• Journal of KIISE
• /
• v.42 no.4
• /
• pp.442-450
• /
• 2015

This paper proposes a method for reducing false conflicts in signature-based eager hardware transactional memory (HTM). The method tracks the information on all cache blocks that are accessed by a transaction. If the information provides evidence that there are no conflicts for a given transactional request from another core, the method prevents the occurrence of a false conflict by forcing the HTM to ignore the decision based on the signature. The method is very effective in reducing false conflicts and the associated unnecessary transaction stalls and aborts, and can be used to improve the performance of the multicore processor that implements the signature-based eager HTM. When running the STAMP benchmark on a 16-core processor that implements the LogTM-SE, the increased speed (decrease in execution time) achieved with the use of the method is 20.6% on average.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.8
• /
• pp.243-250
• /
• 2020

Knowledge graph-based question answering not only requires deep understanding of the given natural language questions, but it also needs effective reasoning to find the correct answers on a large knowledge graph. In this paper, we propose a deep neural network model for effective reasoning on a knowledge graph, which can find correct answers to complex questions requiring multi-hop inference. The proposed model makes use of highly expressive bilinear graph neural network (BGNN), which can utilize context information between a pair of neighboring nodes, as well as allows bidirectional feature propagation between each entity node and one of its neighboring nodes on a knowledge graph. Performing experiments with an open-domain knowledge base (Freebase) and two natural-language question answering benchmark datasets(WebQuestionsSP and MetaQA), we demonstrate the effectiveness and performance of the proposed model.

• The Journal of the Korea Contents Association
• /
• v.10 no.4
• /
• pp.84-90
• /
• 2010

DBMS is used for a range of applications from data warehousing through on-line transaction processing. As a result of this demand, DBMS has continued to grow in terms of its size. This growth invokes the most important issue of manually tuning the performance of DBMS. The DBMS tuning should be adaptive to the type of the workload put upon it. But, identifying workloads in mixed database applications might be quite difficult. Therefore, a method is necessary for identifying workloads in the mixed database environment. In this paper, we propose a SVM workload classifier to automatically identify a DBMS workload. Database workloads are collected in TPC-C and TPC-W benchmark while changing the resource parameters. Parameters for SVM workload classifier, C and kernel parameter, were chosen experimentally. The experiments revealed that the accuracy of the proposed SVM workload classifier is about 9% higher than that of Decision tree, Naive Bayes, Multilayer perceptron and K-NN classifier.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.4
• /
• pp.574-579
• /
• 2009

There have been much difficulties to construct an optimized neural network in complex nonlinear regression problems such as selecting the networks structure and avoiding overtraining problem generated by noise. In this paper, we propose a stepwise constructive method for neural networks using a flexible incremental algorithm. When the hidden nodes are added, the flexible incremental algorithm adaptively controls the number of hidden nodes by a validation dataset for minimizing the prediction residual error. Here, the ELM (Extreme Learning Machine) was used for fast training. The proposed neural network can be an universal approximator without user intervene in the training process, but also it has faster training and smaller number of hidden nodes. From the experimental results with various benchmark datasets, the proposed method shows better performance for real-world regression problems than previous methods.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.395-395
• /
• 2000

Optimization of existing processes becomes more important than the past as environmental problems and concerns about energy savings stand out. When we can model a process mathematically, we can easily optimize it by using the model as constraints. However, modeling is very difficult for most chemical processes as they include numerous units together with their correlation and we can hardly obtain parameters. Therefore, optimization that is based on the process models is, in turn, hard to perform. Especially, f3r unknown processes, such as bioprocess or microelectronics materials process, optimization using mathematical model (first principle model) is nearly impossible, as we cannot understand the inside mechanism. Consequently, we propose a few optimization method using empirical model evolutionarily instead of mathematical model. In this method, firstly, designing experiments is executed fur removing unecessary experiments. D-optimal DOE is the most developed one among DOEs. It calculates design points so as to minimize the parameters variances of empirical model. Experiments must be performed in order to see the causation between input variables and output variables as only correlation structure can be detected in historical data. And then, using data generated by experiments, empirical model, i.e. response surface is built by PLS or MLR. Now, as process model is constructed, it is used as objective function for optimization. As the optimum point is a local one. above procedures are repeated while moving to a new experiment region fur finding the global optimum point. As a result of application to the pulp digester benchmark model, kappa number that is an indication fur impurity contents decreased to very low value, 3.0394 from 29.7091. From the result, we can see that the proposed methodology has sufficient good performance fur optimization, and is also applicable to real processes.

