• Korean Journal of Artificial Intelligence
• /
• v.5 no.1
• /
• pp.29-37
• /
• 2017

The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the Gradient Descent Optimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2003.10a
• /
• pp.3-3
• /
• 2003

With biomedical literature expanding so rapidly, there is an urgent need to discover and organize knowledge extracted from texts. Although factual databases contain crucial information the overwhelming amount of new knowledge remains in textual form (e.g. MEDLINE). In addition, new terms are constantly coined as the relationships linking new genes, drugs, proteins etc. As the size of biomedical literature is expanding, more systems are applying a variety of methods to automate the process of knowledge acquisition and management. In my talk, I focus on the project, GENIA, of our group at the University of Tokyo, the objective of which is to construct an information extraction system of protein - protein interaction from abstracts of MEDLINE. The talk includes (1) Techniques we use fDr named entity recognition (1-a) SOHMM (Self-organized HMM) (1-b) Maximum Entropy Model (1-c) Lexicon-based Recognizer (2) Treatment of term variants and acronym finders (3) Event extraction using a full parser (4) Linguistic resources for text mining (GENIA corpus) (4-a) Semantic Tags (4-b) Structural Annotations (4-c) Co-reference tags (4-d) GENIA ontology I will also talk about possible extension of our work that links the findings of molecular biology with clinical findings, and claim that textual based or conceptual based biology would be a viable alternative to system biology that tends to emphasize the role of simulation models in bioinformatics.

• Journal of Korea Society of Industrial Information Systems
• /
• v.21 no.5
• /
• pp.11-17
• /
• 2016

In this paper, we propose a method of correcting depth image by the plane modeling and then improving the coding performance. We model a plane by using the least squares method to the horizontal and vertical directions including the target pixel, and then determine that the predicted plane is suitable from the estimate error. After that, we correct the target pixel by the plane mode. The proposed method can correct not only the depth image composed the plane but also the complex depth image. From the simulation result that measures the entropy power, which can estimate the coding performance, we can see that the coding performance by the proposed method is improved up to 80.2%.

• Environmental Engineering Research
• /
• v.17 no.3
• /
• pp.125-132
• /
• 2012

The removal of Cd(II) and Cu(II) from aqueous solution by an agricultural solid waste biomass prepared from Moringa oleifera bark (MOB) was investigated. The biosorbent was characterized by Fourier transform infrared spectroscopy and elemental analysis. Furthermore, the effect of initial pH, contact time, biosorbent dosage, initial metal ion concentration and temperature on the biosorption of Cd(II) and Cu(II) were studied using the batch sorption technique. Kinetic studies indicated that the biosorption process of the metal ions followed the pseudo-second order model. The biosorption data was analyzed by the Langmuir, Freundlich, Dubinin-Radushkevich, and Temkin isotherm models. Based on the Langmuir isotherm, the maximum biosorption capacities for Cd(II) and Cu(II) onto MOB were 39.41 and 36.59 mg/g at 323 K, respectively. The thermodynamic parameters, Gibbs free energy (${\Delta}G^o$), enthalpy (${\Delta}H^o$), and entropy (${\Delta}S^o$) changes, were also calculated, and the values indicated that the biosorption process was endothermic, spontaneous and feasible in the temperature range of 303-323 K. It was concluded that MOB powder can be used as an effective, low cost, and environmentally friendly biosorbent for the removal of Cd(II) and Cu(II) ions from aqueous solution.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.11
• /
• pp.2068-2086
• /
• 2011

Based on the quincunx sub-sampling grid, the New Interleaved Hierarchical INTerpolation (NIHINT) method is recognized as a superior pyramid data structure for the lossless and progressive coding of natural images. In this paper, we propose a new image interpolation algorithm, Edge Adaptive Hierarchical INTerpolation (EAHINT), for a further reduction in the entropy of interpolation errors. We compute the local variance of the causal context to model the strength of a local edge around a target pixel and then apply three statistical decision rules to classify the local edge into a strong edge, a weak edge, or a medium edge. According to these local edge types, we apply an interpolation method to the target pixel using a one-directional interpolator for a strong edge, a multi-directional adaptive weighting interpolator for a medium edge, or a non-directional static weighting linear interpolator for a weak edge. Experimental results show that the proposed algorithm achieves a better compression bit rate than the NIHINT method for lossless image coding. It is shown that the compression bit rate is much better for images that are rich in directional edges and textures. Our algorithm also shows better rate-distortion performance and visual quality for progressive image transmission.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2002.07b
• /
• pp.939-944
• /
• 2002

