DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 윤철희 | - |
dc.contributor.author | Cho, Hanseul | - |
dc.contributor.author | 조한슬 | - |
dc.date.accessioned | 2024-07-25T19:30:47Z | - |
dc.date.available | 2024-07-25T19:30:47Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1045735&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/320547 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 김재철AI대학원, 2023.8,[iv, 53 p. :] | - |
dc.description.abstract | however, there are few theoretical results on this approach for minimax algorithms, especially outside the easier-to-analyze (strongly-)monotone setups. To narrow this gap, we study the convergence bounds of SGDA with random reshuffling (SGDA-RR) for smooth nonconvex-nonconcave objectives with Polyak-Łojasiewicz (PŁ) geometry. We analyze both simultaneous and alternating SGDA-RR for nonconvex-PŁ and primal-PŁ-PŁ objectives, and obtain convergence upper bounds faster than with-replacement SGDA. Our rates also extend to mini-batch SGDA-RR, recovering known rates for full-batch gradient descent-ascent (GDA). Lastly, we present a comprehensive lower bound for two-time-scale GDA, which matches the full-batch rate for primal-PŁ-PŁ case. | - |
dc.description.abstract | Stochastic gradient descent-ascent (SGDA) is one of the main workhorses for solving finite-sum minimax optimization problems. Most practical implementations of SGDA randomly reshuffle components and sequentially use them (i.e., without-replacement sampling) | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 최대최소화▼a확률적 경사 하강 상승법▼a비복원추출▼a셔플링▼a폴랴크-워야시에비치(PŁ) 조건 | - |
dc.subject | Minimax optimization▼aSGDA▼aWithout-replacement sampling▼aRandom reshuffling▼aPolyak-Łojasiewicz | - |
dc.title | Improved convergence rate of sgda by shuffling: focusing on the nonconvex-PŁ minimax problems | - |
dc.title.alternative | 셔플링을 이용한 확률적 경사 하강 상승법의 수렴 속도 분석: 비볼록-PŁ 최대최소화 문제를 중심으로 | - |
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.