• 농업과학연구
• /
• 제47권4호
• /
• pp.1029-1037
• /
• 2020

The machine learning algorithm has been widely used in water-related fields such as water resources, water management, hydrology, atmospheric science, water quality, water level prediction, weather forecasting, water discharge prediction, water quality forecasting, etc. However, water quality prediction studies based on the machine learning algorithm are limited compared to other water-related applications because of the limited water quality data. Most of the previous water quality prediction studies have predicted monthly water quality, which is useful information but not enough from a practical aspect. In this study, we predicted the dissolved oxygen (DO) using recurrent neural network with long short-term memory model recurrent neural network long-short term memory (RNN-LSTM) algorithms with hourly- and daily-datasets. Bugok Bridge in Oncheoncheon, located in Busan, where the data was collected in real time, was selected as the target for the DO prediction. The 10-month (temperature, wind speed, and relative humidity) data were used as time prediction inputs, and the 5-year (temperature, wind speed, relative humidity, and rainfall) data were used as the daily forecast inputs. Missing data were filled by linear interpolation. The prediction model was coded based on TensorFlow, an open-source library developed by Google. The performance of the RNN-LSTM algorithm for the hourly- or daily-based water quality prediction was tested and analyzed. Research results showed that the hourly data for the water quality is useful for machine learning, and the RNN-LSTM algorithm has potential to be used for hourly- or daily-based water quality forecasting.

• 인지과학
• /
• 제19권4호
• /
• pp.369-396
• /
• 2008

멀티미디어 학습에서 시공간 작업기억의 용량에 따라 그림 유형과 그림 내용의 효과가 달라지는지를 두 개의 실험에서 알아보았다. 실험 1에서는 영어단어 설명을 단어로만 주는 조건, 동영상과 같이 주는 조건, 그림과 같이 주는 조건을 비교하였다. 단어검사에서 동영상은 시공간 작업기억 용량이 큰 집단에서만 효과가 있었고, 이해검사에서는 단어설명만 주어진 조건에서 가장 수행이 좋았다. 실험 2에서는 동영상의 내용이 미치는 영향을 알아보았다. 단어검사에서는 전형적인 내용을 담은 동영상이 효과가 있었으나, 이해검사에서는 맥락에 적합한 내용의 동영상으로 학습한 집단이 이해 점수가 높았고, 이 경향은 시공간 작업기억 용량이 작은 집단에서 더 두드러지게 나타났다. 두 실험의 결과는 멀티미디어 학습의 효과는 인지부하의 정도에 달려 있다는 인지부하이론과 멀티미디어 학습을 여러 수준의 활동의 결과로 설명한 생성이론의 의해 잘 설명되는 것으로 해석되었다.

• Animal cells and systems
• /
• 제15권2호
• /
• pp.115-121
• /
• 2011

Previous studies have demonstrated that N-methyl-D-aspartate (NMDA) receptors and acetylcholine receptors are related to learning and memory in rat and mice. In this study, we examined the effects of MK-801, a non-competitive NMDA receptor antagonist, on learning and memory in zebrafish using a passive avoidance test. We further tested whether or not nicotine, a nicotinic acetylcholine receptor agonist, and physostigmine, an acetylcholinesterase inhibitor, reverse the effects of MK-801. Crossing time was increased significantly in the training and test sessions for the controls. When 20 ${\mu}M$ MK-801 was administered prior to the training session, the crossing time did not increase in either session. The MK-801-induced learning deficit was rescued by pretreatment with 20 ${\mu}M$ physostigmine, and crossing time was increased in the training and test sessions compared to the MK-801-treated zebrafish. Further, the MK-801-induced learning deficit was prevented by pretreatment with 20 ${\mu}M$ nicotine, and crossing time was increased in the training session but not in the test session. These results show that MK-801 induced a learning deficit in zebrafish that was prevented by pretreatment with nicotine and physostigmine.

• International Journal of Computer Science & Network Security
• /
• 제22권4호
• /
• pp.119-130
• /
• 2022

Crime is a common social problem that affects the quality of life. As the number of crimes increases, it is necessary to build a model to predict the number of crimes that may occur in a given period, identify the characteristics of a person who may commit a particular crime, and identify places where a particular crime may occur. Data privacy is the main challenge that organizations face when building this type of predictive models. Federated learning (FL) is a promising approach that overcomes data security and privacy challenges, as it enables organizations to build a machine learning model based on distributed datasets without sharing raw data or violating data privacy. In this paper, a federated long short- term memory (LSTM) model is proposed and compared with a traditional LSTM model. Proposed model is developed using TensorFlow Federated (TFF) and the Keras API to predict the number of crimes. The proposed model is applied on the Boston crime dataset. The proposed model's parameters are fine tuned to obtain minimum loss and maximum accuracy. The proposed federated LSTM model is compared with the traditional LSTM model and found that the federated LSTM model achieved lower loss, better accuracy, and higher training time than the traditional LSTM model.

