• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.5
• /
• pp.467-474
• /
• 2019

In this paper, we propose a novel method to improve FPS while maintaining the accuracy of YOLO v2 model in NVIDIA Jetson TX1. In general, in order to reduce the amount of computation, a conversion to an integer operation or reducing the depth of a network have been used. However, the accuracy of recognition can be deteriorated. So, we use methods to reduce computation and memory consumption through adjustment of the filter size and integrated computation of the network The first method is to replace the $3{\times}3$ filter with a $1{\times}1$ filter, which reduces the number of parameters to one-ninth. The second method is to reduce the amount of computation through CBR (Convolution-Add Bias-Relu) among the inference acceleration functions of TensorRT, and the last method is to reduce memory consumption by integrating repeated layers using TensorRT. For the simulation results, although the accuracy is decreased by 1% compared to the existing YOLO v2 model, the FPS has been improved from the existing 3.9 FPS to 11 FPS.

• Journal of the Architectural Institute of Korea Structure & Construction
• /
• v.35 no.2
• /
• pp.45-52
• /
• 2019

BEMS(Building Energy Management Systems) have been applied to office buildings and collect relevant building energy data, e.g. temperatures, mass flow rates and energy consumptions of building mechanical systems and indoor spaces. The aforementioned measured data can be beneficially utilized for developing data-driven machine learning models which can be then used as part of MPC(Model Predictive Control) and/or optimal control strategies. In this study, the authors developed ANN(Artificial Neural Network) models of an AHU (Air Handling Unit) and a chiller for a real-life office building using BEMS data. Based on the ANN models, the authors developed optimal control strategies, e.g. daily operation schedule with regard to optimal start and stop of the AHU and the chiller (500 RT). It was found that due to the optimal start and stop of the AHU and the chiller, 4.5% and 16.4% of operation hours of the AHU and the chiller could be saved, compared to an existing operation.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.15 no.6
• /
• pp.281-287
• /
• 2020

In this paper, we propose a method for estimating a walking direction by which a mobile robots follows a person using TRT (Tensor RT) pose, which is motion recognition based on deep learning. Mobile robots can measure individual movements by recognizing key points on the person's pelvis and determine the direction in which the person tries to move. Using these information and the distance between robot and human, the mobile robot can follow the person stably keeping a safe distance from people. The TRT Pose only extracts key point information to prevent privacy issues while a camera in the mobile robot records video. To validate the proposed technology, experiment is carried out successfully where human walks away or toward the mobile robot in zigzag form and the robot continuously follows human with prescribed distance.

• Journal of Ginseng Research
• /
• v.47 no.5
• /
• pp.662-671
• /
• 2023

Background: 20(S)-protopanaxadiol (PPD), a ginsenoside metabolite, has prominent benefits for the central nervous system, especially in improving learning and memory. However, its transcriptional targets in brain tissue remain unknown. Methods: In this study, we first used mass spectrometry-based drug affinity responsive target stability (DARTS) to identify the potential proteins of ginsenosides and intersected them with the transcription factor library. Second, the transcription factor PURA was confirmed as a target of PPD by biolayer interferometry (BLI) and molecular docking. Next, the effect of PPD on the transcriptional levels of target genes of PURA in brain tissues was determined by qRT-PCR. Finally, bioinformatics analysis was used to analyze the potential biological features of these target proteins. Results: The results showed three overlapping transcription factors between the proteomics of DARTS and transcription factor library. BLI analysis further showed that PPD had a higher direct interaction with PURA than parent ginsenosides. Subsequently, BLI kinetic analysis, molecular docking, and mutations in key amino acids of PURA indicated that PPD specifically bound to PURA. The results of qRT-PCR showed that PPD could increase the transcription levels of PURA target genes in brain. Finally, bioinformatics analysis showed that these target proteins were involved in learning and memory function. Conclusion: The above-mentioned findings indicate that PURA is a transcription target of PPD in brain, and PPD upregulate the transcription levels of target genes related to cognitive dysfunction by binding PURA, which could provide a chemical and biological basis for the study of treating cognitive impairment by targeting PURA.

