• Journal of Korea Multimedia Society
• /
• v.25 no.3
• /
• pp.493-501
• /
• 2022

This paper is a study on context-sensitive spelling error correction and uses the Korean WordNet (KorLex)[1] that defines the relationship between words as a graph to improve the performance of the correction[2] based on the vector information of the word embedded in the correction technique. The Korean WordNet replaced WordNet[3] developed at Princeton University in the United States and was additionally constructed for Korean. In order to learn a semantic network in graph form or to use it for learned vector information, it is necessary to transform it into a vector form by embedding learning. For transformation, we list the nodes (limited number) in a line format like a sentence in a graph in the form of a network before the training input. One of the learning techniques that use this strategy is Deepwalk[4]. DeepWalk is used to learn graphs between words in the Korean WordNet. The graph embedding information is used in concatenation with the word vector information of the learned language model for correction, and the final correction word is determined by the cosine distance value between the vectors. In this paper, In order to test whether the information of graph embedding affects the improvement of the performance of context- sensitive spelling error correction, a confused word pair was constructed and tested from the perspective of Word Sense Disambiguation(WSD). In the experimental results, the average correction performance of all confused word pairs was improved by 2.24% compared to the baseline correction performance.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.5
• /
• pp.518-524
• /
• 2022

In this paper, we propose a method for combining convolutional neural networks and attention mechanism to improve the recognition performance of emotional stress from Electroencephalogram (EGG) signals. In the proposed method, EEG signals are decomposed into five frequency domains, and spatial information of EEG features is obtained by applying a convolutional neural network layer to each frequency domain. As a next step, salient frequency information is learned in each frequency band using a gate transformer-based attention mechanism, and complementary frequency information is further learned through inter-frequency mapping to reflect it in the final attention representation. Through an EEG stress recognition experiment involving a DEAP dataset and six subjects, we show that the proposed method is effective in improving EEG-based stress recognition performance compared to the existing methods.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.1
• /
• pp.27-38
• /
• 2023

In this paper, we propose a conversational AI agent based on continual learning that can continuously learn and grow with new data over time. A continual learning-based conversational AI agent consists of three main components: Task manager, User attribute extraction, and Auto-growing knowledge graph. When a task manager finds new data during a conversation with a user, it creates a new task with previously learned knowledge. The user attribute extraction model extracts the user's characteristics from the new task, and the auto-growing knowledge graph continuously learns the new external knowledge. Unlike the existing conversational AI agents that learned based on a limited dataset, our proposed method enables conversations based on continuous user attribute learning and knowledge learning. A conversational AI agent with continual learning technology can respond personally as conversations with users accumulate. And it can respond to new knowledge continuously. This paper validate the possibility of our proposed method through experiments on performance changes in dialogue generation models over time.

• Journal of the Korean School Mathematics Society
• /
• v.25 no.4
• /
• pp.307-332
• /
• 2022

This action research was carried out to help students learn fractions computation by making and using semi-concrete fraction manipulatives that can be used continuously in math classes. For this purpose, the researcher and students made semi-concrete fraction manipulatives and learned how to use these through reviewing the previously learned fraction contents over 4 class sessions. Afterward, through the 14 classes (7 classes for learning to reduce fractions and to a common denominator, 7 classes for adding and subtracting fractions with different denominators) in which the principle inquiry learning model was applied, students actively engaged in learning activities with fraction manipulatives and explored the principles underneath the manipulations of fraction manipulatives. Students could represent various fractions using fraction manipulatives and solve fraction computation problems using them. The achievement evaluation after class found that the students could connect the semi-concrete fraction manipulatives with fraction representation and symbolic formulas. Moreover, the students showed interest and confidence in mathematics through the classes using fraction manipulatives.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.18 no.1
• /
• pp.125-138
• /
• 2022

Companies want organizational members who take e-learning courses to enjoy the advantages of transcending time and space that e-learning has, but also want what they have learned to help the organization, the work they perform, or their future careers. In addition, while enjoying the effect of reducing education costs compared to offline education through e-learning, it is expected that executives and employees will apply the knowledge and skills learned to the field and perform tasks to achieve results. As COVID-19 continues, many education programs that have been conducted offline at corporate sites have been converted to e-learning, with a larger number of e-learning operations than in the past. This study was conducted based on the perception that learners' learning satisfaction is important for the successful operation of e-learning education, and that learners' own self-directed learning ability and self-determination are important as well as corporate efforts. As a result of the study, hypotheses 1-1, 1-2, 1-3-1, and 1-3-2 that the better the self-determination (autonomy, competence, full-time support, and peer support) is, the higher the learning satisfaction will be. Both Hypothesis 2-1 and Hypothesis 2-2 were adopted that the better self-directed learning (subjectivity, execution ability) is, the higher the learning satisfaction will increase. In conclusion, it is necessary to properly introduce the concepts of self-determination and self-directed learning in corporate education while operating with the corporate education system.

