DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 윤철희 | - |
dc.contributor.author | Oh, Junsoo | - |
dc.contributor.author | 오준수 | - |
dc.date.accessioned | 2024-07-30T19:30:41Z | - |
dc.date.available | 2024-07-30T19:30:41Z | - |
dc.date.issued | 2024 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1096073&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/321368 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 김재철AI대학원, 2024.2,[iv, 63 p. :] | - |
dc.description.abstract | perhaps surprisingly, the task of finding optimal decision boundaries becomes harder for more separable distributions. For Mixup training, we show that Mixup mitigates this problem by significantly reducing the sample complexity. To this end, we develop new concentration results applicable to $n^2$ pair-wise augmented data points constructed from $n$ independent data, by carefully dealing with dependencies between overlapping pairs. Lastly, we study other masking-based Mixup-style techniques and show that they can distort the training loss and make its minimizer converge to a suboptimal classifier in terms of test accuracy. | - |
dc.description.abstract | We investigate how pair-wise data augmentation techniques like Mixup affect the sample complexity of finding optimal decision boundaries in a binary linear classification problem. For a family of data distributions with a separability constant~$\kappa$, we analyze how well the optimal classifier in terms of training loss aligns with the optimal one in test accuracy (i.e., Bayes optimal classifier). For vanilla training without augmentation, we uncover an interesting phenomenon named the \emph{curse of separability}. As we increase $\kappa$ to make the data distribution more separable, the sample complexity of vanilla training increases exponentially in $\kappa$ | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 믹스업▼a최적의 결정 경계▼a샘플 복잡도▼a이론▼a일반화 | - |
dc.subject | Mixup▼aOptimal decision boundary▼aSample complexity▼aTheory▼aGeneralization | - |
dc.title | Theoretical insights into mixup: perspectives on decision boundaries | - |
dc.title.alternative | 믹스업에 대한 이론적 통찰: 결정 경계의 관점으로 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :김재철AI대학원, | - |
dc.contributor.alternativeauthor | Yun, Chulhee | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.