• Journal of the Korean Geotechnical Society
• /
• v.16 no.2
• /
• pp.71-77
• /
• 2000

In this paper, consolidation settlements with time at vertical drain sites were predicted by artificial neural networks. Laboratory test results and field measurements of two vertical drain sites were used for training and testing neural networks. Predicted consolidation settlements by trained artificial neural networks were compared with measured settlements by field instrumentation. To improve the prediction accuracy, modular artificial neural networks were studied. From the results of applying artificial neural networks to the same situation, it was shown that modular artificial neural network model was more accurate for the prediction of the consolidation settlements than the general model.

• Journal of Korean Society on Water Environment
• /
• v.27 no.4
• /
• pp.533-541
• /
• 2011

Algal blooms have caused problems for drinking water as well as eutrophication. However it is difficult to control algal blooms by current warning manual in rainy season because the algal blooms happen in a few days. The water quality data, which have high correlations with Chlorophyll-a on Daecheongho station, were analyzed and chosen as input data of Artificial Neural Networks (ANN) for training pattern changes. ANN was applied to early forecasting of algal blooms, and ANN was assessed by forecasting errors. Water temperature, pH and Dissolved oxygen were important factors in the cross correlation analysis. Some water quality items like Total phosphorus and Total nitrogen showed similar pattern to the Chlorophyll-a changes with time lag. ANN model (No. 3), which was calibrated by water temperature, pH and DO data, showed lowest error. The combination of 1 day, 3 days, 7 days forecasting makes outputs more stable. When automatic monitoring data were used for algal bloom forecasting in Daecheong reservoir, ANN model must be trained by just input data which have high correlation with Chlorophyll-a concentration. Modular type model, which is combined with the output of each model, can be effectively used for stable forecasting.

• Journal of the Korean Society of Industry Convergence
• /
• v.27 no.1
• /
• pp.87-95
• /
• 2024

This study is on modulo scheduling algorithms for multicore processor in machine learning applications. Machine learning algorithms are designed to perform a large amount of operations such as vectors and matrices in order to quickly process large amounts of data stream. To support such large amounts of computations, processor architectures to support applications such as artificial intelligence, neural networks, and machine learning are designed in the form of parallel processing such as multicore. To effectively utilize these multi-core hardware resources, various compiler techniques are being used and studied. In this study, among these compiler techniques, we analyzed the modular scheduler, which is especially important in one core's computation pipeline. This paper looked at and compared the iterative modular scheduler and the swing modular scheduler, which are the most widely used and studied. As a result, both schedulers provided similar performance results, and when measuring register pressure as an indicator, it was confirmed that the swing modulo scheduler provided slightly better performance. In this study, a technique that divides recurrence edge is proposed to improve the minimum initiation interval of the modulo schedulers.

• Nuclear Engineering and Technology
• /
• v.53 no.8
• /
• pp.2547-2555
• /
• 2021

The frequency of reactor coolant leakage is expected to increase over the lifetime of a nuclear power plant owing to degradation mechanisms, such as flow-acceleration corrosion and stress corrosion cracking. When loss of coolant accidents (LOCAs) occur, several parameters change rapidly depending on the size and location of the cracks. In this study, leak flow during LOCAs is predicted using a deep fuzzy neural network (DFNN) model. The DFNN model is based on fuzzy neural network (FNN) modules and has a structure where the FNN modules are sequentially connected. Because the DFNN model is based on the FNN modules, the performance factors are the number of FNN modules and the parameters of the FNN module. These parameters are determined by a least-squares method combined with a genetic algorithm; the number of FNN modules is determined automatically by cross checking a fitness function using the verification dataset output to prevent an overfitting problem. To acquire the data of LOCAs, an optimized power reactor-1000 was simulated using a modular accident analysis program code. The predicted results of the DFNN model are found to be superior to those predicted in previous works. The leak flow prediction results obtained in this study will be useful to check the core integrity in nuclear power plant during LOCAs. This information is also expected to reduce the workload of the operators.

• Journal of KIISE
• /
• v.44 no.1
• /
• pp.51-56
• /
• 2017

The research area for solving question answering (QA) problems using artificial intelligence models is in a methodological transition period, and one such architecture, the dynamic memory network (DMN), is drawing attention for two key attributes: its attention mechanism defined by neural network operations and its modular architecture imitating cognition processes during QA of human. In this paper, we increased accuracy of the inferred answers, by adapting an automatic data augmentation method for lacking amount of training data, and by improving the ability of time perception. The experimental results showed that in the 1K-bAbI tasks, the modified DMN achieves 89.21% accuracy and passes twelve tasks which is 13.58% higher with passing four more tasks, as compared with one implementation of DMN. Additionally, DMN's word embedding vectors form strong clusters after training. Moreover, the number of episodic passes and that of supporting facts shows direct correlation, which affects the performance significantly.