• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2022년도 제66차 하계학술대회논문집 30권2호
• /
• pp.63-66
• /
• 2022

본 논문에서는 분자의 특성을 정확하게 예측하기 위해 효과적인 사전학습(pretraining) 전략과 트랜스포머(Transformer) 모델을 활용한 방법을 제시한다. 딥러닝을 활용한 분자의 성능을 예측하는 연구는 그동안 레이블이 부족한 분자데이터의 특성에 의해 학습 때 사용된 데이터이외의 분자데이터에 대해 일반화 능력이 떨어지는 어려움을 겪었다. 이 논문에서 제시한 모델은 사전학습(pretraining)을 수행할 때 자기지도학습(self-supervised training)을 사용하여 부족한 레이블에 의한 문제점을 피할 수 있다. 대규모 분자 데이터셋으로부터 학습된 이 모델은 4가지 다운스트림 데이터셋에 대해 모두 우수한 성능을 보여주어 일반화 성능이 뛰어나며 효과적인 분자표현을 얻을 수 있음을 보인다.

• 반도체디스플레이기술학회지
• /
• 제8권4호
• /
• pp.21-24
• /
• 2009

We investigated a nanoscale inertia sensor based on telescoping carbon nanotubes, using classical molecular dynamics simulations. The position of the telescoping nanotubes is controlled by the centrifugal force exerted by the rotation platform, thus, position shifts are determined by the capacitance between carbon nanotubes and the electrode, and the operating frequency of the carbon nanotube oscillator. This measurement system, tracking oscillations of the carbon nanotube oscillator, can be used as the sensor for numerous types of devices, such as motion detectors, accelerometers and acoustic sensors.

• 반도체디스플레이기술학회지
• /
• 제6권1호
• /
• pp.27-30
• /
• 2007

The gigahertz oscillator behavior of double-walled boron-nitride nanotube (BNNT) was investigated by using classical molecular dynamics simulations. The BNNT oscillator characteristics were compared to carbon-nanotube (CNT) and hybrid-C@BNNT oscillators. The results show that the BNNT oscillators are higher than the van der Waals force of the CNT oscillator. Since the frictional effects of BNNT oscillators are higher than that of a CNT oscillator, the damping factors of BNNT and hybrid oscillators are higher than that of a CNT oscillator.

• 정보과학회지
• /
• 제23권5호
• /
• pp.41-56
• /
• 2005

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 하계학술대회 논문집 C
• /
• pp.2102-2104
• /
• 2004

In this study, sheathless electrospray from PDMS/glass microchips with conducting metal emitter tip is described. A chip-based capillary electrophoresis/mass spectrometry (CE/MS) system has advantages of the CE separation and on-line electrospray detection of peptide solution. We have fabricated a new electrospray ionization(ESI) device composed of the metal emitter tip and CE separation channel monolithically in a glass microchip. The separation channel and metal emitter tip are fabricated using a glass wet etching and gold electro plating process, respectively. The fabricated micro electrospray chip was tested by spraying peptide sample for mass spectrometric analysis. Singlely-charged peak and doublely-charged peak of peptide were detected and further MS/MS fragmentation was performed in each peak. Direct comparisons with conventional glass or fused silica emitters showed very similar performance with respect to signal strength and stability.

• Biotechnology and Bioprocess Engineering:BBE
• /
• 제10권3호
• /
• pp.212-217
• /
• 2005

Desorption/ionization on silicon mass spectrometry (DIOS-MS) is a relatively new laser desorption/ionization technique for mass spectrometry without employing an organic matrix. This present study was carried to survey the experimental factors to improve the efficiency of DIOS-MS through electrochemical etching condition in structure and morphological properties of the porous silicon. The porous structure of silicon structure and its properties are crucial for the better performance of DIOS-MS and they can be controlled by the suitable selection of electrochemical conditions. The fabrication of porous silicon and ion signals on DIOS-MS were examined as a function of silicon orientation, etching time, etchant, current flux, irradiation, pore size, and pore depth. We have also examined the effect of pre- and post-etching conditions for their effect on DIOS-MS. Finally, we could optimize the electrochemical conditions for the efficient performance of DIOS-MS in the analysis of small molecule such as amino acid, drug and peptides without any unknown noise or fragmentation.