• Journal of Energy Engineering
• /
• v.24 no.2
• /
• pp.91-102
• /
• 2015

A thermal-hydraulic design and performance analysis computer code for a once-through steam generator using straight tubes is developed. To benchmark the developed physical models and computer code, an once-through steam generator developed by other designer is simulated and the calculated results are compared with the design data. Also, the same steam generator is analyzed with the best-estimate thermal-hydraulic system code, MARS, for the code-to-code validation. The overall characteristics of heat transfer area, pressure and temperature distributions calculated by the developed code show general agreements with the published design data as well as the analysis results of MARS. It is demonstrated that the developed code can be utilized for diverse purposes, such as, sensitivity analyses and optimum thermal design of a once-through steam generator.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.10
• /
• pp.1187-1191
• /
• 2011

In this paper, the development of an adjustable load limiter is presented, which is a component of the seat belt. The adjustable load limiter is loaded at different levels for varied weights and heights of occupant. The recent regulation FMVSS 208 demands strict safety standards for different percentiles of dummy size. In this work, high- and low-level load conditions are proposed according to dummy scale and thoracic injury criteria. The suggested load conditions were verified by performing a sled test using the benchmark model. A dual-level load limiter has been developed on the basis of these tests. Experiments were conducted on the product performance, and finite element analysis was carried out; the results confirmed the points for improvement.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.7
• /
• pp.3093-3115
• /
• 2020

Network anomaly detection system plays an essential role in detecting network anomaly and ensuring network security. Anomaly detection system based machine learning has become an increasingly popular solution. However, due to the unbalance and high-dimension characteristics of network traffic, the existing methods unable to achieve the excellent performance of high accuracy and low false alarm rate. To address this problem, a new network anomaly detection method based on data balancing and recursive feature addition is proposed. Firstly, data balancing algorithm based on improved KNN outlier detection is designed to select part respective data on each category. Combination optimization about parameters of improved KNN outlier detection is implemented by genetic algorithm. Next, recursive feature addition algorithm based on correlation analysis is proposed to select effective features, in which a cross contingency test is utilized to analyze correlation and obtain a features subset with a strong correlation. Then, random forests model is as the classification model to detection anomaly. Finally, the proposed algorithm is evaluated on benchmark datasets KDD Cup 1999 and UNSW_NB15. The result illustrates the proposed strategies enhance accuracy and recall, and decrease the false alarm rate. Compared with other algorithms, this algorithm still achieves significant effects, especially recall in the small category.

• Korean Journal of Ecology and Environment
• /
• v.50 no.1
• /
• pp.70-78
• /
• 2017

The Hwasan wetland vegetation is observed at mountain basin (644~780 m a.s.l.) where had become a potential land for indigenous people since prehistoric period. We phytosociologically investigated old-growth alder (Alnus japonica) forests using the $Z\ddot{u}rich$-Montpellier School's method and analyzed their spatial distribution pattern by actual vegetation map. Species performance was determined by using coverage and r-NCD. Viburnum opulus var. calvescens-Alnus japonica community syntaxonomically belonging to the Alnetea japonicae was first described and composed of three subunits: Salix koreensis subcommunity, typical subcommunity, and Pyrus ussuriensis subcommunity. Present plant community was compared with vicariant syntaxa such as Molinia japonica-Alnus japonica community, Rhamno nipponicae-Alnetum japonicae, and Aceri-Salicetum koreensis. Hwasan's alder forest, an alluvial terrace vegetation type on valley fan in the montane zone, is evaluated as vegetation class [I], which is a sort of benchmark plant community potentially on mountain wetlands in southeastern part of the Korean Peninsula. Simultaneously we suggested an establishment of the national strategy for habitat conservation free from hydrologically radical transform due to military utilization.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.11-17
• /
• 2007

One of the popular methods used for pattern classification is the MBR (Memory-Based Reasoning) algorithm. Since it simply computes distances between a test pattern and training patterns or hyperplanes stored in memory, and then assigns the class of the nearest training pattern, it cannot explain how the classification result is obtained. In order to overcome this problem, we propose an incremental teaming algorithm based on RPA (Recursive Partition Averaging) to extract IF-THEN rules that describe regularities inherent in training patterns. But rules generated by RPA eventually show an overfitting phenomenon, because they depend too strongly on the details of given training patterns. Also RPA produces more number of rules than necessary, due to over-partitioning of the pattern space. Consequently, we present the IREA (Incremental Rule Extraction Algorithm) that overcomes overfitting problem by removing useless conditions from rules and reduces the number of rules at the same time. We verify the performance of proposed algorithm using benchmark data sets from UCI Machine Learning Repository.