Development of accurate virtual engines and effective ranking systems (AVENGERS) for drug design and protein engineering

임호철 Graduate School, Yonsei University 2023 국내박사

This dissertation consists of six chapters: Chapter 1. Development of accurate virtual engines and effective ranking systems (AVENGERS) for drug design and protein engineering, Chapter 2. Hot spot profiles of SARS-CoV-2 and human ACE2 receptor protein-protein interaction obtained by density functional tight binding fragment molecular orbital method [1], Chapter 3. Assessment of the ANI-2X deep learning neural network potentials for ranking the docking poses in molecular docking simulations, Chapter 4. Identification of novel natural product inhibitors against matrix metalloproteinase 9 using the quantum mechanical fragment molecular orbital-based virtual screening methods [2], Chapter 5. Quantum computational study of chloride ion attack on chloromethane for chemical accuracy and quantum noise effects with UCCSD and k-UpCCGSD ansatzes [3], and Chapter 6. Evaluation of protein descriptors in computer-aided rational protein engineering tasks and its application in property prediction in SARS-CoV-2 spike glycoprotein [4]. Drug discovery is an initial step in process of identification of novel drug candidates for their therapeutic target. The global drug discovery market is predicted to reach a compound annual growth rate (CAGR) of nearly 8.21% during the period (2019 ~ 2027). Whereas drug discovery comes to be progressively more expensive and costs on average USD 2.6 billion with about 13 years to develop a single drug. To reduce the progressively expensive cost, computer-aided drug design (CADD) can be one of the effective and less costly ways for drug discovery. Quantum chemistry and artificial intelligence in many CADD methods developed over the past decades have been in the spotlight because quantum mechanics (QM) can ab initio provide the most accurate descriptions and machine learning (ML) can effectively provide the predictions from training data. Therefore, accurate scoring functions and ranking systems with QM and/or ML would be powerful in drug discovery. The prevalence of a novel β-coronavirus (SARS-CoV-2) was declared a public health emergency of international concern on 30 January 2020 and a global pandemic on 11 March 2020 by WHO. The spike glycoprotein of SARS-CoV-2 is regarded as a key target for the development of vaccines and therapeutic antibodies. To develop antiviral therapeutics for SARS-CoV-2, it is crucial to find amino acid pairs that strongly attract each other at the interface of the spike glycoprotein and the human angiotensin-converting enzyme 2 (hACE2) complex. To find hot spot residues, the strongly attracting amino acid pairs at the protein-protein interaction (PPI) interface, we introduce a reliable inter-residue interaction energy calculation method, FMO-DFTB3/D/PCM/3D-SPIEs. In addition to the SARS-CoV-2 spike glycoprotein/hACE2 complex, the hot spot residues of the SARS-CoV-1 spike glycoprotein/hACE2 complex, SARS-CoV-1 spike glycoprotein/antibody complex, and HCoV-NL63 spike glycoprotein/hACE2 complex were obtained using the same FMO method. Following this, a 3D-SPIEs-based interaction map was constructed with hot spot residues for the hACE2/SARS-CoV-1 spike glycoprotein, hACE2/HCoV-NL63 spike glycoprotein, and hACE2/SARS-CoV-2 spike glycoprotein complexes. Finally, the three 3D-SPIEs-based interaction maps were combined and analyzed to find the consensus hot spots among the three complexes. As a result of the analysis, two hot spots were identified between hACE2 and the three spike proteins. In particular, E37, K353, G354, and D355 of the hACE2 receptor strongly interact with the spike proteins of coronaviruses. The 3D-SPIEs-based map would provide valuable information to develop anti-viral therapeutics that inhibit PPIs between the spike protein of SARS-CoV-2 and hACE2. Computational molecular simulations have provided useful insight into understanding the structures of protein-ligand complexes with the balance between accuracy and cost. Molecular docking requires a balance because it is generally used in virtual screening methods by selecting the efficient subset from large virtual compound libraries. Machine learning-based neural network