• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.12
• /
• pp.81-90
• /
• 2016

The objective of this study was to investigate the effect of an education method applying the flipped learning technique for college students. Both self-directed learning readiness and educational performance before and after applying the flipped learning were examined. After applying the flipped learning technique, teacher-student interaction, learning satisfaction, and learning motivation were identified. The correlation of each variable was examined after applying the flipped learning technique to investigate its influence on learning motivation. A total of 68 second-year nursing students enrolled in E University were analyzed. A difference between before and after applying the flipped learning was analyzed by the paired t-test; a correlation between the variables was analyzed via Pearson's correlation coefficient; and an influence on the dependent variable learning motivation was analyzed using the stepwise multiple regression analysis. The results showed that self-directed learning readiness increased before and after applying the flipped learning technique with statistical significance, and the difference of educational performance was not significant. After an education session applying the flipped learning technique, a learning motivation demonstrated a significantly positive correlation with self-directed learning readiness (r=0.33, p=.006), college student educational performance (r=0.51, p<.001), teacher-student interaction (r=0.72, p<.001), and learning satisfaction (r=0.79, p<.001). A significantly positive correlation was also observed between the other variables. Factors influencing learning motivation were learning satisfaction and teacher-student interaction. The explanatory power for learning motivation in the regression model considering these two variables was 71.3% (F=80.66, p<.001). Therefore, to enhance learning motivation in applying the flipped learning technique, it is necessary to increase learning satisfaction and to establish a strategy that further vitalizes the teacher-student interaction.

• The Journal of Korean Medicine
• /
• v.30 no.2
• /
• pp.1-16
• /
• 2009

Objectives: This study investigated the effect of Yukmijihwangtang on cerebral ischemia-induced learning and memory impairment by middle cerebral artery (MCA) occlusion in rats. Methods: The ability of learning and memory of rats was measured using the eight-arm radial maze and the passive avoidance test, and profile of cholinergic neuron was assessed in the medial septum and hippocampus region by immuno-histochemistry. Results: 1. No differences were found between groups in the number of correct choices in acquisition performance during the eight-arm radial maze task. 2. No differences were found between groups on day 1 in the error rate in acquisition performance, which is defined as the number of enters into the same arm more than once within five minutes. After 5 to 6 days of test, the number of errors was significantly reduced in the Yukmijihwangtang group (forebrain ischemia group with Yukmijihwangtang treatment), compared with the ischemia group. 3. The memory processes significantly improved in the Yukmijihwangtang group according to results of the passive avoidance test. 4. The appearance of AchE (acetylcholinesterase) in the CA1 region of hippocampus significantly decreased in the ischemia group, compared with the sham group (untreated group). The appearance of AchE in the same region significantly increased in the Yukmijihwangtang group, compared with the ischemia group. 5. The appearance of ChAT (choline acetyltransferase) in the CA1 region of the hippocampus and medial septum decreased in the ischemia group, compared with the sham group. The appearance of ChAT in the same region significantly increased in the Yukmijihwangtang group, compared with the ischemia group Conclusions: This study provides evidence that Yukmijihwangtang is effective for reviving the ability of learning and memory and damaged neurons in rats with experimental cerebral ischemia.

• Journal of Korea Multimedia Society
• /
• v.11 no.3
• /
• pp.353-363
• /
• 2008

A multi-agent system has a longer learning period and larger state-spaces than a sin91e agent system. In this paper, we suggest a new method to reduce the learning time of Nash Q-learning in a multi-agent environment. We apply Macro-actions to Nash Q-learning to improve the teaming speed. In the Nash Q-teaming scheme, when agents select actions, rewards are accumulated like Macro-actions. In the experiments, we compare Nash Q-learning using Macro-actions with general Nash Q-learning. First, we observed how many times the agents achieve their goals. The results of this experiment show that agents using Nash Q-learning and 4 Macro-actions have 9.46% better performance than Nash Q-learning using only 4 primitive actions. Second, when agents use Macro-actions, Q-values are accumulated 2.6 times more. Finally, agents using Macro-actions select less actions about 44%. As a result, agents select fewer actions and Macro-actions improve the Q-value's update. It the agents' learning speeds improve.