• 한국멀티미디어학회논문지
• /
• 제20권5호
• /
• pp.769-781
• /
• 2017

Deep learning networks like Convolutional Neural Networks (CNNs) show successful performances in many computer vision applications such as image classification, object detection, and so on. For implementation of deep learning networks in embedded system with limited processing power and memory, deep learning network may need to be simplified. However, simplified deep learning network cannot learn every possible scene. One realistic strategy for embedded deep learning network is to construct a simplified deep learning network model optimized for the scene images of the installation place. Then, automatic training will be necessitated for commercialization. In this paper, as an intermediate step toward automatic training under fisheye camera environments, we study more precise human localization in fisheye images, and propose an accurate human localization method, Automatic Ground-Truth Labelling Method (AGTLM). AGTLM first localizes candidate human object bounding boxes by utilizing GoogLeNet-LSTM approach, and after reassurance process by GoogLeNet-based CNN network, finally refines them more correctly and precisely(tightly) by applying saliency object detection technique. The performance improvement of the proposed human localization method, AGTLM with respect to accuracy and tightness is shown through several experiments.

• 혜화의학회지
• /
• 제12권1호
• /
• pp.11-26
• /
• 2003

Alzheimer's disease (AD) is characterized by amyloid deposition and associated loss of neunons in brain regions involved in learning and memory processes. Several causes of evidence support that the congnitive disturbance is closed associated with the deficit of cerebral acetylcholine neurotransmission, and the effect of carboxyl terminal 105 amino acid fragment (CT105) of the amyloid precursor protein (APP) on the gene expression of proinflammatory cytokines. We tested it on the scopolamine-induced amnesia model of the ICR mouse using the Morris water maze with repeated orally administration of 1st Green-Tea extract (200 mg/kg) and 2nd Green-Tea extract (200 mg/kg). The Green-Tea prevents impairment of learning and memory and neuronal loss in mouse models of cognitive disturbance and it demonstrated selectivity for inhibition of acetylcholinesterase (AChE). Furthermore, the repeated administration of Green-Tea, CT105-induced alzheimer's mouse model showed central cholinergic activity by ameliorates learning and memory impairment, and isolation of CD14 microglia showed significantly decreases intracellular release of the proinflammatory cytokines tumor necrosis factor-${\alpha}$, interleukin-$1{\beta}$ and reactive oxygen species (ROS). Because of its composite profile, oral therapeutic index and a prophylactic, Green-Tea is considered the better therapeutic candidate for the treatment of Alzheimer's disease.

• The Korean Journal of Physiology and Pharmacology
• /
• 제18권3호
• /
• pp.201-209
• /
• 2014

The present study was designed to investigate the efficacy of selective $ET_A$ receptor antagonist, ambrisentan on hyperhomocysteinemia-induced experimental vascular dementia. L-methionine was administered for 8 weeks to induce hyperhomocysteinemia and associated vascular dementia in male rats. Ambrisentan was administered to L-methionine-treated effect rats for 4 weeks (starting from $5^{th}$ to $8^{th}$ week of L-methionine treatment). On $52^{nd}$ day onward, the animals were exposed to the Morris water maze (MWM) for testing their learning and memory abilities. Vascular endothelial function, serum nitrite/nitrate levels, brain thiobarbituric acid reactive species (TBARS), brain reduced glutathione (GSH) levels, and brain acetylcholinesterase (AChE) activity were also measured. L-methionine-treated animals showed significant learning and memory impairment, endothelial dysfunction, decrease in/serum nitrite/nitrate and brain GSH levels along with an increase in brain TBARS levels and AChE activity. Ambrisentan significantly improved hyperhomocysteinemia-induced impairment of learning, memory, endothelial dysfunction, and changes in various biochemical parameters. These effects were comparable to that of donepezil serving as positive control. It is concluded that ambrisentan, a selective $ET_A$ receptor antagonist may be considered as a potential pharmacological agent for the management of hyperhomocysteinemia-induced vascular dementia.