potentials (NNP) have been considered as an alternative to achieve balance with as high accuracy as quantum mechanical methods and as fast speed as force fields. In this work, we demonstrate the applicability of the NNP (ANI-2x) to molecular docking simulations, especially to ranking docking poses. Although the ANI-2x was not trained to describe protein structures and ionic molecules, it can be used for scoring docking poses even in not-zero net charge protein-ligand systems. The pose ranking method by ANI-2x can be incorporated directly into existing molecular docking methods to find the most favorable docking poses. As the development of new and improved NNP continues, the versatile approach by NNP would enable accurate free energy estimation in molecular simulations. Matrix metalloproteinases (MMPs) are calcium-dependent zinc-containing endopeptidases involved in multiple cellular processes. Among the MMP isoforms, MMP-9 regulates cancer invasion, rheumatoid arthritis, and osteoarthritis by degrading extracellular matrix proteins present in the tumor microenvironment and cartilage and promoting angiogenesis. Here, we identified two potent natural product inhibitors of the non-catalytic hemopexin domain of MMP-9 using a novel quantum mechanical fragment molecular orbital (FMO)-based virtual screening workflow. The workflow integrates qualitative pharmacophore modeling, quantitative binding affinity prediction, and raw material search of natural product inhibitors with the BMDMS-NP library. In binding affinity prediction, we made a scoring function with the FMO method and applied the function to two protein targets (acetylcholinesterase and fibroblast growth factor 1 receptor) from DUD-E benchmark sets. In the two targets, the FMO method outperformed the Glide docking score and MM/PBSA methods. By applying this workflow to MMP-9, we proposed two potent natural product inhibitors (laetanine 9 and genkwanin 10) that interact with hotspot residues of the hemopexin domain of MMP-9. Laetanine 9 and genkwanin 10 bind to MMP-9 with a dissociation constant (KD) of 21.6 and 0.614 μM, respectively. Overall, we present laetanine 9 and genkwanin 10 for MMP-9 and demonstrate that the novel FMO-based workflow with a quantum mechanical approach is promising to discover potent natural product inhibitors of MMP-9, satisfying the pharmacophore model and good binding affinity. Quantum computing is expected to play an important role in solving the problem of huge computational costs in various applications by utilizing the collective properties of quantum states, including superposition, interference, and entanglement, to perform computations. Quantum mechanical (QM) methods are candidates for various applications and can provide accurate absolute energy calculations in structure-based methods. QM methods are powerful tools for describing reaction pathways and their potential energy surfaces (PES). In this study, we applied quantum computing to describe the PES of the bimolecular nucleophilic substitution (SN2) reaction between chloromethane and chloride ions. We performed noiseless and noise simulations using quantum algorithms and compared the accuracy and noise effects of the ansatzes. In noiseless simulations, the results from UCCSD and k-UpCCGSD are similar to those of full configurational interaction (FCI) with the same active space, which indicates that quantum algorithms can describe the PES of the SN2 reaction. In noise simulations, UCCSD is more susceptible to quantum noise than k-UpCCGSD. Therefore, k-UpCCGSD can serve as an alternative to UCCSD to reduce quantum noisy effects in the noisy intermediate-scale quantum era, and k-UpCCGSD is sufficient to describe the PES of the SN2 reaction in this work. The results showed the applicability of quantum computing to the SN2 reaction pathway and provided valuable information for structure-based molecular simulations with quantum computing. The importance of protein engineering in the research and development of biopharmaceuticals and biomaterials has increased. Machine learning in computer-aided protein engineering can markedly reduce the experimental effort in identifying optimal sequences that satisfy the desired properties from a large number of possible protein sequences. To develop general protein