• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.368-371
• /
• 1999

In this paper, we propose a neural network processor architecture with on-chip learning and with reconfigurability according to the data dependencies of the algorithm applied. For the neural network model applied, the proposed architecture can be configured into either SIMD or SRA(Systolic Ring Array) without my changing of on-chip configuration so as to obtain a high throughput. However, changing of system configuration can be controlled by user program. To process activation function, which needs amount of cycles to get its value, we design it by using PWL(Piece-Wise Linear) function approximation method. This unit has only single latency and the processing ability of non-linear function such as sigmoid gaussian function etc. And we verified the processing mechanism with EBP(Error Back-Propagation) model.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.2
• /
• pp.45-52
• /
• 2004

The sliding mode controller with a boundary layer implemented by simplified indirect inference method (SIIM) fuzzy logic was proposed. The components of the sliding line function are used for the inputs of the SIIM fuzzy logic. The proposed control system is simple because there is no need to derive the sigmoid function and there are only four rules. The overall stability of the proposed system and the boundness of the tracking error are proved easily using the Lyapunov theory. We apply the proposed controller to control a nonlinear time-varying system. The computer simulation showed the validity of the proposed control system.

• Computers and Concrete
• /
• v.26 no.5
• /
• pp.411-420
• /
• 2020

Deep Convolution neural network (DCNN) has been widely used in the healthy maintenance of civil infrastructure. Using DCNN to improve crack detection performance has attracted many researchers' attention. In this paper, a light-weight spatial attention network module is proposed to strengthen the representation capability of ResNet and improve the crack detection performance. It utilizes attention mechanism to strengthen the interested objects in global receptive field of ResNet convolution layers. Global average spatial information over all channels are used to construct an attention scalar. The scalar is combined with adaptive weighted sigmoid function to activate the output of each channel's feature maps. Salient objects in feature maps are refined by the attention scalar. The proposed spatial attention module is stacked in ResNet50 to detect crack. Experiments results show that the proposed module can got significant performance improvement in crack detection.

• The Journal of Korean Medicine
• /
• v.22 no.3
• /
• pp.179-188
• /
• 2001

Diarrhea is the frequent passage of loose, watery stool (frequency: ${\geq}4/day$, weight: ${\geq}250g/day$) Most antibiotics can cause inflammatory change of the colon or Pseudomembranous colitis (PMC). Typical presentations of PMC are watery diarrhea, abdominal pain, fever, leukocytosis ($12,000~20,000/\textrm{mm}^3$), hypoalbuminemia, hypovolemia and recent or concurrent use of antibiotics. Diagnostic methods of PMC are stool assay, sigmoid scopy, abdominal CT, abdominal US, etc. The age-related susceptibility noted with PMC is impressive but unexplained. Two stroke patients had diarrhea, abdominal pain, fever hypoalbuminemia and a history of recent or concurrent use of antibiotics. By use of Shirhyung- Tang, we could improve clinical symptoms (diarrhea, abdominal pain, fever hypoalbuminemia, etc.) and so report clinical course of two stroke patients with antibiotics-associated PMC.

• Journal of Korean Neurosurgical Society
• /
• v.43 no.5
• /
• pp.250-252
• /
• 2008

Meningiomas frequently invade cerebral venus sinuses, especially parasagittal meningioma to superior sagittal sinus. However, most invasions do not reach internal jugular vein. We present a case of parasagittal meningioma extending into the internal jugular vein through the sinuses. Radiological investigation revealed that the tumor was invading the sagittal, transverse, sigmoid sinus and junction of the internal jugular vein to subclavian vein, which was filled with tumor. The histopathological examinations revealed that both the cerebral tumor and mass in the internal jugular vein contributed to the transitional meningioma. This is a rare case of a meningioma extending into the internal jugular vein through the sinuses. According to this case, the frontal parasagittal meningioma could invade directly the internal jugular vein. The significance of this association to cerebral venus sinuses and internal jugular vein are discussed.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2001.05a
• /
• pp.456-459
• /
• 2001

In this paper, a wavelet neural network for adaptive control was proposed. The structure of this network is similar to that of the multilayer perceptron(MLP), except that here the sigmoid functions are replated by mother wavelet function in the hidden units. The simulation result showed the effectiveness of using the wavelet neural network structure in the adaptive control of one-link manipulator.

• Speech Sciences
• /
• v.15 no.4
• /
• pp.85-96
• /
• 2008

Automatic speech recognition has severe performance degradation under noisy environments. To cope with the noise problem, many methods have been proposed. Most of them focused on noise-robust features or model adaptation. However, researchers have overlooked utterance verification (UV) under noisy environments. In this paper we discuss UV problems based on the normalized confidence measure. First, we show that UV performance is also degraded in noisy environments with the experiments of an isolated word recognition. Then we observe how the degradation of UV performances is suffered. Based on the UV experiments we propose a modeling method of the statistics of phone confidences using sigmoid functions. For obtaining the parameters of the sigmoidal models, the particle swarm optimization (PSO) is adopted. The proposed method improves 20% rejection performance. Our experimental results show that the PSO-NCM can apply noise speech recognition successfully.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.518-518
• /
• 2000

We have researched quantitative structure activity relationships between molecular structure of medicines and physiological activity. Since they are non-linear, neural networks are useful tool to research them. There are many ranks for the non-linearity; therefore, the neuron function must be selected carefully. As the results of some trial calculations, Ire find the sigmoid-linear functions pair. We call the neural network constructed of the pair as ANN. The inter- or extrapolation abilities of the ANN are excellent; therefere, ANN is a superior predictor for the relationships. We evaluated the anticarcinogenic medicines, Carboquinone derivatives, by the developed ANN and leave-one-out method.

• The Korean Journal of Applied Statistics
• /
• v.25 no.4
• /
• pp.633-640
• /
• 2012

Biological methods are described for the assay of certain substances and preparations whose potency cannot be adequately assured by chemical or physical analysis. The principle applied through these assays is of a comparison with a standard preparation to determine how much of the examined substance produces the same biological effects as a given quantity (the Unit) of the standard preparation. In these dilution assays, to estimate the relative potencies of the unknown preparations to the standard preparations, it is necessary to compare dose-response relationships of standard and unknown preparations. The dose-response relationship in the dilution assay is non-linear and sigmoid when a wide range of doses is applied. The parallel line model (applied to the dose region with the steepest slope) is used to estimate the relative potency. In this paper, the statistical theory in the parallel line model is explained with an application to a dilution assay data. The parallel line method is implemented in a SAS program and is available at the author's homepage(http://cafe.daum.net/go.analysis).

• Journal of the Korean Society of Tobacco Science
• /
• v.3 no.2
• /
• pp.109-114
• /
• 1981

This experiment was conducted with three varieties (Hicks, Burley 21, Sohyang) and cropping systems (Improved mulching, Mulching, Non mulching) of NC 2326 to analyze growth kinetics by means of growth function involving its velocity and accelerated velocity. The basic growth data were obtained by harvest method at interval of ten days from transplanting to hundred days and analyzed by , regression equation, determinant of matrix, and differentiation. The plot of total dry weight of leaves, stalk and roots per a plant vs. time forms a sigmoid curve and its function fitted logistic satisfactorily. Tobacco plant grows at an accelerated velocity. And growth velocity, symmetric about an inflection point, is proportional to biomass attained and to the difference between biomass attained and the maximum, and to the decrease according to the biomass. Of varieties and cropping systems, the most maximum velocity was 9.58g per day per plant in mulching cultivation of NC 2326 and maximum accelerated velocity was 264mg per $day^2$ per plant in Burley 21.