• Journal of Intelligence and Information Systems
• /
• v.29 no.4
• /
• pp.31-49
• /
• 2023

With the growth of the food-catering industry, consumer preferences and the number of dine-in restaurants are gradually increasing. Thus, personalized recommendation services are required to select a restaurant suitable for consumer preferences. Previous studies have used questionnaires and star-rating approaches, which do not effectively depict consumer preferences. Online reviews are the most essential sources of information in this regard. However, previous studies have aggregated online reviews into long documents, and traditional machine-learning methods have been applied to these to extract semantic representations; however, such approaches fail to consider the surrounding word or context. Therefore, this study proposes a novel review textual-based restaurant recommendation model (RT-RRM) that uses deep learning to effectively extract consumer preferences from online reviews. The proposed model concatenates consumer-restaurant interactions with the extracted high-level semantic representations and predicts consumer preferences accurately and effectively. Experiments on real-world datasets show that the proposed model exhibits excellent recommendation performance compared with several baseline models.

• Journal of Life Science
• /
• v.25 no.9
• /
• pp.1043-1050
• /
• 2015

Glutamate is one of the principle transmitters in the CNS. Ionotropic receptors of glutamate, selectively activated by N-methyl-D-aspartate (NMDA), play an important role in the processes of cell development, learning, memory, and etc. On the other hand, many studies discovered that over-activation of glutamate receptors leads to neurodegeneration and are known to be implicated in major areas of brain pathology. Any sustained effect of a transient NMDA receptor activation is likely to involve signaling to the nucleus and to trigger coordinated changes in gene expression. Classically, a set of immediate-early genes are induced first; some of genes are by themselves transcription factors that control expression of other target genes. This study provides understanding of changes of inducible transcription factors mRNA levels with RT-PCR by inducing over-activation of NMDA receptor with intraperitoneal NMDA injection. The experimental conditions were varied by 1, 5, 25, and 125 g/ of body weight NMDA and measured transcription factors mRNA levels are Egr-1, c-Jun, JunB, and FosB. Based on result obtained, inducible transcription factors mRNA in NMDA injection to mice with 5 g/body weight showed the greatest change. And ITF mRNA showed greatest change 24 hr after injection. The expression level of JunB mRNA was markedly changed. Up to the present days, no study clearly understood how ITF mRNA affected the apoptosis of purkinje cells in the cerebellum. The current study improves the understanding of the mechanism of apoptosis of purkinje cells in the cerebellum.

• Information Systems Review
• /
• v.23 no.4
• /
• pp.45-67
• /
• 2021

As the transaction volume of the C2C second-hand market is growing, the number of frauds, which intend to earn unfair gains by sending products different from specified ones or not sending them to buyers, is also increasing. This study explores the model that can identify frauds in the online C2C second-hand market by examining the postings for transactions. For this goal, this study collected 145,536 field data from actual C2C second-hand market. Then, the model is built with the characteristics from postings such as the topic and the linguistic characteristics of the product description, and the characteristics of products, postings, sellers, and transactions. The constructed model is then trained by the machine learning algorithm XGBoost. The final analysis results show that fraudulent postings have less information, which is also less specific, fewer nouns and images, a higher ratio of the number and white space, and a shorter length than genuine postings do. Also, while the genuine postings are focused on the product information for nouns, delivery information for verbs, and actions for adjectives, the fraudulent postings did not show those characteristics. This study shows that the various features can be extracted from postings written in C2C second-hand transactions and be used to construct an effective model for frauds. The proposed model can be also considered and applied for the other C2C platforms. Overall, the model proposed in this study can be expected to have positive effects on suppressing and preventing fraudulent behavior in online C2C markets.