descriptors for computer-aided protein engineering tasks, we devised new protein descriptors, one sequence-based descriptor (PCgrades), and three structure-based descriptors (PCspairs, 3D-SPIEs_5.4Å, and 3D-SPIEs_8Å). While the PCgrades and PCspairs include general and statistical information in physicochemical properties in single and pairwise amino acids respectively, the 3D-SPIEs include specific and quantum-mechanical information with parameterized quantum mechanical calculations (FMO2-DFTB3/D/PCM). To evaluate the protein descriptors, we made prediction models with the new descriptors and previously developed descriptors for diverse protein datasets including protein expression and binding affinity change in SARS-CoV-2 spike glycoprotein. As a result, the newly devised descriptors showed a good performance in diverse datasets, in which the PCspairs showed the best performance (R^2=0.783 for protein expression and R^2=0.711 for binding affinity). As a result, the newly devised descriptors showed a good performance in diverse datasets, in which the PCspairs showed the best performance. Similar approaches with those descriptors would be promising and useful if the prediction models are trained with sufficient quantitative experimental data from high-throughput assays for industrial enzymes or protein drugs. In this work, accurate virtual engines and effective ranking systems (AVENGERS) were developed based on quantum mechanics and machine learning for drug design and protein engineering. The hotspot analysis study for SARS-CoV-2 highlights that the 3D-SPIEs-based interaction map from the QM results can be used to quantitatively analyze hotspot information in protein-protein interactions. The pose ranking study by ANI-2x highlights that NNP can be used to rank the docking poses to find the most favorable docking poses. The ligand ranking study by FMO highlights that FMO can be used for scoring functions to quantitatively predict binding affinities of ligands. Three case studies show that the QM methods enable accurate free energy estimation in molecular simulations in the hotspot analysis, pose ranking, and binding affinity prediction. The quantum computational SN2 reaction analysis study highlights that quantum computers can be applied to the QM methods and would dramatically accelerate the QM methods in the future. The evaluation study of protein descriptors for predicting diverse protein properties highlights that the newly devised protein descriptors can be used to make prediction models for diverse protein properties and show the best performance in the evaluation. As a result, AVENGERS are developed with the QM and ML methods for drug design and protein engineering and would accelerate drug discovery through accurate virtual screening by helping quantitatively predict diverse properties of small molecules and proteins with quantitative scoring models. 본 학위 논문은 6개의 장으로 이루어져 있다. 첫번째 장. 신약 개발과 단백질 개량을 위한 정확한 가상엔진과 효과적인 평가시스템(AVENGERS)의 개발; 두번째 장. 밀착 겹합 밀도 함수를 사용한 조각 단위 분자 궤도 함수를 통해 얻은 SAR-CoV-2와 안지오텐신 전환효소 2와의 단백질-단백질 상호작용의 핫스팟 프로필; 세번째 장. 분자도킹에서의 결합상태 평가를 위한 ANI-2X 딥러닝 신경망 포텐셜의 평가; 네번째 장. 양자역학 조각 단위 분자 궤도 함수 기반의 가상 탐색 방법을 이용한 기질금속단백질 분해효소 9의 새로운 천연물 저해제의 발견; 다섯번째 장. UCCSD와 k-UpCCGSD ansatz 을 이용한 화학적 정확성과 양자 잡음 효과 분석을 위한 클로로메테인에 대한 염화물의 공격에 대한 양자컴퓨터적인 연구; 여섯번째 장. 컴퓨터 기반의 합리적인 단백질 공학 문제에 있어서 단백질 표현자들의 평가와 SARS-CoV-2 스파이크 당단백질의 특성 예측에서의 적용. 신약개발은 치료제 표적에 대한 새로운 약물 후보군을 발견하는 과정에 있어서 초기 과정이다. 전 세계 신약개발 시장은 2019년에서 2027년까지 연평균 성장률이 8.21%일 것으로 예측된다. 반면, 신약개발은 점점 더 비싸지고 있고, 1개의 약물을 만들기 위해서 26억 달러의 비용이 들고 13년의 시간이 소요된다. 점점 더 비싸지는 신약개발 비용을 줄이기 위해서 컴퓨터를 활용한 신약개발은 효과적이고 가격이 저렴한 방법들 중 하나이다. 지난 10년 동안 개발되었던 많은 컴퓨터를 활용한 신약개발 방법들 중 양자역학과 기계학습은 가장 세간의 이목을 집중받았던 기술이며, 이는 양자역학은 가장 정확한 묘사력을 지녔고 기계학습은 효과적인 예측이 가능했기 때문이다. 따라서 양자역학과 기계학습을 이용하여 정확한 평가 시스템을 개발한다면 신약개발에 강력한 기능을 수행할 것으로 기대된다. 새로운 베타-코로나 바이러스인 SARS-CoV-2의 출현은 2020년 1월 30일에 국제적으로 우려된 공중보건 비상사태로 지정되었으며, WHO에 의해 2020년 3월 11일에 전세계 펜데믹으로 지정되었다. SARS-CoV-2의 스파이크 당단백질은 백신과 항체 치료제 개발에 있어서 중요한 표적으로 여겨진다. SARS-CoV-2을 위한 항바이러스 치료제를 개발하기 위해서 스파이크 당단백질와 사람 내의 안지오텐신 변환효소 2와의 접촉면에서 서로 강하게 끌어당기는 아미노산 쌍을 찾는 것이 중요하다. 단백질-단백질 접촉면에서 강하게 끌어당기는 아미노산 쌍인 핫스팟 잔기들을 찾기 위하여, 잔기들 간의 상호작용 에너지 계상 방법인 FMO-DFTB3/D/PCM/3D-SPIEs 을 도입하였다. SARS-CoV-2 스파이크 당단백질와 안지오텐신 변환효소 2 복합체뿐 아니라, SARS-CoV-1 스파이크 당단백질/안지오텐신-변환효소-2 복합체, SARS-CoV-1 스파이크 당단백질/항체 복합체, HCoV-NL63 스파이크 당단백질/안지오텐신-변환효소-2 복합체들에 대한 핫스판 잔기들도 분석하였다. 