• Journal of Distribution Science
• /
• v.13 no.6
• /
• pp.105-118
• /
• 2015

Purpose - Currently, strengthening small and medium-sized enterprises (SME) in terms of competitiveness is a key economic issue. However, the problem is that many SMEs lack the internal competence required to cope with a rapidly changing market structure. Such problems can act as an obstacle to economic development, yet most SMEs in Korea are dealing with this problem today. A company's source of competitive advantage is changing from quantity to quality, facility to knowledge, and hardwork to creativity. Under such circumstances, a company should place learning and sharing of knowledge and continuously creating new knowledge as its priority. This study aims to identify the effect of a chief executive officer's (CEO) compassion and forgiveness - positive factors in organizational emotion - on learning organization activities and organizational performance, through a theoretical comparison. Research design, data, and methodology - For this study, SMEs based in Daejeon and Chungcheong area were selected. To secure credibility of the data, the subjects were selected among those who have been working at the business for six months or longer. The survey was conducted for 30 days from March 5, 2015 to April 5, 2015. Both offline and online surveys were conducted. Fifty companies were chosen and 700 questionnaires were distributed, with 506 used for analysis. Fifty subject companies (25 from Daejeon, 10 from Chungnam, 10 from Chungbuk, and five from Sejong) were selected and the objective, target, and survey content were explained to a manager at each company either face-to-face or on the phone. Of the total of 700 questionnaires distributed via mail or e-mail, 78.6% or 550 copies were returned. Excluding 44 insufficient questionnaires, the remainder, 506 questionnaires, were used for analysis. Results - This study analyzed how the CEO's compassion and forgiveness affects learning organization activities and organizational performance. First, compassion of the CEO at the SMEs directly affected the learning organization activities and indirectly affected the organizational performance. Second, forgiveness of the CEO at the SMEs did not affect the learning organization activities and organizational performance directly or indirectly. Conclusions - The study conclusions are as follows. First, CEO compassionate behavior at the SMEs was a significant variable that directly and indirectly affected learning organization activities and organizational performance. Therefore, the CEO of an SME can create a positive organizational atmosphere through compassionate behaviors in the organization. Second, the forgiving behavior of the CEO did not have direct or indirect effects on learning organization activities and organizational performance. However, the reason for a CEO to continue his or her forgiving behavior is because it strengthens employee resilience, commitment, and self-efficacy to protect the organization from negative influences such as layoffs, risks, and wrongdoings. The action of forgiveness does not have direct or indirect effects. However, the CEO shall continue such behavior to strengthen members' physiological resilience, commitment, and self - effectiveness, and to protect the organization from risks including layoff and external negative factors.

• The Journal of Society for e-Business Studies
• /
• v.25 no.1
• /
• pp.1-12
• /
• 2020

Recently, research using machine learning have shown remarkable results in various domains, leading to the fact that leaning-based dispatchers have intrigued interest in both academia and industry. To improve the performance of the dispatcher, each dispatch decision needs to be evaluated in detail. However, existing studies on visualization techniques for manufacturing lines have mainly focused on illustrating the performance indicators or abnormal patterns. In this paper, we propose a monitoring system that displays a variety of information about the manufacturing line along with alternatives at the time of each dispatching decision being made. Furthermore, the proposed system effectively represents the cause of the idle time of resources and the change of the performance index over time.

