http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Machine Learning Directed Prediction of Saturation Magnetization
Chunghee Nam 한국자기학회 2021 한국자기학회 학술연구발표회 논문개요집 Vol.31 No.2
Research to predict the various physical properties of inorganic materials using material data has been actively conducted in recent years. Among them, magnetic properties are important features for functional material applications and have been predicted, such as the Curie temperature, coercivity and magnetocaloric properties. In this study, saturation magnetization value was predicted using the DFT quantum calculation-based material data (open source) via machine learning. Ensemble algorithms have been used to compare the performance of models with two metrics of R² (coefficient of determination) and RMSE (root-mean-squared error), as shown in Fig. 1. Detailed results will be presented. 〈그림 본문참조〉
Nam, Chunghee,Jang, Youngman,Lee, Ki-Su,Shim, Jungjin,Cho, B. K. American Scientific Publishers 2006 Journal of Nanoscience and Nanotechnology Vol.6 No.11
<P>We have studied the influence of the insertion of a nano-oxide layer (NOL) into a magnetic GMR spin-valve. It was found that the spin-valve with NOL has a higher GMR ratio than that of the normal spin-valve without NOL. Naturally formed NOL without vacuum break shows a uniform layer, which effectively suppresses the current shunt, resulting in the reduction of the sheet resistance of GMR. The NOL spin-valve also shows a lower interlayer coupling (<I>H</I>in) than that of the optimal normal spin-valve, which is consistent with AFM measurement showing lower roughness of NOL formed CoFe surface. Based on the advantage of NOL, we succeeded in lowering <I>H</I>in while maintaining GMR ratio by insertion of NOL inside the CoFe free layer, where the free layer consists of CoFe/NOL/CoFe/NOL/Capping layer.</P>
Machine learning based Human Activity Recognition with mobile 3-axis magnetometers
Chunghee Nam 한국자기학회 2021 한국자기학회 학술연구발표회 논문개요집 Vol.31 No.1
Recently, human activity recognition (HAR) plays an important role in well-being life and context-aware IoT systems. HAR can be carried out in real-time by using sensory data collected from embedded sensor networks in mobile smart phones. Recent HAR investigations have shown that is solely based on 3-axis accelerometers, which is the most energy-efficient approach. In this presentation, I propose a simple approach for HAR process with built-in 3-axis magnetometers in a smart phone. Based on deep learning with convolutional neural network(CNN), I found 98% accuracy for 4 -classes (standing, sitting, jogging, and walking).
Characteristics of domain wall pinning and depinning in a three-terminal magnetic Y-junction
Nam, Chunghee,Jang, Youngman,Lee, Ki-Su,Cho, B K IOP Pub 2008 Nanotechnology Vol.19 No.1
<P>The characteristics of domain wall (DW) pinning and propagation in a three-terminal magnetic Y-junction were investigated, where the junction consisted of two input and one output wires. The output switching depends strongly on the junction angle (α). Junctions with high angles of α>9.5° lead to DW pinning at the junction, whereas junctions with low angles of α<9.5° have no DW pinning effect. At the critical angle of α = 9.5°, the Y-junction showed a multimode DW propagation, which was ascribed to a moderate transverse field effect.</P>
Magnetostatic interaction between magnetic domain walls in dual Co rings
Chunghee Nam 한국물리학회 2016 Current Applied Physics Vol.16 No.7
Magnetostatic interaction of magnetic domain walls (DWs) is investigated by using magnetic force microscopy (MFM) at the remanent state in closely placed Co dual rings. The Co dual rings are 10 and 20 nm in thickness and the spacing between them is varied. In an array of dual rings varying in steps of 10 against the direction of an applied field, the angular dependence of DW interaction shows an obvious change from coupling to decoupling of the DWs in the MFM measurements. It is found that the strong interaction between DWs at smaller angles and smaller spacing is owed to the surface magnetic charge attraction. On the other hand, monopole-like magnetic charge attraction is related with weak DW interaction at higher angles and a larger spacing. In addition, the dependence of the Co thickness on the magnetostatic DW interaction can be explained by the volume magnetic effects.
머신러닝 기반 모바일 스마트폰 자기센서를 이용한 인간행동인식
남충희(Chunghee Nam) 한국자기학회 2021 韓國磁氣學會誌 Vol.31 No.4
As the performance of sensors embedded in mobile smart phones has improved, many studies using data collected from sensors are being conducted. In this study, using the data obtained from the 3-axis magnetic sensor mounted on the smartphone, a study on the recognition of four human activities was performed using machine learning. From the total data of the 3-axis magnetic sensor, the data was bundled into frames for 2 seconds, divided into several frames, and then supervised learning was carried out using it as an input to the convolutional neural network. The operation of the magnetic sensor depending on the direction was confirmed, and the human activity recognition for standing, sitting, walking, and jogging was verified.
Byung-Nam Cho,Jung Mi Ahn,Hoi Kyung Jung,Chunghee Cho,최돈찬,Kelly E Mayo 한국분자세포생물학회 2004 Molecules and cells Vol.18 No.1
Inhibin is a gonadal hormone composed of an α- subunit and one of two β-subunits (βA, βB), and its primary role is to inhibit FSH secretion by the pituitary. To investigate the roles of inhibin α in the reproductive system, an expression plasmid, pCMV-rINA, with the rat inhibin α cDNA fused to the cytomegalovirus promoter, was introduced into muscle by direct injection. Inhibin α mRNA was detected in the muscle by RT-PCR and Southern blot analysis. Inhibin protein was also detected, and Western blot analysis revealed a relatively high level of serum inhibin, but not of activin βA. The estrous cycle of the pCMVrINA- injected mice was extended, but there was no change in levels of pituitary FSH mRNA or serum FSH and no ovarian cysts were observed. When injected female mice were mated with males of proven fertility, litter size increased. Surprisingly, the embryos of pregnant females injected with pCMV-rINA, were retarded in growth and had defects in internal organs. When male mice were injected, testicle weight increased slightly without any noticeable change in the histology of the seminiferous tubules. Taken together, our data indicate that the inhibin α subunit influences a number of the reproductive functions of female mice.