이다음에 3D-SPIEs 기반의 상호작용 지도를 만들어서 여러 복합체들에 대한 핫스팟 잔기들을 파악하였다. 결과적으로 이러한 상호작용 지도는 복학체들 간의 공통점을 찾는데 사용될 수 있다. 분석에 의하면, 안지오텐신-변환효소-2와 스파이크 당단백질은 2개의 핫스팟 영역을 가지고 있고, 안지오텐신-변환효소-2에 위치한 E37, K353, G354, D355 은 3가지 코로나 바이러스 유래 스파이크 당단백질과의 상호작용에 중요한 역할을 하였다. 3D-SPIEs 기반의 지도는 SARS-CoV-2 스파이크 당단백질과 안지오텐신-변환효소-2 와의 단백질-단백질 상호작용을 억제한 항바이러스 치료제를 개발하는데 중요한 정보를 제공할 것으로 기대된다. 컴퓨터를 활용한 분자 시뮬레이션은 계산 속도와 정확도와의 균형을 이루면서 단백질과 리간드 복합체의 구조를 이해하는데 중요한 정보를 제공해왔다. 분자도킹은 가상 화학물 라이브러리로부터 효과적인 부분집합을 선택하는 가상탐색에 주로 이용되기 때문에, 특히 계산 속도와 정확도의 균형이 매우 중요하다. 기계학습 기반의 신경망 포텐셜이 양자역학 기반의 방법들만큼 정확하고 분자역학 기반의 방법들만큼 빠르기 때문에 기존 방법의 대체제로 여겨지고 있다. 본 연구를 통해 신경망 포텐셜 중 하나인 ANI-2X를 사용하여 분자도킹 시뮬레이션에 적용하여 결합상태를 평가를 진행하였다. ANI-2X는 단백질과 전하를 가지는 분자를 묘사하도록 학습되지 않았음에도, 전하를 가지고 있는 단백질-리간드 시스템에서도 결합 상태를 평가하는데 사용될 수 있었다. 이러한 결합 상태를 평가하는 방법은 기존 분자도킹 방법에 결합되기 쉽고, 가장 적절한 결합 상태를 찾는데 활용될 수 있을 것이다. 신경망 포텐셜의 연구가 계속될수록, 신경망 포텐셜 방법을 활용한 접근은 분자 시뮬레이션에서 정확한 에너지 추정을 가능하게 할 것으로 기대된다. 기질 금속단백질분해효소은 칼슘-의존적이자 아연을 포함하는 펩타이드내부분해효소이고, 다양한 세포 신호전달체계에 참여하고 있다. 기질 금속단백질분해효소 이성질체 중에서 MMP-9 은 암의 미세환경 속에 존재하는 세포막 단백질과 연골을 분해하고 혈관싱생을 촉진함으로써 암전이, 류마티스 관절염, 그리고 골관절염을 조절한다. 여기서 우리는 새로운 양자역학 조각단위분자궤도함수(FMO)를 기반으로 가상 탐색 방법을 통해 MMP-9 의 비촉매 헤모펙신 도메인에 결합하는 2개의 천연물을 발굴하였다. 이 방법론은 정성적인 약물작용발생단 기법, 정량적인 결합력 예측, 그리고 천연물 저해제의 원료 추적방법을 통합하여 개발되었다. 결합력 예측에 있어서, 우리는 FMO 방법을 사용한 평가 함수를 만들었고, 이 함수를 DUD-E 벤치마킹에 있는 두 개의 단백질 표적 (아세틸콜린 에스테라제와 섬유 아세포 생장인자 수용체 1)에 적용하였다. 두 단백질 표적에서 FMO 방법은 Glide 분자도킹과 MM/PBSA 방법보다 우월한 성능을 보였다. 이 방법론을 MMP-9 에 적용하여 MMP-9 의 헤모펙신 도메인의 핫스팟 잔기와 상호작용하는 2개의 천연물 억제제 (laetanine 과 genkwanin)를 발굴하였다. Laetanine 과 genkwanin은 21.6 와 0.614 μM 의 결합력으로 MMP-9 에 결합하였다. 결과적으로는 본 연구를 통해 우리는 MMP-9 에 대한 laetanine 과 genkwanin 을 발굴하였고, 양자역학을 사용한 새로운 FMO 기반 방법론이 약물작용발생단과 좋은 결합력을 만족시키는 MMP-9 의 새로운 천연물 저해제를 발굴하는데 유망하다는 것을 보여주었다. 양자컴퓨팅은 양자 상태의 중첩, 간섭 및 얽힘 등과 같은 집합적인 속성을 활용하여 연산을 수행함으로써 다양한 응용 분야에서 막대한 연산 비용 문제를 해결하는 데 중요한 역할을 할 것으로 기대된다. 양자역학 기법은 양자컴퓨팅의 다양한 응용 분야의 하나이며, 구조 기반의 신약개발 방법에서 절대 에너지 계산을 제공할 수 있다. 양자역학 방법론은 또한 반응 경로와 포텐셜 에너지 표면을 기술하는데 강력한 도구이기도 하다. 본 연구에서 클로로메탄과 염화물 이온 사이의 이분자 친핵성 치환 반응 (SN2 반응)의 포텐셜 에너지 표면을 설명하기 위해 양자컴퓨팅을 적용하였다. 양자컴퓨팅 알고리즘을 사용하여 잡음이 없는 시뮬레이션에서 UCCSD 및 k-UpCCGSD 의 결과는 동일한 active space 을 활용한 full configurational interaction (FCI) 결과와 유사했다. 이는 양자 알고리즘이 SN2 반응의 포텐셜 에너지 표면을 설명할 수 있음을 보여준다. 양자 잡음이 있는 시뮬레이션에서 UCCSD 는 k-UpCCGSD 보다 양자 잡음에 더 취약하였다. 따라서 k-UpCCGSD 는 잡음이 있는 중간 규모의 양자 시대에 있어서 양자 잡음 효과를 줄이기 위해 UCCSD 의 대안으로 사용할 수 있고, k-UpCCGSD 는 본 연구에서 SN2 반응의 포텐셜 에너지 표면을 설명하기 충분하였다. 본 연구의 결과는 SN2 반응 경로에 양자 컴퓨팅의 적용 가능성을 보여주었고, 양자 컴퓨팅을 사용한 구조 기반의 분자 시뮬레시션에 유용한 정보를 제공할 것으로 기대된다. 생물의약과 생물소재를 개발하는데 있어서 단백질 공학의 중요성을 점점 커지고 있다. 컴퓨터를 이용한 단백질 공학에 있어서 기계학습은 많은 가상의 단백질 서열들로부터 원하는 특성을 가지는 최적의 단백질 서열을 찾는 과정 속에서 실험적인 노력을 현격히 줄일 수 있다. 컴퓨터를 활용한 단백질 공학 문제를 해결하는데 중요한 일반적인 단백질 표현자를 위해, 본 연구를 통해 서열 기반의 표현자인 PCgrades, 구조 기반의 표현자인 PCsparis, 3D-SPIEs 표현자를 개발하였다. PCgrades 와 PCspairs 는 아미노산과 아미노산 쌍에 있어서 물리화학적인 특성들로부터 추출한 통계적인 정보를 기반으로 하기에 일반적인 정보를 주로 담고 있다. 3D-SPIEs 는 FMO2-DFTB3/D/PCM 기반의 조각 단위 양자역학적인 정보를 기반으로 하기에 특이적인 정보를 주로 담고 있다. 본 연구에서 SARS-CoV-2 스파이크 당단백질의 단백질 발현량과 결합력을 포함한 다양한 데이터들에 적용하여 예측 모델을 만들어서 본 연구에서 개발한 단백질 표현자를 평가하였다. 결과적으로 새롭게 개발된 단백질 표현자들은 다양한 데이터들에서도 좋은 성능을 보였고, PCspairs 는 단백질 발현량 예측에 있어서 R2=0.783 의 성능과 결합력 예측에 있어서 R2=0.711 의 성능을 보여주며 가장 좋은 표현자로 나타났다. 따라서 산업적인 효소와 단백질 치료제를 위해 주로 사용하고 있는 대량 신속처리 실험들로부터 나오는 데이터들을 이용하여 학습한다면, 새롭게 개발한 단백질 표현자들을 활용한 비슷한 접근 방식은 유용할 것으로 기대된다. 본 학위논문에서 신약개발과 단백질 개량을 목적으로 양자역학과 기계학습을 기반으로 한 정확하고 효과적인 평가시스템인 AVENGERS 을 개발하였다. 두번째 장에서 소개한 SARS-CoV-2 에 대한 핫스팟 분석 연구는 양자역학적인 결과를 바탕으로 만든 3D-SPIEs 기반의 핫스팟 지도는 단백질-단백질 상호작용을 정량적으로 분석하는데 적합하다는 것을 보여주었다. 세번째 장에서 소개한 ANI-2X 를 기반으로 한 결합상태 평가 연구는 신경망 포텐셜을 이용하여 분자도킹에서의 결합상태를 평가하는 방법은 가장 좋은 결합 상태를 찾는데 유용하다는 것을 보여주었다. 네번째 장에서 소개한 조각 단위 분자궤도함수를 사용한 리간드 평가 연구는 조각 단위 분자궤도함수는 리간드의 결합력을 정량적으로 예측하기 위한 평가 함수를 만드는데 적합하는 것을 보여주었다. 앞서 소개한 신약개발을 위한 3가지 연구 사례들은 양자역학적인 방법들은 분자 시뮬레이션에 있어서 정확한 자유 에너지 추정을 가능하게 한다는 것을 보여주었다. 다섯번째 장에서 소개란 SN2 반응에 대한 양자컴퓨터적인 분석 연구는 양자컴퓨터는 양자역학적인 방법을 가속화할 뿐 아니라 고전적인 컴퓨터에서 얻은 양자역학적인 결과를 화학적 정확성 내로 결과를 얻을 수 있다는 것을 보여주었다. 여섯번째 장에서 소개한 다양한 단백질 특성들을 예측하기 위한 단백질 표현자 개발 연구는 새롭게 개발된 단백질 표현자는 기계학습을 통해 다양한 단백질 특성 예측 모델을 만들었을 때 좋은 성능을 보인다는 것을 보여주었다. 결과적으로 AVENGERS 는 양자역학과 기계학습 방법들을 기반으로 신약개발과 단백질 개량을 위해 개발되었다. 이는 정량적인 예측 모델들을 기반으로 저분자 화합물과 단백질의 여러 특징을 정량적으로 예측하는 도움을 줌으로써 정확학 가상탐색을 통해 신약개발과 단백질 개량을 가속화 할 것이다.

