서브메뉴
검색
Essays on Fair Operations
Essays on Fair Operations
Detailed Information
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211151957
- ISBN
- 9798383705827
- DDC
- 519
- 저자명
- Xia, Shangzhou.
- 서명/저자
- Essays on Fair Operations
- 발행사항
- [Sl] : Columbia University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 165 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-02, Section: B.
- 주기사항
- Advisor: Balseiro, Santiago.
- 학위논문주기
- Thesis (Ph.D.)--Columbia University, 2024.
- 초록/해제
- 요약Fairness emerges as a vital concern to decision makers as crucial as efficiency, if not more important. Fair operations decisions are aimed at distributive justice in various scenarios. In this dissertation, we study two examples of distributively fair decision making in operations research, a dynamic fair allocation problem and a subpopulational robustness assessment problem for machine learning models.We first study a dynamic allocation problem in which T sequentially arriving divisible resources are to be allocated to a number of agents with concave utilities. The joint utility functions of each resource to the agents are drawn stochastically from a known joint distribution, independently and identically across time, and the central planner makes immediate and irrevocable allocation decisions. Most works on dynamic resource allocation aim to maximize the utilitarian welfare, i.e., the efficiency of the allocation, which may result in unfair concentration of resources on certain high-utility agents while leaving others' demands under-fulfilled. In this work, aiming at balancing efficiency and fairness, we instead consider a broad collection of welfare metrics, the Holder means, which includes the Nash social welfare and the egalitarian welfare. To this end, we first study a fluid-based policy derived from a deterministic surrogate to the underlying problem and show that for all smooth H\\"older mean welfare metrics it attains an O(1) regret over the time horizon length T against the hindsight optimum, i.e., the optimal welfare if all utilities were known in advance of deciding on allocations. However, when evaluated under the non-smooth egalitarian welfare, the fluid-based policy attains a regret of order Θ(√T). We then propose a new policy built thereupon, called Backward Infrequent Re-solving ($\\mathsf{BIR}$), which consists of re-solving the deterministic surrogate problem at most O(log T) times. We show under a mild regularity condition that it attains a regret against the hindsight optimal egalitarian welfare of order O(1) when all agents have linear utilities and O(log T) otherwise. We further propose the Backward Infrequent Re-solving with Thresholding (BIRT) policy, which enhances the BIR policy by thresholding adjustments and performs similarly without any assumption whatsoever. More specifically, we prove the BIRT policy attains an O(1) regret independently of the horizon length T when all agents have linear utilities and O(log2+ ϵ T) otherwise. We conclude by presenting numerical experiments to corroborate our theoretical claims and to illustrate the significant performance improvement against several benchmark policies.The performance of ML models degrades when the training population is different from that seen under operation. Towards assessing distributional robustness, we study the worst-case performance of a model over all subpopulations of a given size, defined with respect to core attributes Z. This notion of robustness can consider arbitrary (continuous) attributes Z, and automatically accounts for complex intersectionality in disadvantaged groups. We develop a scalable yet principled two-stage estimation procedure that can evaluate the robustness of state-of-the-art models. We prove that our procedure enjoys several finite-sample convergence guarantees, including dimension-free convergence. Instead of overly conservative notions based on Rademacher complexities, our evaluation error depends on the dimension of Z only through the out-of-sample error in estimating the performance conditional on Z. On real datasets, we demonstrate that our method certifies the robustness of a model and prevents deployment of unreliable models.
- 일반주제명
- Applied mathematics
- 기타저자
- Columbia University Business
- 기본자료저록
- Dissertations Abstracts International. 86-02B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017162306
■00520250211151957
■006m o d
■007cr#unu||||||||
■020 ▼a9798383705827
■035 ▼a(MiAaPQ)AAI31329272
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a519
■1001 ▼aXia, Shangzhou.
■24510▼aEssays on Fair Operations
■260 ▼a[Sl]▼bColumbia University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a165 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-02, Section: B.
■500 ▼aAdvisor: Balseiro, Santiago.
■5021 ▼aThesis (Ph.D.)--Columbia University, 2024.
■520 ▼aFairness emerges as a vital concern to decision makers as crucial as efficiency, if not more important. Fair operations decisions are aimed at distributive justice in various scenarios. In this dissertation, we study two examples of distributively fair decision making in operations research, a dynamic fair allocation problem and a subpopulational robustness assessment problem for machine learning models.We first study a dynamic allocation problem in which T sequentially arriving divisible resources are to be allocated to a number of agents with concave utilities. The joint utility functions of each resource to the agents are drawn stochastically from a known joint distribution, independently and identically across time, and the central planner makes immediate and irrevocable allocation decisions. Most works on dynamic resource allocation aim to maximize the utilitarian welfare, i.e., the efficiency of the allocation, which may result in unfair concentration of resources on certain high-utility agents while leaving others' demands under-fulfilled. In this work, aiming at balancing efficiency and fairness, we instead consider a broad collection of welfare metrics, the Holder means, which includes the Nash social welfare and the egalitarian welfare. To this end, we first study a fluid-based policy derived from a deterministic surrogate to the underlying problem and show that for all smooth H\\"older mean welfare metrics it attains an O(1) regret over the time horizon length T against the hindsight optimum, i.e., the optimal welfare if all utilities were known in advance of deciding on allocations. However, when evaluated under the non-smooth egalitarian welfare, the fluid-based policy attains a regret of order Θ(√T). We then propose a new policy built thereupon, called Backward Infrequent Re-solving ($\\mathsf{BIR}$), which consists of re-solving the deterministic surrogate problem at most O(log T) times. We show under a mild regularity condition that it attains a regret against the hindsight optimal egalitarian welfare of order O(1) when all agents have linear utilities and O(log T) otherwise. We further propose the Backward Infrequent Re-solving with Thresholding (BIRT) policy, which enhances the BIR policy by thresholding adjustments and performs similarly without any assumption whatsoever. More specifically, we prove the BIRT policy attains an O(1) regret independently of the horizon length T when all agents have linear utilities and O(log2+ ϵ T) otherwise. We conclude by presenting numerical experiments to corroborate our theoretical claims and to illustrate the significant performance improvement against several benchmark policies.The performance of ML models degrades when the training population is different from that seen under operation. Towards assessing distributional robustness, we study the worst-case performance of a model over all subpopulations of a given size, defined with respect to core attributes Z. This notion of robustness can consider arbitrary (continuous) attributes Z, and automatically accounts for complex intersectionality in disadvantaged groups. We develop a scalable yet principled two-stage estimation procedure that can evaluate the robustness of state-of-the-art models. We prove that our procedure enjoys several finite-sample convergence guarantees, including dimension-free convergence. Instead of overly conservative notions based on Rademacher complexities, our evaluation error depends on the dimension of Z only through the out-of-sample error in estimating the performance conditional on Z. On real datasets, we demonstrate that our method certifies the robustness of a model and prevents deployment of unreliable models.
■590 ▼aSchool code: 0054.
■650 4▼aApplied mathematics
■653 ▼aFair operations decisions
■653 ▼aNash social welfare
■653 ▼aState-of-the-art models
■690 ▼a0796
■690 ▼a0364
■71020▼aColumbia University▼bBusiness.
■7730 ▼tDissertations Abstracts International▼g86-02B.
■790 ▼a0054
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17162306▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
Preview
Export
ChatGPT Discussion
AI Recommended Related Books
Подробнее информация.
- Бронирование
- не существует
- моя папка
- Первый запрос зрения
- Non-Book Loan Application
- Nighttime Book Loan Application
Available after logging in.