• The Plant Pathology Journal
• /
• 제33권6호
• /
• pp.529-542
• /
• 2017

Control of plant diseases is largely dependent on use of agrochemicals. However, there are widening gaps between our knowledge on plant diseases gained from genetic/mechanistic studies and rapid translation of the knowledge into target-oriented development of effective agrochemicals. Here we propose that the time is ripe for computer-aided drug discovery/design (CADD) in molecular plant pathology. CADD has played a pivotal role in development of medically important molecules over the last three decades. Now, explosive increase in information on genome sequences and three dimensional structures of biological molecules, in combination with advances in computational and informational technologies, opens up exciting possibilities for application of CADD in discovery and development of agrochemicals. In this review, we outline two categories of the drug discovery strategies: structure- and ligand-based CADD, and relevant computational approaches that are being employed in modern drug discovery. In order to help readers to dive into CADD, we explain concepts of homology modelling, molecular docking, virtual screening, and de novo ligand design in structure-based CADD, and pharmacophore modelling, ligand-based virtual screening, quantitative structure activity relationship modelling and de novo ligand design for ligand-based CADD. We also provide the important resources available to carry out CADD. Finally, we present a case study showing how CADD approach can be implemented in reality for identification of potent chemical compounds against the important plant pathogens, Pseudomonas syringae and Colletotrichum gloeosporioides.

• 미생물학회지
• /
• 제35권2호
• /
• pp.115-120
• /
• 1999

3개 이상의 DNA 혹은 단백질의 염기서열을 정렬하는 복수 염기서열 정렬은 염기서열들 사이의 진화관계, gene regulation, 단백질의 구조와 기능에 관한 연구에 필수적인 도구이다. 복수 염기서열 정렬을 얻기 위한 기존의 방법은 progressive pairwise alignment 와같이 빠른 실행시간 내에 만족할 만한 복수 염기서열 정렬을 제공하는 방법과, 최적의 복수 여기서열 정렬을 제공하나 실행시간이 상대적으로 긴 dynamic programming 과 같은 방법 등이 있다. 본 논문에서는 진화 알고리즘을 사용하여 기존의 방법에서 제공하는 복수 염기서열 정렬을 짧은 시간내에보다 개선된 복수 염기서열 정렬을 획득하게 하는 방법을 제시하였으며, 진화 알고리즘의 구성내용을 설명하였으며, 실제의 염기서열을 사용하여 이 방법의 장점을 보였다.

• 한국진공학회지
• /
• 제6권S1호
• /
• pp.70-74
• /
• 1997

The study of the thickness dependence of the electron energy structure of Fe, Ni, Ti and Zr sublayers in Ni/Ti and Fe/Zr multilayers by using the experimental and computer simulated optical spectroscopy has been performed. A series of Ni/Ti and Fe/Ze multiayered films (MLF) with a bilayer period of 0.5 - 30 nm and constant (Ni/Ti) / different (Fe/Zr) sublayer thickness ratios were prepared by using computer-controlled double-pair target face-to-face sputtering onto a glass substrate at room temperature (RT) Computer simulation of the resulting optical properties of these MLF was carried out by solving of multireflection problem with a matrix method assuming either "sharp" interfaces resulting in rectangular depth profiles of the components or "mixed" (alloy-like) interfaces of variable thickness between pure-metal sublayers. Optical constants of pure bulk metals as well as equiatomic alloy interfaces were employed in these simulations. It was shown that the difference between experimental and simulated optical properties of the investigated MLF increases with decrease in sublayer thickness. This result allows to conclude that the electronic structures of sublayers below 4-5 nm thickness in mlf differ from the corresponding bulk metals.ponding bulk metals.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2020년도 추계학술발표대회
• /
• pp.1036-1038
• /
• 2020

This paper proposes a novel automated liver segmentation using Multi-Dilated U-Nets. The proposed multidilation segmentation model has the advantage of considering both local and global shapes of the liver image. We use the CT images subject-wise, every 2D image is concatenated to 3D to calculate the IOU score and DICE score. The experimental results on Jeonbuk National University hospital dataset achieves better performance than the conventional U-Net.