공학 기술 기반의 자동차 디지털 디자인 : Digital Design of A Passenger Car based on Engineering Technology

맹주원 인하대학교 대학원 2009 국내박사

Every car manufacturer desires to reduce the new car development time spent in improving the safety, NVH, lightweight, reliability and environment friendly features of the car at the same time. Other considerations such as planning, exterior and interior styling, packaging, color, trim, and material selection increase the complexity of the car design process. Also the importance of the concept stage is becoming a greater priority in new car development. This dissertation proposes a digital design process to utilize the engineering analysis and design/styling software in early engineering stages of car design. Proposed digital design process can be efficiently used by a team of car research center or a studio with small number of engineers, helping ordinary engineers becoming ambidextrous in design as well as engineering applications. The concept model starts from feasibility study, idea and concept sketch, rendering, and 3D surface model with computer aided styling (CAS) to the final safety estimation by using proposed digital design process based on engineering technology (CAD, CAE, CFD, CNC). As an illustration for the concept model proposed a hydrogen fuel cell sports coupe equipped wheel hub motor drive system which could be available within next 10 years. The first objective is define a suitable style for the concept model. This concept model is suggested to environment friendly, lightweight, minimal style, and high technology. After conceiving the idea, the concept sketch is drawn and developed into 2D digital sketch and rendering by using Autodesk Sketchbook Pro, and Adobe Photoshop. The 2D digital data is exported to Autodesk Alias Studio for creating 3D surface model. This 3D surface model is used to review and realization the concept model after realistic rendering which applied material, color, texture, surface treatment, and environment, etc. Once the digital styling, engineering analysis step is performed. Dassault systemes CATIA V5 imported the 3D surface model data (STP format) from Alias Studio for computer aided engineering (CAE) application; checking engineering tolerance. Torsional rigidity, bending rigidity, and normal modes of body-in-white (BIW) model are analyzed for evaluating structural performances by using ANSYS 11. For predicting the crashworthiness of the concept model based on US NCAP (100% frontal crash at 56 km/h) and AMS (50% offset frontal crash at 55 km/h) by using LSTC LS-DYNA. The aerodynamic forces on body surface is computed by using ANSYS CFX. Moreover, to reduce the aerodynamic forces (drag and lift) the car body shape is modified without changing BIW. Finally, a 1/10 scale physical model is created by computer numerical control (CNC) milling machine (DECKEL MAHO DMU 60T) for final realization and inspection. It is concluded from this study that proposed digital design process can improve the concept model development within 12 weeks. This proposed digital design process can not only reduce the new car development time and cost of the final product but also be adopted into design of varied products design such as aircraft, yacht, electrical equipment and sports gear. Also proposed digital design process will serve as a valuable tool to teach systems engineering issues to beginning car designers and engineers. 세계의 자동차 제조 회사들은 안전, 소음 및 진동, 경량과 친환경 등을 동시에 고려하여 신차 개발 기간을 단축하는데 많은 연구와 투자를 하고 있다. 자동차 디자인 프로세스에 있어서는 기획, 외부/내부 스타일링, 패키징, 컬러, 트림과 재료의 선택 등이 고려된다. 또한 신차 개발에 있어서 초기 컨셉 단계의 중요성이 매우 높아지고 있다. 본 논문은 자동차 디자인의 초기 공학 단계에서 공학 해석과 디자인/스타일링 소프트웨어를 사용한 디지털 디자인 프로세스를 제안하였다. 제안된 디지털 디자인 프로세스는 완성차의 단일팀, 소수의 공학자로 이루어진 스튜디오와 소규모 주문 생산 업체에서 효율적으로 사용할 수 있으며 일반 공학자가 공학적 활용뿐만 아니라 디자인도 함께 진행할 수 있다. 직접적인 사례를 통하여 제안된 디지털 디자인 프로세스의 효율성을 확인하였다. 컨셉 모델은 향후 10년 후 양산 가능한 휠 허브 모터 구동 시스템을 장착한 친환경 수소 연료 전지 스포츠 쿠페를 제안하였다. 컨셉 모델 개발 과정은 타당성 연구, 아이디어와 컨셉 스케치, 렌더링과 CAS 소프트웨어를 이용한 3차원 서피스 모델 생성의 디자인 단계부터 공학 기술 (CAD, CAE, CFD, CNC)을 적용한 차체 강성, 충돌 안전성, 공기 역학 해석과 최종 물리적 모델 제작의 공학 단계까지 진행하였다. 먼저, 컨셉 모델의 외부 스타일링은 친환경, 경량, 미니멀 스타일링과 첨단 기술을 키워드로 하여 접근하였다. 아이디어와 컨셉 스케치는 손 스케치뿐만 아니라 Wacom Intuos3 tablet를 입력장치로 Autodesk Sketchbook Pro와 Abode Photoshop 소프트웨어를 이용하여 2차원 디지털 스케치를 하였다. Autodesk Alias Studio 소프트웨어를 이용하여 컨셉 모델의 2차원 사면도 (정면, 후면, 측면, 평면)를 불러온 후 3차원 서피스 모델을 생성하였다. 생성된 3차원 서피스 모델은 각 파트에 재료, 색깔, 질감, 표면 처리와 주위 환경 등을 적용하여 중간 품평에 사용하였다. 이후 공학 기술 적용 단계에서는 Dassault systems의 CATIA V5를 이용하여 Alias Studio에서 생성한 3차원 서피스 모델 (STP 형식)을 공학 해석 (CAE)에 적용 가능한 모델로 수정하였다. 상기의 과정을 통하여 생성한 BIW 모델을 상용 유한요소해석 소프트웨어인 ANSYS 11을 이용하여 차체 강성 (굽힘, 휨)과 고유진동수 해석을 수행하였다. 충돌 안전성 해석은 상용 비선형해석 소프트웨어인 LSTC LS-DYNA를 이용하여 US NCAP (100% 정면 충돌, 56 km/h)과 AMS (50% 정면 옵셋 충돌, 55 km/h)를 수행하였다. 차체의 항력과 양력을 구하기 위하여 상용 CFD 해석 소프트웨어인 ANSYS CFX를 이용하였으며 CFD 해석 모델은 BIW 모델은 변경하지 않고 항력과 양력을 감소시키도록 외부 스타일링만을 수정하였다. 마지막으로 최종 품평을 위하여 3차원 서피스 모델을 CAM 소프트웨어인 Delcam PowerMILL (IGES 형식)에서 툴 경로를 지정하고 CNC 밀링 머신 (DECKEL MAHO DMU 60T)을 이용하여 1/10 스케일의 3차원 물리적 모델을 제작하였다. 본 디지털 디자인 프로세스를 적용하여 컨셉 모델의 아이디어 스케치 단계에서 1/10 스케일의 3차원 물리적 모델 제작까지 총 12주가 소요되었다. 제안된 디지털 디자인 프로세스는 신차 개발 기간과 비용 절감뿐만 아니라 항공기, 요트, 전자 제품과 운동 기구 등의 제품 디자인에도 적용할 수 있다. 또한 자동차 디자이너, 제품 디자이너, 공학자 등을 위해 제품 개발의 디지털 디자인 프로세스를 경험할 수 있는 교육용 프로그램으로도 활용 가능하다.

Novel Computer Vision-Based Vehicle Non-Contact Weigh-In-Motion System

Leung, Ryan Yuen Hung Columbia University ProQuest Dissertations & These 2022 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

grant a,b*/ gue-induced deterioration and critical fracture to civil infrastructure, among many other purposes. Developing a cost-effective weigh-in-motion (WIM) system remains challenging. This doctoral research describes the creation and experimental validations of a computer vision- based non-contact vehicle WIM system. The underlining physics is that the force exerted by each tire onto the roadway is the product of the two key vehicle parameters: tire-roadway contact pressure and contact area. Computer vision is applied (1) to measure the several tire parameters so that the tire-roadway contact area can be accurately estimated; and (2) to recognize the marking texts on the tire sidewall so that the manufacturer-recommended tire pneumatic pressure can be found. Consequently, a computer vision system is developed in this research. The computer vision system comprises a camera and computer vision software/techniques for measuring the tire parameters and recognizing the tire sidewall markings from individual tire images of a moving vehicle. Computer vision techniques, such as edge detection and optical character recognition (OCR), are applied to enhance the measurements and recognition accuracy. Numerous laboratory and field experiments were conducted on a sport utility vehicle and fully loaded or empty concrete trucks to demonstrate the feasibility of this novel method. The vehicle weights estimated by this novel computer vision-based non-contact vehicle WIM system agreed well with the curb weights verified by static weighing, demonstrating the potential of this computer vision- based method as a non-contact means for weighing vehicles in motion. To further illustrate and exemplify the versatility of this novel computer vision-based WIM system, this research explores the potential application capability of the system for structural health monitoring (SHM) in civil engineering. This work aims to investigate the potential of this proposed and prototyped computer vision-based vehicle WIM system to acquire vehicle weight and location information as well as to obtain corresponding bridge responses simultaneously for later structural model updating analysis and damage detection and identification framework. In order to validate the concept, a laboratory vehicle-bridge model was constructed. Subsequently, numerous experiments were carried out to demonstrate how the computer vision-based WIM system can be utilized as a resourceful application to (1) extract bridge responses, (2) estimate vehicle weight, and (3) localize the input force simultaneously. This doctoral research delivers a unique, pioneering, and innovative design and development of a computer vision-based non-contact vehicle WIM method and system that can remotely perform vehicle weight estimation. It also demonstrates a novel application of computer vision technology to solve challenging weigh-in-motion (WIM) and civil engineering problems.