• The Journal of Korea Robotics Society
• /
• v.1 no.2
• /
• pp.142-150
• /
• 2006

This study examines whether the reinforcement theory would be effectively applied to teaching assistant robots between a robot and a student in the same way as it is applied to teaching methods between a teacher and a student. Participants interact with a teaching assistant robot in a 3 (types of robots: positive reinforcement vs. negative reinforcement vs. both reinforcements) by 2 (types of participants: honor students vs. backward students), within-subject experiment. Three different types of robots, such as 'Ching-chan-ee' which gives 'positive reinforcement', 'Um-bul-ee' which gives 'negative reinforcement', and 'Sang-bul-ee' which gives both 'positive and negative reinforcement' are designed based on the reinforcement theory and the token reinforcement system. Participants' task performance and reaction rate are measured according to the types of robots and the types of participants. In task performance, the negative reinforcement robot is more effective than the other two types, but regarding the number of stimulus, the less the stimulus is, the more effective the task performance is. Also, participants showed the highest reaction rate on the negative reinforcement robot which implies that the negative reinforcement robot is most effective to motivate students. The findings demonstrate that the participants perceive the teaching assistant robot not as a toy but as a teaching assistant and the reinforcement interaction is important and effective for teaching assistant robots to motivate students. The results of this study can be implicated as an effective guideline to interaction design of teaching assistant robots.

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.25 no.1
• /
• pp.37-47
• /
• 2011

This study was aimed to investigate the memory enhancing effect of Jeongjihwan against scopolamine-induced amnesia in C57BL/6 mice. To determine the effect of Jeongjihwan on the memory and cognitive function, we have injected scopolamine (1 mg/kg, i.p.) into C57BL/6 mice 30 min before beginning of behavior tests. We have conducted Y-maze, Morris water-maze, passive avoidance and fear conditioning tests to compare learning and memory functions. Scopolamine-induced behavior changes of memory impairment were significantly restored by oral administration of Jeongjihwan (100 or 200 mg/kg/day). To elucidate the molecular mechanism underlying the memory enhancing effect of Jeongjihwan, we have examined the antioxidant defense system and neurotrophic factors. Jeongjihwan treatment attenuated intracellular accumulation of reactive oxygen species and up-regulated mRNA and protein expression of antioxidant enzymes as assessed by RT-PCR and western blot analysis, respectively. Jeongjihwan also increased protein levels of brain-derived neurotrophic factor (BDNF) compared with those in the scopolamine-treated group. Furthermore, as an upstream regulator, the activation of cAMP response element-binding protein (CREB) via phosphorylation was assessed by Western blot analysis. Jeongjihwan elevated the phosphorylation of CREB (p-CREB), which seemed to be mediated partly by extracellular signal-regulated kinase1/2 (ERK1/2) and protein kinase B/Akt. These findings suggest that Jeongjihwan may have preventive and therapeutic potential in the management of amnesia.

• Korean Journal of Biological Psychiatry
• /
• v.8 no.1
• /
• pp.85-95
• /
• 2001

This study was designed to examine the effects of two antidepressant drugs on the expression of c-fos mRNA in cultured embryonic rat hippocampal neurons. The drugs used were imipramine and amitriptyline. On the fourth day of culture, hippocampal neurons were treated with variable concentrations of each drug. Competitive RT-PCR(Reverse Transcriptase-PCR) analysis was used to quantify the c-fos mRNA expression induced by each drug. Experimental results showed that acute and direct treatment with imipramine and amitriptyline with relatively low concentrations(imipramine ${\leq}10{\mu}M$, amitriptylne ${\leq}10{\mu}M$) had no inductive effect on the expression of c-fos mRNA in the rat hippocampal neurons. However, after treatment with relatively high concentrations(imipramine ${\geq}100{\mu}M$, amitriptyline ${\geq}100{\mu}M$) c-fos mRNA was not detected. These findings suggest the followings. Firstly, the action mechanisms of these drugs on the hippocampal neurons might not be mediated by c-fos but by other immediate-early genes(IEGs). Secondly, their actions may be mediated indirectly via other areas of the brain. Thirdly, the expression of c-fos might be inhibited by high concentrations of these drugs, or the high concentrations could induce cell death. Finally, though cell death remains to be confirmed, the inhibition of c-fos induction or cell death could play a role in the cognitive impairments known to be adverse effects of some antidepressants. This study is believed to be a first step toward understanding the mechanisms of learning and memory. Further studies are needed to investigate the expression of various IEGs and changes in the hippocampal neurons of rat resulting from chronic treatment with antidepressant drugs.