• Journal of the Korean Society of Food Culture
• /
• v.23 no.6
• /
• pp.693-704
• /
• 2008

In this study, the preference for Korean Kimchi by Chinese people in Shandong Province was evaluated. Specifically, this study was conducted to aid in the introduction of Kimchi to China by providing information and developing local types designed to meet regional taste preferences. The subjects were comprised of 298 Chinese (male 108, female 190) residents of Weihai, Yantai and Qingdao, in Shandong province, China. The subjects were provided with a self administered questionnaire form designed to evaluate their views on Korean Kimchi. The collected data were then analyzed using the SAS software package. The results revealed that 95.3% of the respondents were aware of Korean Kimchi. In addition, 100% of the respondents who had visited Korea and 98.1% of the respondents who had an interest in Korea were aware of Kimchi. With regard to the origins of their interest in Kimchi, 26.8% of the subjects answered 'through mass media', while 23.9% reported that they learned about Kimchi 'through friends'. Most subjects recognized Kimchi as a 'Korean traditional food' (92.6%), a 'delicious food' (53.2%), and a 'fermented food' (38.0%). Baechu Kimchi was found to be the most well-known Kimchi, followed by Kkakdugi, Oi Kimchi, Yoelmu Kimchi and Nabak Kimchi. Additionally, 69.1% of the subjects knew how it was prepared, most of whom reported that they learned how Kimchi was prepared through 'Korean movie and/or drama'. Moreover, 88.9% of the subjects had eaten Kimchi. Overall, 43.8% of the subjects reported that they ate Kimchi $1{\sim}2$ times per month, while 32.1% reported that they ate Kimchi $1{\sim}2$ time per year. The most common places that Kimchi was eaten were a 'Korean restaurant' (67.6%) or with a 'colleague' (32.8%). The primary reasons for not having eaten Kimchi were 'no knowledge or dislike of Kimchi by family' (30.3%), 'difficulty purchasing Kimchi' (21.2%), 'high priced Kimchi' (21.2%), and 'dislike the smell and shape of Kimchi' (12.1%).

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.29 no.6
• /
• pp.659-665
• /
• 2023

The exciter of a ship generator adjusts the magnetic flux through excitation current control to maintain the output terminal voltage constant. The voltage controller inside the exciter typically uses a proportional integral control method. however, the response characteristics determined by the gain and time constant produce unwanted output owing to an inappropriate setting value that can reduce the quality and stability of power within the ship. In this study, a neural network circuit is learned using stable input/output data that can be obtained through the AC4A type exciter model provided by IEEE, and the simulation is performed by replacing the existing proportional integral control type voltage controller with the learned neural network circuit controller. Consequently, overshooting was improved by up to 9.63% compared with that of the previous model, and excellence in stable response characteristics was confirmed.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.4
• /
• pp.1-10
• /
• 2024

In the fields of computer animation and robotics, reaching a destination while avoiding obstacles has always been a difficult task. Moreover, generating appropriate motions while planning a route is even more challenging. Recently, academic circles are actively conducting research to generate character motions by modifying and utilizing VAE (Variational Auto-Encoder), a data-based generation model. Based on this, in this study, the latent space of the MVAE model is learned using a reinforcement learning method[1]. With the policy learned in this way, the character can arrive its destination while avoiding both static and dynamic obstacles with natural motions. The character can easily avoid obstacles moving in random directions, and it is experimentally shown that the performance is improved, and the learning time is greatly reduced compared to existing approach.

• Journal of radiological science and technology
• /
• v.30 no.4
• /
• pp.419-426
• /
• 2007

In this paper, the input pattern vectors are extracted and the learning algorithms is designed to recognize taste(bitter, sweet, sour and salty) pattern vectors. The signal intensity of taste are used to compose the input pattern vectors. The SOM(Self Organizing Maps) algorithm for taste pattern recognition is used to learn initial reference vectors and the ot-star learning algorithm is used to determine the class of the output neurons of the sunclass layer. The weights of the proposed algorithm which is between the input layer and the subclass layer can be learned to determine initial reference vectors by using SOM algorithm and to learn reference vectors by using LVQ(Learning Vector Quantization) algorithm. The pattern vectors are classified into subclasses by neurons in the subclass layer, and the weights between subclass layer and output layer are learned to classify the classified subclass, which is enclosed a class. To classify the pattern vectors, the proposed algorithm is simulated with ones of the conventional LVQ, and it is confirmed that the proposed learning method is more successful classification than the conventional LVQ.

• Journal of Educational Research in Mathematics
• /
• v.20 no.2
• /
• pp.185-202
• /
• 2010

This study was designed to investigate the possibility that the optimization problem solving activities based on the instrumented CAS can have an influence on the sequence of mathematics curriculum in secondary mathematics education. Some optimization problem solving activities based on CAS were constructed and executed to eleventh grade(the penultimate year of Korean high school) 7 students for nine class hours. They have experienced using CAS in mathematics class for three months, but never learned calculus. The data which consists of classroom observations(audio and video taped) and post-unit interviews with students were analyzed. In the analysis, with CAS, students can highly deal with the applied optimization problems made up of calculus, cubic equation, solution of radical equation, and graph analysis which never learned. This result shows CAS may have an influence on the sequence of mathematics curriculum in secondary mathematics education.