DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 강남우 | - |
dc.contributor.author | Shin, Seungyeon | - |
dc.contributor.author | 신승연 | - |
dc.date.accessioned | 2024-07-25T19:31:29Z | - |
dc.date.available | 2024-07-25T19:31:29Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1045982&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/320747 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 조천식모빌리티대학원, 2023.8,[iii, 38 p. :] | - |
dc.description.abstract | The product design process in manufacturing involves iterative design modeling and analysis to achieve the target engineering performance, but such an iterative process is time consuming and computationally expensive. Recently, deep learning-based engineering performance prediction models have been proposed to accelerate design optimization. However, they only guarantee predictions on training data and may be inaccurate when applied to new domain data. In particular, 3D design data have complex features, which means domains with various distributions exist. Thus, the utilization of deep learning has limitations due to the heavy data collection and training burdens. We propose a bi-weighted unsupervised domain adaptation approach that considers the geometry features and engineering performance of 3D design data. It is specialized for deep learning-based engineering performance predictions. Domain-invariant features can be extracted through an adversarial training strategy by using hypothesis-discrepancy, and a multi-output regression task can be performed with the extracted features to predict the engineering performance. In particular, we present a source instance weighting method suitable for 3D design data to avoid negative transfers. The developed bi-weighting strategy based on the geometry features and engineering performance of engineering structures is incorporated into the training process. The proposed model is tested on a wheel impact analysis problem to predict the magnitude of the maximum von Mises stress and the corresponding location of 3D road wheels. This mechanism can reduce the target risk for unlabeled target domains on the basis of weighted multi-source domain knowledge and can efficiently replace conventional finite element analysis. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 딥러닝▼a비지도 도메인 적응▼a공학 성능 예측▼a기하 특성▼a인스턴스 가중치▼a다중 출력 회귀▼a휠 충격 시험 | - |
dc.subject | Deep learning▼aUnsupervised domain adaptation▼aEngineering performance▼aGeometry features▼aInstance weighting▼aMulti-output regression▼aWheel impact test | - |
dc.title | Weighted unsupervised domain adaptation considering geometry feature and engineering performance of 3D design data | - |
dc.title.alternative | 3D 설계 데이터의 기하 특성과 공학 성능을 고려한 가중 비지도 도메인 적응 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :조천식모빌리티대학원, | - |
dc.contributor.alternativeauthor | Kang, Namwoo | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.