DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 주재걸 | - |
dc.contributor.author | Lee, Jungsoo | - |
dc.contributor.author | 이정수 | - |
dc.date.accessioned | 2024-08-08T19:30:58Z | - |
dc.date.available | 2024-08-08T19:30:58Z | - |
dc.date.issued | 2024 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1098136&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/321981 | - |
dc.description | 학위논문(박사) - 한국과학기술원 : 김재철AI대학원, 2024.2,[vi, 56 p. :] | - |
dc.description.abstract | however, they often suffer from notable performance degradation when confronted with distribution shifts between the train set and the test set. Addressing this challenge without the need for additional data collection has become a recent research focus. A primary hurdle in handling such shifts is the dataset bias, where models overly rely on the unwanted correlation between peripheral attributes and labels. This bias can lead models to learn irrelevant features, hindering their ability to generalize to various data distributions. Another challenge is the domain shift, which encompasses differences in style, object sizes, or sources of datasets. Among various techniques for addressing the domain shift, test-time adaptation (TTA) has gained traction for its practicality in mitigating these shifts. This thesis mainly tackles such two challenges in computer vision and further suggests the future work direction of addressing distribution shifts. | - |
dc.description.abstract | In computer vision, deep neural networks have made significant progresses | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 분포 변화▼a견고함▼a편향 완화▼a테스트 시간 적응 | - |
dc.subject | Distribution shift▼aRobustness▼aDebiasing▼aTest-time adaptation | - |
dc.title | Addressing distribution shift in computer vision | - |
dc.title.alternative | 컴퓨터 비전에서 발생하는 분포 변화 완화 | - |
dc.type | Thesis(Ph.D) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :김재철AI대학원, | - |
dc.contributor.alternativeauthor | Choo, Jaegul | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.