Engineering a Control System for a Logical Qubit-Scale Trapped Ion Quantum Computer

Risinger, Andrew Russ University of Maryland, College Park ProQuest Diss 2023 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Quantum computing is a promising field for continuing to develop new computing capabilities, both in its own right and for continued gains as Moore's Law growth ends. Trapped ion quantum computing is a leading technology in the field of quantum computing, as it combines the important characteristics of high fidelity operations, individual addressing, and long coherence times. However, quantum computers are still in their infancy; the first quantum computers to have more than a handful of quantum bits (qubits) are less than a decade old. As research groups push the boundaries of the number of qubits in a system, they are consistently running into engineering obstacles preventing them from achieving their goals. There is effectively a knowledge gap between the physicists who have the capability to push the field of quantum computing forward, and the engineers who can design the large-scale & reliable systems that enable pushing those envelopes. This thesis is an attempt to bridge that gap by framing trapped ion quantum computing in a manner accessible to engineers, as well as improving on the state-of-the-art in quantum computer digital and RF control systems.We also consider some of the practical and theoretical engineering challenges that arise when developing a leading-edge trapped ion quantum computer capable of demonstrating error-corrected logical qubits, using trapped 171Yb+ qubits. There are many fundamental quantum operations that quantum information theory assumes, yet which are quite complicated to implement in reality. First, we address the time cost of rearranging a chain of ions after a scrambling collision with background gases. Then we consider a gate waveform generator that reduces programming time while supporting conditional quantum gates. Next, we discuss the development of a digital control system custom-designed for quantum computing and quantum networking applications. Finally, we demonstrate experimental results of the waveform generator executing novel gate schemes on a chain of trapped ions. These building blocks together will unlock new capabilities in the field of trapped ion quantum computers.

Four-dimensional and vision-based framework for infrastructure maintenance management

Zhang, Zixiao Stanford University 2009 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Billions of dollars are spent each year to address the ever-increasing demands of maintaining large civil infrastructure assets in the United States. Rapid and reliable updating of the status of tens of thousands of structures and making the information readily accessible for maintenance engineers has become a big challenge for infrastructure management. Four-dimensional (4D) and computer vision technology are two relatively new and promising technologies in the field of civil engineering. By linking construction schedules with 3D geometric models, 4D has shown its strength in managing and visualizing time-related data in construction planning. However, the capability of 4D tools to manage maintenance information through the lifecycle of infrastructure has not yet been much explored. A 4D and vision-based framework to acquire and manage the maintenance information of infrastructures is proposed to fully explore and combine the advantages of both 4D modeling and computer vision technology. In the framework, 4D modeling is used to manage and provide easy access to maintenance data through multiple interfaces. Computer vision technology is used to automate the process of acquiring cracking from reinforced concrete structures and provide timely updates about the conditions of structures for maintenance prioritization. The developed 4D-based bridge/infrastructure maintenance information management system was evaluated with a group of maintenance engineers and engineering students. Experiment results showed that with the help of the 4D-based system, access to maintenance information was made easier and the engineers were able to complete assigned tasks with more accuracy, efficiency and consistency than using traditional paper-based reports. A series of computer vision-based techniques were developed for acquiring condition information, in this case, cracking from pictures of reinforced concrete structures. The techniques developed include a self-adaptive segmentation and registration algorithm for reinforced concrete beams, crack-detection enhancing algorithms for reinforced concrete, a method to deduce load configuration from crack patterns using probabilistic inference, a method to recover cracks that are not captured in the images and a technique to estimate the remaining capacity of reinforced concrete structures using the obtained crack data. The developed techniques were evaluated with images and data from three laboratory tests of reinforced-concrete beams. Experiment results showed that the segmentation algorithm worked consistently and successfully with all 159 images taken during the beam tests. The quality and speed of crack detection was improved with the enhancing algorithms. Based on the obtained crack pattern, the algorithm was able to successfully recover cracks that were not captured in the images, deduce the underlying load configuration and estimate the sustained plastic damages of the reinforced concrete beams.

Joint processing of audio-visual information for the recognition of emotional expressions in human-computer interaction

Chen, Lawrence Shao-Hsien University of Illinois at Urbana-Champaign 2000 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Recent technological advances have enabled human users to interact with computers in ways previously unimaginable. Beyond the confines of the keyboard and mouse, new modalities to control the computer such as voice, gesture, and force-feedback are emerging. Among these, voice and vision are two natural modalities in human-to-human communication. Automatic speech recognition (ASR) technology has matured enough to allow users to dictate to a word processor or operate the computer using voice commands. Computer vision techniques have enabled the computer to see. Interacting with computers in these modalities is much more natural for people, and the progression is towards the kind of interaction between humans. Despite these advances, one necessary ingredient for natural interaction is still missing—emotions. Emotions play an important role in human-to-human communication and interaction, allowing people to express themselves beyond the verbal domain. The ability to understand human emotions is desirable for the computer in some applications such as computer-aided learning or user-friendly online help. This thesis addresses the problem of detecting human emotional expressions by computer from the voice and facial motions of the user. The computer is equipped with a microphone to listen to the user's voice, and a video camera to look at the user. Prosodic features in the audio and facial motions exhibited on the face can help the computer make some inferences about the user's emotional state, assuming the users are willing to show their emotions. Another problem it addresses is the coupling between voice and the facial expression. Sometimes the user moves the lips to produce the speech, and sometimes the user only exhibits facial expression without speaking any words. Therefore, it is important to handle these two modalities accordingly. In particular, a pure “facial expression detector” will not function properly when the person is speaking, and a pure “vocal emotion recognizer” is useless when the user is not speaking. In this thesis, a complementary relationship between audio and video is proposed. Although these two modalities do not couple strongly in time, they seem to complement each other. In some cases, similar facial expressions may have different vocal characteristics, and vocal emotions having similar properties may have distinct facial behaviors.