• 동의신경정신과학회지
• /
• 제10권1호
• /
• pp.39-51
• /
• 1999

The effects of Hyangbujapalmultang on the enhancement of learning and memory of AD model rats were studied with Morris water maze. Sample group was electrolytically lesioned on nucleus basalis of Meynert(nbM), and then daily treated with the medicine for two months. Control group with nbM lesion, and sham group with the sham operation were treated the vehicle for the same duration. The following results were observed. 1. As the learning trials of Morris water maze were proceeded, sham group showed the escape latency from $54.7{\pm}2.28$ seconds in 1st trial to $13.3{\pm}3.27$ seconds in 7th. The control group showed the escape latency from $58.0{\pm}1.78$ seconds in 1st trial to $51.3{\pm}3.52$ seconds in 7th. The sample group showed the escape latency from $57.0{\pm}2.21$ seconds in 1st trial to $28.4{\pm}4.82$ seconds in 7th. Therefore, these data shows that all three groups were improved in learning capacity as trials were proceeded, but the sample group showed more prominent improvement in learning compared with the control group(p<0.05). 2. In memory retention test of Morris water maze that counts the staying time in the target area during 30 seconds of freely swimming period, sham group stayed for $4.81{\pm}1.15$ seconds, the control group stayed for $1.27\pm}0.78$ seconds, and the sample group stayed for $4.17{\pm}1.47$ seconds. The analysis of the memory retention data shows that the sample group marked more improvement in memory retention compared with the control group, but could not obtained statistically significant result(p<0.05). With the experimental results above, Hyangbujapalmultang can be supposed to have the improving effects on the learning of AD model rats induced by electrolytic lesion of nbM.

• 대한본초학회지
• /
• 제21권4호
• /
• pp.189-195
• /
• 2006

Objectives : It has been reported previously that the roots of Scutellaria baicalensis (known as Huang-Gum in Korean, henceforth referred to as S. baicalensis) could prevent neuronal cell death after global cerebral ischemia. In Genuine Korean medicine, S. baicalensis is known to relieve fever in upper body, and it was thus thought to be able to alleviate deteriorations in brain function. Methods : The protective effects of S. baicalensis against post-stroke memory retardation using 4-vessel occlusion model were examined in the present study. Results : S. baicalensis was shown to significantly alleviate the deficits in learning and memory by increasing the fraction of time spent in the quadrant in which the platform was initially placed ($34.9\;{\pm}\;3.2%$, p < 0.05) compared to that of the ischemia group ($28.0\;{\pm}\;2.5%$). The cytoprotective effect of S. baicalensis on CA1 hippocampal neurons was evaluated by measuring the neuronal cell density. Neuronal cell density in S. baicalensis extracts-treated ischemia group ($138.0\;{\pm}\;13.6\;cells/mm^2$) was significantly increased compared to saline-treated ischemia group ($22.1 \;{\pm}\;9.3\;cells/mm^2$, p < 0.05). In the study of OX-42 immunohistochemistry, S. baicalensis could decrease the micrgial activation in hippocampus after brain ischemia. Conclusion : These results may provide experimental support for the use of S. baicalensis in treating post-stroke memory impairment.

• 구강회복응용과학지
• /
• 제23권2호
• /
• pp.179-186
• /
• 2007

The decrease of masticatory function caused by tooth loss leads to a decrease of cerebral blood flow volume resulting in impairment of cognitive function and learning memory disorder. However, the reduced mastication-mediated morphological alteration in the central nervous system (CNS) responsible for senile deficit of cognition, learning and memory has not been well documented. In this study, the effect of the loss of the molar teeth (molarless condition) on the hippocampal expression of glial fibrillary acidic protein (GFAP) protein was studied by immunohistochemical techniques. The results were as follows : 1. The molarless mice showed a lower density of pyramidal cells in the cornu ammonis 1 (CA1) and dentate gyrus (DG) region of the hippocampus than control mice. 2. Immunohistochemical analysis showed that the molarless condition enhanced the time-dependent increase in the cell density and hypertrophy of GFAP immunoreactivity in the CA1 region of the hippocampus. The molarless condition enhanced an time-dependent decrease in the number of neurons in the hippocampal formation and the time-dependent increase in the number and hypertrophy of GFAP-labeled cells in the same region. The data suggest a possible link between reduced mastication and histological changes in hippocampal formation that may be one risk factor for senile impairment of cognitive function and spatial learning memory.