POSBIOTM—NER: a trainable biomedical named-entity recognition system

Song, Yu,Kim, Eunju,Lee, Gary Geunbae,Yi, Byoung-kee Oxford University Press 2005 Bioinformatics Vol.21 No.11

<P><B>Summary:</B> POSBIOTM–NER is a trainable biomedical named-entity recognition system. POSBIOTM–NER can be automatically trained and adapted to new datasets without performance degradation, using CRF (conditional random field) machine learning techniques and automatic linguistic feature analysis. Currently, we have trained our system on three different datasets. GENIA–NER was trained based on GENIA Corpus, GENE–NER based on BioCreative data and GPCR–NER based on our own POSBIOTM/NE corpus, respectively, which would be used in GPCR-related pathway extraction.</P><P><B>Availability:</B> http://isoft.postech.ac.kr/Research/BioNER/POSBIOTM/NER/main.html</P><P><B>Contact:</B> songyu@postech.ac.kr</P>

Intrusion Detection in Aviation Terminal Region Petri Net with Non-arc and Unchanged Library

Song Yu,He Xiaoli,Ralf Volker Binsack 보안공학연구지원센터 2016 International Journal of Security and Its Applicat Vol.10 No.5

Due to the complex structure and the large number of the running aircrafts in the application of the intrusion detection in terminal regions, the defense conflict dangers and the influence on aviation safety exist in such terminal region. Therefore, Petri net framework with non-arc and unchanged library is constructed in this paper to propose the terminal intrusion detection scheme. Firstly, the aviation operation structure of the terminal region is analyzed and the constraint model of the terminal region is constructed as the terminal intrusion detection basis according to Petri net model framework; secondly, the aviation constraint model of the terminal region constructed thereby and the principle of non-arc and unchanged library are adopted to establish the terminal intrusion detection control strategy, and the control decision is made according to the transition activation for the aviation instructions; finally, the experimental analysis of the practical cases shows that the proposed terminal intrusion detection scheme can effectively handle the terminal intrusion detection problem and reduce the workload of the controllers.

Rock bridge fracture model and stability analysis of surrounding rock in underground cavern group

Song Yu,Wei-shen Zhu,Wei-min Yang,Dun-fu Zhang,Qing-song Ma 국제구조공학회 2015 Structural Engineering and Mechanics, An Int'l Jou Vol.53 No.3

Many hydropower stations in southwest China are located in regions of brittle rock mass with high geo-stresses. Under these conditions deep fractured zones often occur in the sidewalls of the underground caverns of a power station. The theory and methods of fracture and damage mechanics are therefore adopted to study the phenomena. First a flexibility matrix is developed to describe initial geometric imperfections of a jointed rock mass. This model takes into account the area and orientation of the fracturedsurfaces of multiple joint sets, as well as spacing and density of joints. Using the assumption of the equivalent strain principle, a damage constitutive model is established based on the brittle fracture criterion. In addition the theory of fracture mechanics is applied to analyze the occurrence of secondary cracks duringa cavern excavation. The failure criterion, for rock bridge coalescence and the damage evolution equation, has been derived and a new sub-program integrated into the FLAC-3D software. The model has then beenapplied to the stability analysis of an underground cavern group of a hydropower station in Sichuan province, China. The results of this method are compared with those obtained by using a conventional elasto-plastic model and splitting depth calculated by the splitting failure criterion proposed in a previous study. The results are also compared with the depth of the relaxation and fracture zone in the surrounding rock measured by field monitoring. The distribution of the splitting zone obtained both by the proposed model and by the field monitoring measurements are consistent to the validity of the theory developed herein.

A gap element for treating non-matching discrete interfaces

Song, Y. U.,Youn, S. K.,Park, K. C. Springer Science + Business Media 2015 Computational mechanics Vol.56 No.3

Virtual gap element approach for the treatment of non-matching interface using three-dimensional solid elements

Song, Y. U.,Youn, S. K.,Park, K. C. Springer Science + Business Media 2017 Computational mechanics Vol.60 No.4

Signal Hunting Algorithm Based on Difference Grouping Mechanism in Sensor Network

Song Yu,He Xiaoli,Ralf Volker Binsack 보안공학연구지원센터 2016 International Journal of Future Generation Communi Vol.9 No.8

In order to solve such problems as the difficult realization of accurate positioning and multidimensional dynamic elastic range measurement and the poor reading of positioning data caches during the range measurement process in wireless sensor network, a wireless sensor network positioning algorithm based on neighbor spin-hop iteration mechanism. Firstly, the neighbor clustering process is executed for the nodes in the wireless sensor network, namely: the first-order neighbor nodes and the second-order neighbor nodes are taken as the network neighborhood. Secondly, other network nodes are positioned according to the rejection and attraction conditions of the neighborhood radius, and meanwhile the iteration mechanism is adopted to realize error minimization so as to obtain the accurate position coordinate of the node. The simulation result shows: compared with currently widely applied two-dimensional non-iterative accurate positioning algorithms, such as RSLM algorithm and SOCP algorithm, the new algorithm is provided with clustering mechanism for accurate positioning and the error correction method based on iteration mechanism, thus to not only effectively reduce the node position error and improve positioning accuracy during multidimensional range measurement, but also improve network packet delivery ratio and reduce network control overhead.