Hybrid Analog-Digital Co-Processing for Scientific Computation

Huang, Yipeng Columbia University ProQuest Dissertations & These 2018 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

In the past 10 years computer architecture research has moved to more heterogeneity and less adherence to conventional abstractions. Scientists and engineers hold an unshakable belief that computing holds keys to unlocking humanity's Grand Challenges. Acting on that belief they have looked deeper into computer architecture to find specialized support for their applications. Likewise, computer architects have looked deeper into circuits and devices in search of untapped performance and efficiency. The lines between computer architecture layers---applications, algorithms, architectures, microarchitectures, circuits and devices---have blurred. Against this backdrop, a menagerie of computer architectures are on the horizon, ones that forgo basic assumptions about computer hardware, and require new thinking of how such hardware supports problems and algorithms. This thesis is about revisiting hybrid analog-digital computing in support of diverse modern workloads. Hybrid computing had extensive applications in early computing history, and has been revisited for small-scale applications in embedded systems. But architectural support for using hybrid computing in modern workloads, at scale and with high accuracy solutions, has been lacking. I demonstrate solving a variety of scientific computing problems, including stochastic ODEs, partial differential equations, linear algebra, and nonlinear systems of equations, as case studies in hybrid computing. I solve these problems on a system of multiple prototype analog accelerator chips built by a team at Columbia University. On that team I made contributions toward programming the chips, building the digital interface, and validating the chips' functionality. The analog accelerator chip is intended for use in conjunction with a conventional digital host computer. The appeal and motivation for using an analog accelerator is efficiency and performance, but it comes with limitations in accuracy and problem sizes that we have to work around. The first problem is how to do problems in this unconventional computation model. Scientific computing phrases problems as differential equations and algebraic equations. Differential equations are a continuous view of the world, while algebraic equations are a discrete one. Prior work in analog computing mostly focused on differential equations; algebraic equations played a minor role in prior work in analog computing. The secret to using the analog accelerator to support modern workloads on conventional computers is that these two viewpoints are interchangeable. The algebraic equations that underlie most workloads can be solved as differential equations, and differential equations are naturally solvable in the analog accelerator chip. A hybrid analog-digital computer architecture can focus on solving linear and nonlinear algebra problems to support many workloads. The second problem is how to get accurate solutions using hybrid analog-digital computing. The reason that the analog computation model gives less accurate solutions is it gives up representing numbers as digital binary numbers, and instead uses the full range of analog voltage and current to represent real numbers. Prior work has established that encoding data in analog signals gives an energy efficiency advantage as long as the analog data precision is limited. While the analog accelerator alone may be useful for energy-constrained applications where inputs and outputs are imprecise, we are more interested in using analog in conjunction with digital for precise solutions. This thesis gives novel insight that the trick to do so is to solve nonlinear problems where low-precision guesses are useful for conventional digital algorithms. The third problem is how to solve large problems using hybrid analog-digital computing. The reason the analog computation model can't handle large problems is it gives up step-by-step discrete-time operation, instead allowing variables to evolve smoothly in continuous time. To make that happen the analog accelerator works by chaining hardware for mathematical operations end-to-end. During computation analog data flows through the hardware with no overheads in control logic and memory accesses. The downside is then the needed hardware size grows alongside problem sizes. While scientific computing researchers have for a long time split large problems into smaller subproblems to fit in digital computer constraints, this thesis is a first attempt to consider these divide-and-conquer algorithms as an essential tool in using the analog model of computation. As we enter the post-Moore's law era of computing, unconventional architectures will offer specialized models of computation that uniquely support specific problem types. Two prominent examples are deep neural networks and quantum computers. Recent trends in computer science research show these unconventional architectures will soon have broad adoption. In this thesis I show another specialized, unconventional architecture is to use analog accelerators to solve problems in scientific computing. Computer architecture researchers will discover other important models of computation in the future. This thesis is an example of the discovery process, implementation, and evaluation of how an unconventional architecture supports specialized workloads.

Supporting Computer Science Education Through Automation and Surveys

Presler-Marshall, Kai North Carolina State University ProQuest Dissertat 2022 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Software engineering is a growing field, with ever-increasing demand for capable engineers who can design, implement, and test the software that is needed for the modern world. With this increasing demand for software engineers, there is a corresponding increase in the demand placed on computer science programs that graduate these engineers. However, the increase in undergraduate enrollment in computer science programs has generally outpaced the increase in instructors. Unfortunately, this can have negative educational impacts by reducing the support that instructors can offer each student. Automation has resulted in significant benefits, allowing developers to work more efficiently and deliver higher-quality software, but automation is not as prevalent within computer science education as it is within industry. To help promote better educational outcomes, particularly by improving the feedback that students receive on their work, I adopt software engineering automation techniques into computer science education and evaluate their efficacy.With ever more students enrolled in computer science programs comes a more widespread use of team-based learning (TBL) and larger teams. While TBL has numerous educational benefits, it is not an educational panacea. Larger teams increases the risk of team challenges, including ineffective communication and non-participation, which has the potential to hamper educational outcomes. To address this, I propose and evaluate using survey techniques to gain insights into how teams work and the challenges that students face in this environment, and enable just-in-time support for struggling teams. This approach can provide instructors with feedback on team challenges, and also encourages self-reflection on the part of students. Together, these approaches support my thesis: Using software engineering automation and survey techniques in computer science education results in improved student learning outcomes, early prediction of struggling teams, and more effective instructional materials.My first two research contributions focus on applying software engineering automation to support individual students. My test flakiness study investigates the impact that configuration options have upon the stability of Selenium tests. This supports improved educational outcomes by giving students more consistent feedback and greater confidence in the code and tests that they write. My automated program repair study investigates the mistakes that students make when learning SQL, and introduces an automated program repair tool for SQL queries. It demonstrates that automated repair can be applied to special-purpose languages such as SQL, and that students find automatically-repaired SQL queries to be understandable, suggesting that they may have promise as an instructional technique.My final three research contributions use survey techniques and software engineering automation to focus on supporting software engineering student teams. My collaboration reflection study investigates the use of a team collaboration reflection survey (TCRS) for identifying software engineering student teams that are struggling to collaborate effectively. It shows that most (89%) teams which later receive poor grades can be flagged through the TCRS, typically by the halfway mark of the project, and students appreciated that the TCRS encourages self-reflection. In my team challenges study, to better understand team challenges that were uncovered through the TCRS, I interview students who had recently completed a team-based software engineering course about their teaming experiences. This provides novel insights into how teams work together, and the types of issues that students face and how they attempt to overcome them.

