• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.56-59
• /
• 2017

CNN(Convolutional Neural Network)을 이용하여 발화 주제 다중 분류 task를 multi-labeling 방법과, cluster 방법을 이용하여 수행하고, 각 방법론에 MSE(Mean Square Error), softmax cross-entropy, sigmoid cross-entropy를 적용하여 성능을 평가하였다. Network는 음절 단위로 tokenize하고, 품사정보를 각 token의 추가한 sequence와, Naver DB를 통하여 얻은 named entity 정보를 입력으로 사용한다. 실험결과 cluster 방법으로 문제를 변형하고, sigmoid를 output layer의 activation function으로 사용하고 cross entropy cost function을 이용하여 network를 학습시켰을 때 F1 0.9873으로 가장 좋은 성능을 보였다.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.3
• /
• pp.115-122
• /
• 1994

The problems in back-propagation(BP) generally are learning speed and misclassification due to lacal minimum. In this paper, to solve these problems, the classical modified methods of BP are reviewed and an extension of the BP to compensate the sigmoide-prime function around the extremity where the actual output of a unit is close to zero or one is proposed. The proposed method is not onlu faster than the conventional methods in learning speed but has an advantage of setting variables easily because it shows good classification results over the vast and uncharted space about the variations of learning rate, etc.. And it is simple for hardware implementation.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.1
• /
• pp.145-154
• /
• 2000

In this paper we proposed the orthogonal neural network(ONN) to control and identify the unknown controlled system. The proposed ONN used the buffer layer in front of the hidden layer and the hidden layer used the sigmoid function and its derivative a derived RBF that is a derivative of the sigmoid function. In order to verify the property of the proposed, it is examined by simulation results of the Narendra model. Controlled system is composed of ONN and confirmed its usefulness through simulation and experimental results.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.16 no.6
• /
• pp.285-292
• /
• 2021

Autonomous vehicles are divided into an upper controller that calculates control value through cognitive judgment and a lower controller that appropriately transmits its control value to an actuator. Here, the longitudinal control in a lower controller has a problem as the road slopes due to the property of the Acceleration sensor to output the acceleration as the slope of the device. Therefore, in this paper, a sigmoid function is proposed to determine the slope to compensate for this problem. Through the experiment, Checked performance by comparing the existing table model with the proposed model.

• International Journal of Control, Automation, and Systems
• /
• v.6 no.6
• /
• pp.818-827
• /
• 2008

In this paper, we propose a new noise estimation and reduction algorithm for stationary and nonstationary noisy environments. This approach uses an algorithm that classifies the speech and noise signal contributions in time-frequency bins. It relies on the ratio of the normalized standard deviation of the noisy power spectrum in time-frequency bins to its average. If the ratio is greater than an adaptive estimator, speech is considered to be present. The propose method uses an auto control parameter for an adaptive estimator to work well in highly nonstationary noisy environments. The auto control parameter is controlled by a linear function using a posteriori signal to noise ratio(SNR) according to the increase or the decrease of the noise level. The estimated clean speech power spectrum is obtained by a modified gain function and the updated noisy power spectrum of the time-frequency bin. This new algorithm has the advantages of much more simplicity and light computational load for estimating the stationary and nonstationary noise environments. The proposed algorithm is superior to conventional methods. To evaluate the algorithm's performance, we test it using the NOIZEUS database, and use the segment signal-to-noise ratio(SNR) and ITU-T P.835 as evaluation criteria.

• Pediatric Gastroenterology, Hepatology & Nutrition
• /
• v.23 no.2
• /
• pp.137-145
• /
• 2020

Purpose: The Oral-anal Transit Test (OTT) is a simple method of obtaining information about colonic transit. We aim to assess the correlation of OTT with the neuromuscular integrity of the colon determined by colonic manometry (CM). Methods: All patients who had OTT followed by CM were evaluated. Less than 6 of 24 markers remaining on OTT was considered normal. CM was performed per previously published guidelines. A normal CM was defined as at least one High Amplitude Propagating Contraction progressing from the most proximal sensor through the sigmoid colon. Results: A total of 34 patients underwent both OTT and CM (44% male, age 4-18 years, mean 11.5 years, 97% functional constipation +/- soiling, Hirschsprung's Disease). Of normal and abnormal OTT patients, 85.7% (6/7) and 18.5% (5/27) respectively had normal CM. When all markers progressed to at least the sigmoid colon, this was 100% predictive against colonic inertia. Greater than 50% of patients with manometric isolated sigmoid dysfunction had markers proximal to the recto-sigmoid. Conclusion: OTT and CM are both valuable studies that assess different aspects of colonic function. OTT can be used as a screening test to rule out colonic inertia. However, the most proximal extent of remaining markers does not predict the anatomical extent of the manometric abnormality, particularly in isolated sigmoid dysfunction.

