서브메뉴
검색
Distributionally Robust Machine Learning
Distributionally Robust Machine Learning
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152745
- ISBN
- 9798342107174
- DDC
- 616.07
- 저자명
- Sagawa, Shiori.
- 서명/저자
- Distributionally Robust Machine Learning
- 발행사항
- [Sl] : Stanford University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 212 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-04, Section: B.
- 주기사항
- Advisor: Liang, Percy.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2024.
- 초록/해제
- 요약Machine learning models can unexpectedly fail in the wild due to distribution shifts:mismatches in data distributions between training and deployment. Distribution shifts are often unavoidable and pose significant reliability challenges in many real-world applications. For example, models can fail on certain subpopulations (e.g., language models can fail on non-English languages) and on new domains unseen during training (e.g., medical models can fail on new hospitals).In this thesis, we aim to build reliable machine learning models that are robust to distribution shifts in the wild. In the first part, we mitigate subpopulation shifts by developing methods that leverage distributionally robust optimization (DRO). These methods overcome the computational and statistical obstacles of applying DRO on modern neural networks and on real-world shifts. In the second part, to tackle domain shifts, we introduce WILDS, a benchmark of real-world shifts, and show that existing methods fail on WILDS even though they perform well on synthetic shifts from prior benchmarks. We then develop a method that successfully mitigates real-world domain shifts. We propose an alternative to domain invariance---a key principle behind the prior methods---to reflect the structure of real-world shifts, and our method achieves state-of-the-art performance on multiple WILDS datasets.Altogether, the algorithms developed in this thesis mitigate real-world distribution shifts by addressing key statistical and computational challenges of training robust models, while anchoring to and leveraging the structure of real-world distribution shifts. These algorithms successfully improve robustness to a wide range of distribution shifts in the wild, from subpopulation shifts in language modeling to domain shiftsin wildlife monitoring and histopathology to spurious correlations.
- 일반주제명
- Histopathology
- 일반주제명
- Deep learning
- 일반주제명
- Demographics
- 일반주제명
- Benchmarks
- 일반주제명
- Demography
- 일반주제명
- Pathology
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 86-04B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017163727
■00520250211152745
■006m o d
■007cr#unu||||||||
■020 ▼a9798342107174
■035 ▼a(MiAaPQ)AAI31520286
■035 ▼a(MiAaPQ)Stanfordjw920ff8426
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a616.07
■1001 ▼aSagawa, Shiori.
■24510▼aDistributionally Robust Machine Learning
■260 ▼a[Sl]▼bStanford University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a212 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-04, Section: B.
■500 ▼aAdvisor: Liang, Percy.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2024.
■520 ▼aMachine learning models can unexpectedly fail in the wild due to distribution shifts:mismatches in data distributions between training and deployment. Distribution shifts are often unavoidable and pose significant reliability challenges in many real-world applications. For example, models can fail on certain subpopulations (e.g., language models can fail on non-English languages) and on new domains unseen during training (e.g., medical models can fail on new hospitals).In this thesis, we aim to build reliable machine learning models that are robust to distribution shifts in the wild. In the first part, we mitigate subpopulation shifts by developing methods that leverage distributionally robust optimization (DRO). These methods overcome the computational and statistical obstacles of applying DRO on modern neural networks and on real-world shifts. In the second part, to tackle domain shifts, we introduce WILDS, a benchmark of real-world shifts, and show that existing methods fail on WILDS even though they perform well on synthetic shifts from prior benchmarks. We then develop a method that successfully mitigates real-world domain shifts. We propose an alternative to domain invariance---a key principle behind the prior methods---to reflect the structure of real-world shifts, and our method achieves state-of-the-art performance on multiple WILDS datasets.Altogether, the algorithms developed in this thesis mitigate real-world distribution shifts by addressing key statistical and computational challenges of training robust models, while anchoring to and leveraging the structure of real-world distribution shifts. These algorithms successfully improve robustness to a wide range of distribution shifts in the wild, from subpopulation shifts in language modeling to domain shiftsin wildlife monitoring and histopathology to spurious correlations.
■590 ▼aSchool code: 0212.
■650 4▼aHistopathology
■650 4▼aDeep learning
■650 4▼aDemographics
■650 4▼aBenchmarks
■650 4▼aDemography
■650 4▼aPathology
■690 ▼a0800
■690 ▼a0938
■690 ▼a0571
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g86-04B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17163727▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
