confirmed (r=.630, p<.01). With the development of technology, it is time to actively introduce new devices such as smart glasses.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.4
• /
• pp.825-840
• /
• 2015

Adaptive bitrate (ABR) streaming technology has become an important and prevalent feature in many multimedia delivery systems, with content providers such as Netflix and Amazon using ABR streaming to increase bandwidth efficiency and provide the maximum user experience when channel conditions are not ideal. Where such systems could see improvement is in the delivery of live video with a closed loop cognitive control of video encoding. In this paper, we present streaming camera system which provides spatially and temporally adaptive video streams, learning the user's preferences in order to make intelligent scaling decisions. The system employs a hardware based H.264/AVC encoder for video compression. The encoding parameters can be configured by the user or by the cognitive system on behalf of the user when the bandwidth changes. A cognitive video client developed in this study learns the user's preferences(i.e. video size over frame rate) over time and intelligently adapts encoding parameters when the channel conditions change. It has been demonstrated that the cognitive decision system developed has the ability to control video bandwidth by altering the spatial and temporal resolution, as well as the ability to make scaling decisions.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.12
• /
• pp.581-588
• /
• 2017

This study was conducted to determine which factors would predict metabolic syndrome (MetS) perception and exercise by applying a machine learning classifier, or Extreme Gradient Boosting algorithm (XGBoost) from July 2014 to December 2015. Data were obtained from the Korean Community Health Survey (KCHS), representing different community-dwelling Korean adults 19 years and older, from 2009 to 2013. The dataset includes 370,430 adults. Outcomes were categorized as follows based on the perception of MetS and physical activity (PA): Stage 1 (no perception, no PA), Stage 2 (perception, no PA), and Stage 3 (perception, PA). Features common to all questionnaires for the last 5 years were selected for modeling. Overall, there were 161 features, categorical except for age and the visual analogue scale (EQ-VAS). We used the Extreme Boosting algorithm in R programming for a model to predict factors and achieved prediction accuracy in 0.735 submissions. The top 10 predictive factors in Stage 3 were: age, education level, attempt to control weight, EQ mobility, nutrition label checks, private health insurance, EQ-5D usual activities, anti-smoking advertising, EQ-VAS, education in health centers for diabetes, and dental care. In conclusion, the results showed that XGBoost can be used to identify factors influencing disease prevention and management using healthcare bigdata.

• Korean Journal of Cognitive Science
• /
• v.29 no.3
• /
• pp.161-192
• /
• 2018

Retrieving temporal information of encoded events is one of the core control processes in episodic memory. Despite much prior neuroimaging research on episodic retrieval, little is known about how large-scale connectivity patterns are involved in the retrieval of sequentially organized episodes. Task-related functional connectivity multivariate pattern analysis was used to distinguish the different sequential retrieval. In this study, participants performed temporal episodic memory tasks in which they were required to retrieve the encoded items in either the forward or backward direction. While separately parsed local networks did not yield substantial efficiency in classification performance, the large-scale patterns of interactivity across the cortical and sub-cortical brain regions implicated in both the cognitive control of memory and goal-directed cognitive processes encompassing lateral and medial prefrontal regions, inferior parietal lobules, middle temporal gyrus, and caudate yielded high discriminative power in classification of temporal retrieval processes. These findings demonstrate that mnemonic control processes across cortical and subcortical regions are recruited to re-experience temporally-linked series of memoranda in episodic memory and are mirrored in the qualitatively distinct global network patterns of functional connectivity.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.12
• /
• pp.491-498
• /
• 2019

Digital twin is a technology that virtualizes physical objects of the real world on a computer. It is used by collecting sensor data through IoT, and using the collected data to connect physical objects and virtual objects in both directions. It has an advantage of minimizing risk by tuning an operation of virtual model through simulation and responding to varying environment by exploiting experiments in advance. Recently, artificial intelligence and machine learning technologies have been attracting attention, so that tendency to virtualize a behavior of physical objects, observe virtual models, and apply various scenarios is increasing. In particular, recognition of each robot's motion is needed to build digital twin for co-robot which is a heart of industry 4.0 factory automation. Compared with modeling based research for recognizing motion of co-robot, there are few attempts to predict motion based on sensor data. Therefore, in this paper, an experimental environment for collecting current and inertia data in co-robot to detect the motion of the robot is built, and a motion prediction model based on the collected sensor data is proposed. The proposed method classifies the co-robot's motion commands into 9 types based on joint position and uses current and inertial sensor values to predict them by accumulated learning. The data used for accumulating learning is the sensor values that are collected when the co-robot operates with margin in input parameters of the motion commands. Through this, the model is constructed to predict not only the nine movements along the same path but also the movements along the similar path. As a result of learning using SVM, the accuracy, precision, and recall factors of the model were evaluated as 97% on average.