• The Korean Journal of Physiology and Pharmacology
• /
• v.14 no.5
• /
• pp.345-352
• /
• 2010

The present study was undertaken to explore the potential of erythropoietin in memory deficits of mice. Memory impairment was produced by scopolamine (0.5 mg/kg, $i.p.$) and intracerebroventricular streptozotocin (i.c.v STZ, 3 mg/kg, $10{\mu}l$, $1^{st}$ and $3^{rd}$ day) in separate groups of animals. Morris water-maze test was employed to assess learning and memory. The levels of brain thio-barbituric acid reactive species (TBARS) and reduced glutathione (GSH) were estimated to assess degree of oxidative stress. Brain acetylcholinesterase enzyme (AChE) activity was also measured. Scopolamine/streptozotocin administration induced significant impairment of learning and memory in mice as indicated by marked decrease in Morris water-maze performance. Scopolamine/streptozotocin administration also produced a significant enhancement of brain AChE activity and brain oxidative stress (an increase in TBARS and a decrease in GSH) levels. Treatment of erythropoietin (500 and 1,000 IU/Kg i.p.) significantly reversed scopolamine- as well as streptozotocin-induced learning and memory deficits along with attenuation of those-induced rise in brain AChE activity and brain oxidative stress levels. It may be concluded that erythropoietin exerts a beneficial effect in memory deficits of mice possibly through its multiple actions including potential anti-oxidative effect.

• Computers and Concrete
• /
• v.16 no.5
• /
• pp.741-757
• /
• 2015

In the last decade, several modeling approaches have been proposed and applied to estimate the high-performance concrete (HPC) slump flow. While HPC is a highly complex material, modeling its behavior is a very difficult issue. Thus, the selection and application of proper modeling methods remain therefore a crucial task. Like many other applications, HPC slump flow prediction suffers from noise which negatively affects the prediction accuracy and increases the variance. In the recent years, ensemble learning methods have introduced to optimize the prediction accuracy and reduce the prediction error. This study investigates the potential usage of bagging (Bag), which is among the most popular ensemble learning methods, in building ensemble models. Four well-known artificial intelligence models (i.e., classification and regression trees CART, support vector machines SVM, multilayer perceptron MLP and radial basis function neural networks RBF) are deployed as base learner. As a result of this study, bagging ensemble models (i.e., Bag-SVM, Bag-RT, Bag-MLP and Bag-RBF) are found superior to their base learners (i.e., SVM, CART, MLP and RBF) and bagging could noticeable optimize prediction accuracy and reduce the prediction error of proposed predictive models.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.20 no.42
• /
• pp.161-169
• /
• 1997

A variety of learning methods are used for neural networks. Among them, the backpropagation algorithm is most widely used in such image processing, speech recognition, and pattern recognition. Despite its popularity for these application, its main problem is associated with the running time, namely, too much time is spent for the learning. This paper suggests a method which maximize the convergence speed of the learning. Such reduction in e learning time of the backpropagation algorithm is possible through an adaptive adjusting of the slope of the activation function depending on total errors, which is named as the variable slope algorithm. Moreover experimental results using this variable slope algorithm is compared against conventional backpropagation algorithm and other variations; which shows an improvement in the performance over pervious algorithms.

• Journal of Multimedia Information System
• /
• v.3 no.4
• /
• pp.155-160
• /
• 2016

The Pulse Coupled Neural Network (PCNN) is a kind of neural network models that consists of spiking neurons and local connections. The PCNN was originally proposed as a model that can reproduce the synchronous phenomena of the neurons in the cat visual cortex. Recently, the PCNN has been applied to the various image processing applications, e.g., image segmentation, edge detection, pattern recognition, and so on. The method for the image matching using the PCNN had been proposed as one of the valuable applications of the PCNN. In this method, the Genetic Algorithm is applied to the PCNN parameter learning for the image matching. In this study, we propose the method of the similar image rating using the PCNN. In our method, the Genetic Algorithm based method is applied to the parameter learning of the PCNN. We show the performance of our method by simulations. From the simulation results, we evaluate the efficiency and the general versatility of our parameter learning method.