安部公房研究在中国

何嵩昱(He-songyu),戴書亞(Dai-shuya) 동아인문학회 2013 동아인문학 Vol.26 No.-

Abe Kobe is a very famous writer in the literary history of Japan. From late 1970s, the Chinese started to introduce and got research about him. Nowadays, there are many research achievements about this writer in China. From the superficiality of the earlier time to the profoundness of today, the Chinese scholars are working hardly on the research of Abe Kobe. However, all of those research results cannot show his literary status. This paper collates the related research achievements of Abe Kobe in Chinese academic. On one hand, it combs the history line of the research. On the other hand, it concludes the focus of the research. And then, it analyze the deficiency of those research and the reason. Finally, the paper is concluded by pointing out some futher research directions on the research of Abe Kobe.

Analysis of Wind Turbine Fatigue Load Data Model

Jian Yang,Songyue Zheng,Dongran Song,Guoxun Xiao,Junye Ma,Young Hoon Joo 전력전자학회 2019 ICPE(ISPE)논문집 Vol.2019 No.5

The mutual coupling between the average wind speed, the turbulence intensity and the active power regulation affects the fatigue load of the wind turbine (WT), and the time series of the wind speed is random, which causes the uncertainty of the fatigue load. In order to clarify the influence of multiple factors on the fatigue load of WT, this paper discusses the fatigue load of WT and its influencing factors based on data mining method. The Bladed software is used to combine various factors, and a large number of simulation experiments are carried out to obtain the fatigue load data distributed in the three directions of the four components of the blade, hub, yaw and tower. Kernel density function is employed to analyzed distribution of fatigue load data. The fatigue load data model is established to find out the influence of average wind speed, turbulence intensity and active power regulation on fatigue load.

Evaluation of Geotechnical Parameters of a Lagoonal Clay Deposit in Jiangsu Lixia River Area of China by Seismic Piezocone Tests

Guojun Cai,Songyu Liu,Anand J. Puppala 대한토목학회 2016 KSCE JOURNAL OF CIVIL ENGINEERING Vol.20 No.5

The use of the Seismic Piezocone Test (SCPTU) in geotechnical site investigation offers field assessment on stratigraphy and soil behavior. Compared with some traditional investigation methods such as drilling, sampling and field inspecting method or laboratory test procedures, SCPTU can greatly accelerate the field work and hereby reduce corresponding operation cost. The results of a site investigation on the soft, normally to slightly overconsolidated, Lixia River lagoonal clay deposit underlying several expressway lines, including a seismic piezocone penetration test, undisturbed sampling from an adjacent borehole, and laboratory testing, are reported. The determination of soil profile and the main geotechnical properties is facilitated by the use of seismic piezocone penetration test that can provide up to four independent reading with depth from a single sounding. The coupled use of normalized cone resistance and pore pressure parameter provides a simple and quick soil classification for highly interbedding of Lixia River lagoon ground. Comparison of the results reveals the validity of SCPTU tests to interpret the engineering properties of Lixia River lagoonal clay. These data will also be particularly useful for future local construction work.

Compaction and Mechanical Characteristics and Stabilization Mechanism of Carbonated Reactive MgO-Stabilized Silt

Guanghua Cai,Songyu Liu 대한토목학회 2017 KSCE JOURNAL OF CIVIL ENGINEERING Vol.21 No.7

The reinforcement technology of carbonation based on reactive magnesia (MgO) and carbon dioxide (CO2) is a low-carbon and high-efficiency foundation treatment method. This paper investigates the compaction, mechanical and microstructural characteristics of carbonated reactive MgO-stabilized silt with varying MgO-soil ratios, carbonation time and water-soil ratios. The results indicate that the maximum dry density of uncarbonated reactive MgO-stabilized silt increases while the optimum moisture content reduces compared to the parent soil. The unconfined compressive strength of reactive MgO-stabilized soil was found to have increased after CO2 carbonation for several hours. With increasing MgO-soil ratio and carbonation time, the failure mode changes from elasticplastic to brittleness, and the failure strain of carbonated specimens mainly ranges between 0.8% and 1.6% and the ratio of the deformation modulus to unconfined compressive strength is about 30 to 200. The water-soil ratio has a slight influence on the evolution of strength. Scanning Electron Microscopy (SEM) and X-ray diffraction (XRD) analyses have indicated that the carbonation products facilitate the strength growth of reactive MgO-stabilized silt. Thermogravimetric analysis (TGA) shows that CO2 uptake increases with increasing carbonation time and achieves the highest under the MgO-soil ratio of 20%, carbonation time of 6 h, and water-soil ratio of 25%. The stabilization mechanism of carbonated reactive MgO-stabilized silt is proposed according to microstructure analyses, providing a deeper understanding of the application of the reactive MgO carbonation technology in the ground reinforcement.