CMOS Ising Processor and Spintronic Memory Solution: From Concept to Implementation

Ahmed, Ibrahim ProQuest Dissertations & Theses University of Minn 2020 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

CMOS based information technology is facing two challenges simultaneously. The first challenge is efficiently solving the increasingly complex problems we are facing every day. Combinatorial optimization problems (COPs) are one such widely used complex problems. Real-world applications such as supply chain management, logistics control, transportation system design, communication network design, and VLSI layout optimization can be mapped to COP. The computational complexity of many of these COPs is NP-complete, NP-hard, or worse. The computation time for finding the optimal or near-optimal solution increases exponentially with the number of variables for a conventional von Neumann computer. This becomes a fundamental bottleneck for the large problem sizes that are associated with real-world problems. The second challenge is that the CMOS technology that powers all our computing devices reaches its physical limits resulting in the slow growth of computing power. Our data- and computation-driven society increasingly demand more computationally powerful and energy-efficient devices. The recent slow growth CMOS technology has driven the research for an alternative technology to replace some or all of CMOS technology. We explored two potential solutions to address the challenges mentioned above. A non-traditional computing method, Ising computer, have shown to solve COPs very efficiently with a small area and energy consumption. In this thesis, we explored a dedicated hardware accelerator based on CMOS Ising computer to address COPs. On the other hand, spintronic devices are a promising alternative to silicon-based CMOS technology. The spintronic memory applications have shown a lot of promise in recent years. We analyzed various competing spin-based memory write schemes and considered the viability of a spintronic memory solution in this thesis. We designed three CMOS Ising computers with an increasing number of spins and features to progress towards a dedicated Ising processor aiming to solve COPs. We designed electrically coupled CMOS ring oscillators as the network of spins, and studied various architectures and coupling mechanisms in this work. Our proof-of-concept design had six CMOS spins coupled with pseudoresistors. The pseudoresistors were controlled using digitally programmable digital to analog converters (DAC). We successfully mapped and solved NP-hard max-cut problems with an average accuracy of 91%, proving the feasibility of a CMOS Ising computer. However, the DAC circuits required prohibitively large current for a larger Ising computer. Additionally, the all-to-all connected architecture was not modular, and the layout complexity was too high for a practical hardware accelerator with hundreds or thousands of spins. Our second Ising computer was designed with 560 spins coupled using a digital latch based coupling. We found that the Ising computer was probabilistically exploring different local minima with similar quality solutions. The probabilistic nature of the Ising computer is essential to solving difficult COPs. We mapped and solved 1000 graph problems in our chip with an accuracy of 82%-100% compared to the solution of a commercial COP solver. Our Ising computer was 10. 4-10. 6 times faster than the software with four orders of magnitude smaller energy requirement. Additionally, our experiments showed the Ising computer solutions are very consistent at various temperatures and supply voltage conditions. The measured results proved CMOS Ising computer is an excellent candidate for a hardware accelerator. Our third Ising computer was designed with 2150 spins coupled using a pass-gate based coupling which included multi-bit resolution coupling and local field bias. The proposed Ising computer can solve even more diverse and complex problems with the added features. Additionally, we designed a global coupling strength control to achieve better annealing to improve the solution quality. Our preliminary results show the Ising computer can solve difficult problems with an accuracy of 95%-100%. For our spintronic memory work, we developed a universal SPICE model for various MTJ write mechanisms, including spin-transfer torque (STT) and spin hall effect (SHE). We ran Monte-Carlo simulations using realistic magnetic and geometric parameters. The simulations showed SHE is less susceptible to thermal fluctuation than STT. SHE-only switching with a larger write current showed 8x and 7x delay and energy reduction, respectively. On the other hand, SHE-assisted STT switching showed 2x and 3x delay and energy reduction, respectively, with a smaller current requirement. Our analysis indicates SHE-assisted STT scheme can be a viable candidate for embedded applications, including the Ising computer.

(A) case study of Korean adult EFL writers’ computer-assisted L2 writing focusing on the use of online resources as L2 writing aids

민주영 서울대학교 대학원 2017 국내박사

The present L2 writing pedagogy is required for considering the impact of technology on L2 writing. The development of computer technologies provides L2 writers with advanced technological writing aids on the Web: online dictionaries, search engines, and web-based corpora. Although these online resources have great potential to assist L2 writing, the combined use of these online resources by L2 writers during the process of L2 writing in the EFL context, beyond the classroom, has been less investigated in previous studies. The present study investigated three Korean adult EFL writers’ use of the online resources in computer-assisted L2 writing. Each writer performed L2 writing tasks on the computer using three online resources—an online dictionary, a search engine, and a web-based corpus—in a series of thirty writing sessions. For each writer, a simultaneous think-aloud and a retrospective interview were also conducted in every session. The writing process on the computer screen with the writer’s think-aloud and the interviews were recorded and transcribed in order to identify the aspects of the writers’ use and perception of the online resources in computer-assisted L2 writing. These aspects of the use of online resources in the writing process were classified and examined according to the purpose of each use, and the interview transcripts were analyzed by themes. The findings revealed each writer’s respective and combined use of the online resources during L2 writing and perceptional aspects related to the use of the online resources and writing in English. As the writers used L1 extensively during their L2 writing process, such an approach to L2 writing led them to use a bilingual online dictionary most frequently as well as to prefer a bilingual interface of the online resources. The writers thought that the dictionary examples provided with Korean translations were useful to them. The writers’ use of a search engine and a corpus demonstrated each writer’s varying perception of the function and utilization of each online resource and the writing task. The writers’ L2 English writing ability and their manner of writing in English appeared to show the emergence of different aspects of using the online resources among the writers. All the writers positively evaluated and appreciated using the online resources in L2 writing and its effect on their affective aspects. They considered the online resources to be useful and helpful in their L2 writing, especially for supporting their affective aspects related to L2 writing confidence and self-efficacy. Thus, it was found that computer-assisted L2 writing with the aid of online resources can assist not only these EFL writers’ L2 writing, but also enhance their affective aspects. Their L2 writing apprehensions were relieved, and their attitude toward L2 writing had improved. The writers were also aware of the limitations of their self-directed computer-assisted L2 writing in terms of detecting and revising their own writing errors. From the case of these three Korean adult EFL writers, the following five issues were discussed. First, computer technologies have a positive influence on the affective dimension of EFL writers. Second, the value of writing experience and practice to Korean adult EFL writers is significant, and thus it is necessary to consider providing EFL writers with an L2 writing experience that incorporates computer technologies related to L2 writing. Third, future technology that could assist EFL writers’ L2 writing was discussed. Fourth, the role of an L2 writing teacher in computer-assisted L2 writing was discussed, and finally, how to define L2 writing ability in the era of technological innovation was discussed. Based on the case analysis and discussion, pedagogical implications and suggestions for future research were presented. The need exists for a balanced and integrated consideration of EFL writers’ cognition, affect, and the use of computer technology in L2 writing pedagogy. While the active use of the online resources during L2 writing is recommended, a systemic approach to utilizing such technologies is required in order to develop and improve EFL writers’ L2 writing skills. L2 writing teachers should be aware of the value of online resources in L2 writing, identify L2 writers’ affective aspects, and find a proper way to support L2 writers. Computer-assisted L2 writing is able to support EFL writers’ affective aspects and provide EFL writers with an instructive L2 writing experience. Further, advanced technologies to support independent EFL writers’ needs and to solve their writing problems are required, and the use of online resources during L2 writing can be considered both in the current as well as future L2 writing assessments.