• The Journal of Engineering Geology
• /
• v.33 no.1
• /
• pp.137-149
• /
• 2023

In order to intuitively grasp the earth pressure direction acting on the structure and displacement state, displacement data in the horizontal and vertical directions were processed using the sigmoid function. A displacement coordinate system was set up for each axis. The system can intuitively check the current displacement and assess the management stage of each point. A displacement path can be compiled from continuously recorded points, allowing trends in the displacement's history and stress direction to be known. Analysis of data measured for excavated ground, found that displacement occurred in the direction of destressing at all points, and that the points' management state steady. Similar behavior trends were found among measurement points with high spatial correlation, whereas differing behavior trends occurred among measurement points with low spatial correlation. If the correlation analysis of the precursor and behavior area is performed using the continuously distributed surface settlement data and displacement coordinate system, it will be possible to predict the failure time and area.

• Journal of Korean Neurosurgical Society
• /
• v.60 no.2
• /
• pp.165-173
• /
• 2017

Objective : To explore and analyze the influencing factors of facial nerve function retainment after microsurgery resection of acoustic neurinoma. Methods : Retrospective analysis of our hospital 105 acoustic neuroma cases from October, 2006 to January 2012, in the group all patients were treated with suboccipital sigmoid sinus approach to acoustic neuroma microsurgery resection. We adopted researching individual patient data, outpatient review and telephone followed up and the House-Brackmann grading system to evaluate and analyze the facial nerve function. Results : Among 105 patients in this study group, complete surgical resection rate was 80.9% (85/105), subtotal resection rate was 14.3% (15/105), and partial resection rate 4.8% (5/105). The rate of facial nerve retainment on neuroanatomy was 95.3% (100/105) and the mortality rate was 2.1% (2/105). Facial nerve function when the patient is discharged from the hospital, also known as immediate facial nerve function which was graded in House-Brackmann : excellent facial nerve function (House-Brackmann I-II level) cases accounted for 75.2% (79/105), facial nerve function III-IV level cases accounted for 22.9% (24/105), and V-VI cases accounted for 1.9% (2/105). Patients were followed up for more than one year, with excellent facial nerve function retention rate (H-B I-II level) was 74.4% (58/78). Conclusion : Acoustic neuroma patients after surgery, the long-term (${\geq}1year$) facial nerve function excellent retaining rate was closely related with surgical proficiency, post-operative immediate facial nerve function, diameter of tumor and whether to use electrophysiological monitoring techniques; while there was no significant correlation with the patient's age, surgical approach, whether to stripping the internal auditory canal, whether there was cystic degeneration, tumor recurrence, whether to merge with obstructive hydrocephalus and the length of the duration of symptoms.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.7 no.4
• /
• pp.74-81
• /
• 1997

This paper proposes an evolutionary learning algorithm to discipline the projection neural nctworks (PNNs) with special type of hidden nodes which can activate radial basis functions as well as sigmoid functions. The proposed algorithm not only trains the parameters and the connection weights hut also c~ptimizes the network structure. Through the structure optimization, the number of hidden node:; necessary to represent a given target function is determined and the role of each hidden node is decided whether it activates a radial basis function or a sigmoid function. To apply the algorithm, PNN is realized by a self-organizing genotype representation with a linked list data structure. Simulations show that the algorithm can build the PNN with less hidden nodes than thc existing learning algorithm using error hack propagation(EE3P) and network growing strategy.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.3
• /
• pp.357-362
• /
• 2005

Iris recognition system is the one of the most reliable biometries recognition system. An algorithm is proposed to determine the localized iris from the iris image received from iris input camera in client. For the first step, the algorithm determines the center of pupil. For the second step, the algorithm determines the outer boundary of the iris and the pupillary boundary. The localized iris area is transformed into polar coordinates. After performing three times Wavelet transformation, normalization was done using a sigmoid function. The converting binary process performs normalized value of pixel from 0 to 255 to be binary value, and then the converting binary process is compared pairs of two adjacent pixels. The binary code of the iris is transmitted to the server by the network. In the server, the comparing process compares the binary value of presented iris to the reference value in the database. The process of recognition or rejection is dependent on the value of Hamming Distance. After matching the binary value of presented iris with the database stored in the server, the result is transmitted to the client.