PD phenomena can be regarded as a deterministic dynamical process where PD should be occurred if the local electric field be reached to be sufficiently high. And thus, its mathematical model can be described by either difference equations or differential equations using several state variables obtained from the time sequential measured data of PD signals. These variables can provide rich and complex behavior of detectable time series, for which Chaos theory can be employed. In this respect, a new PD pattern recognition method is proposed and named as 'Chaotic Analysis of Partial Discharges (CAPD)' for this work. For this purpose, six types of specimen are designed and made as the models of the possible defects that may cause sudden failures of the underground power transmission cables under service, and partial discharge signals, generated from those samples, are detected and then analyzed by means of CAPD. Throughout the work, qualitative and quantitative properties related to the PD signals from different defects are analyzed by use of attractor in phase space, information dimensions ($D_0$ and D2), Lyapunov exponents and K-S entropy as well. Based on these results, it could be pointed out that the nature of defect seems to be identified more distinctively when the CAPD is combined with traditional statistical method such as PRPDA. Furthermore, the relationship between PD magnitude and the occurrence timing is investigated with a view to simulating PD phenomena.

• Korean Journal of Materials Research
• /
• v.9 no.4
• /
• pp.341-344
• /
• 1999

We reported compositional separation(CS) into Co-enriched and Cri-enriched components inside the grains of Co-Cr based thin films prepared by rf sputtering. CS strongly depends on the sputtering conditions of substrate temperature and target composition. Tuning the microstructure of the Co-Cr films is important in order to employ the CS for high-density magnetic recording. We investigated the origin of CS from thermodynamic viewpoint. We employ a spinodal decomposition-like model to describe the origin of the CS in Co-Cr films. We consider the total free energy of the Co-Cr films as the sum of several free energies of; 1) thermodynamic mixing entropy of a binary solid solution, 2) magnetic ordering interaction(MOI) energy below the Curie temperature, and 3) excess interaction energy(XS) caused by the sputtering process as a function of temperature and composition. Those energies distorted the total free energy like the spinodal decomposition and caused the compositionally separated fine microstructure inside the grains. If the second derivative of the total free energy with respect to Cr composition becomes negative at a given substrate temperature, we may observe a metastable compositional separation inside the Co-Cr alloy films. We expect to exploit the microstructure of CS for ultra-high density magnetic recording.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.10
• /
• pp.4952-4975
• /
• 2018

An induced hesitant linguistic aggregation operator is investigated in the paper, in which, hesitant fuzzy linguistic evaluation values are associated with probabilistic information. To deal with these hesitant fuzzy linguistic information, an induced hesitant fuzzy linguistic probabilistic ordered weighted averaging (IHFLPOWA) operator is proposed, monotonicity, boundary and idempotency of IHFLPOWA are proved. Then andness, orness and the entropy of dispersion of IHFLPOWA are analyzed, which are used to characterize the weighting vector of the operator, these properties show that IHFLPOWA is extensions of the induced linguistic ordered weighted averaging operator and linguistic probabilistic aggregation operator. In this paper, IHFLPOWA is utilized to gather linguistic information and create fuzzy ontologies, and a movie fuzzy ontology as an illustrative case study is used to show the elaboration of the proposed method and comparison with the existing linguistic aggregation operators, it seems that the IHFLPOWA operator is an useful and alternative operator for dealing with hesitant fuzzy linguistic information with probabilistic information.

• The Journal of Korean Association of Computer Education
• /
• v.10 no.3
• /
• pp.67-73
• /
• 2007

In this paper, we review essential constituents of a dialogue interface system and propose practical methods to implement the each constituent. The implemented system consists of a discourse manager, an intention analyzer, a named entity recognizer, a SQL query generator, and a response generator. In the progress of implementation, the intention analyzer uses a maximum entropy model based on statistics because the domain dependency of the intention analyzer is comparatively low. The others use a simple pattern matching method because they needs high domain portability. In the experiments in a schedule arrangement domain, the implemented system showed the precision of 88.1% in intention analysis and the success rate of 83,4% in SQL query generation.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.27 no.3
• /
• pp.389-397
• /
• 2003

Effect of tip leakage flow on through flow and performance of a centrifugal compressor impeller was numerically studied using CFX-TASC flow. Seven different tip clearances were used to consider the influence of tip clearance on performance. Secondary flow and loss factor were evaluated to understand the loss mechanism inside the impeller due to tip leakage flow. The calculated results were circumferentially averaged along the passage and at the impeller exit for quantitative discussion. Tip clearance effect on Performance could be decomposed into inviscid and viscous components using one dimensional equation. The inviscid component is related with the specific work reduction and the viscous component is related with the additional entropy generation. Two components affected Performance equally. while efficiency drop was mainly influenced by viscous loss. Performance and efficiency drop due to tip clearance were proportional to the ratio of tip clearance to exit blade height. A simple model suggested in the present study predict performance and efficiency drop quite successfully.