서브메뉴
검색
Epistemic Limits of Trustworthy Machine Learning
Epistemic Limits of Trustworthy Machine Learning
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202104726
- ISBN
- 9798265409294
- DDC
- 519
- 서명/저자
- Epistemic Limits of Trustworthy Machine Learning
- 발행사항
- [Sl] : Harvard University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 293 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
- 주기사항
- Advisor: du Pin Calmon, Flavio.
- 학위논문주기
- Thesis (Ph.D.)--Harvard University, 2025.
- 초록/해제
- 요약Theoretical understanding of a system's limits has long driven technological breakthroughs. Carnot delineated the fundamental limits of heat engine efficiency, paving the way for the design of modern state-of-the-art engines. More than a century later, Claude Shannon unraveled the fundamental limit of communication, known as channel capacity. This insight revolutionized communication systems, enabling continual improvements that ultimately led to wireless communication as we know it today. This thesis discusses the epistemic limits of machine learning (ML) and leverages them to improve the trustworthiness of ML systems. ML models have an epistemic limit when proving one of their properties is impossible. Epistemic refers to the impossibility of providing theoretical guarantees (knowledge) about a model's property. Epistemic limits are information-theoretic converse results on the hypothesis test that checks a model's property. First, we prove a limit on how much information personalized models can use while ensuring reliable test for performance gains across all users -- epistemic limits of personalization. We leverage this limit to develop a tool to help with feature selection. Second, we show a limit for reliably testing if model performance is equitable across multiple demographic groups --epistemic limit of fairness testing. We exploit this limit to design a metric for efficient algorithmic bias detection. Third, we prove a limit for testing if one model outperforms another on average -- epistemic limit of model selection. We use this result to delineate the set of indistinguishably good models --Rashomon set. Finally, we argue that the epistemic limits in model selection imply that explaining the predictions of ML models is necessary. Then, we develop efficient methods for explaining the content produced by large language models.
- 일반주제명
- Applied mathematics
- 일반주제명
- Statistics
- 키워드
- Explainability
- 키워드
- Fairness
- 기타저자
- Harvard University Engineering and Applied Sciences - Applied Math
- 기본자료저록
- Dissertations Abstracts International. 87-05B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017358612
■00520260202104726
■006m o d
■007cr#unu||||||||
■020 ▼a9798265409294
■035 ▼a(MiAaPQ)AAI32122281
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a519
■1001 ▼aMonteiro Paes, Lucas W.
■24510▼aEpistemic Limits of Trustworthy Machine Learning
■260 ▼a[Sl]▼bHarvard University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a293 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-05, Section: B.
■500 ▼aAdvisor: du Pin Calmon, Flavio.
■5021 ▼aThesis (Ph.D.)--Harvard University, 2025.
■520 ▼aTheoretical understanding of a system's limits has long driven technological breakthroughs. Carnot delineated the fundamental limits of heat engine efficiency, paving the way for the design of modern state-of-the-art engines. More than a century later, Claude Shannon unraveled the fundamental limit of communication, known as channel capacity. This insight revolutionized communication systems, enabling continual improvements that ultimately led to wireless communication as we know it today. This thesis discusses the epistemic limits of machine learning (ML) and leverages them to improve the trustworthiness of ML systems. ML models have an epistemic limit when proving one of their properties is impossible. Epistemic refers to the impossibility of providing theoretical guarantees (knowledge) about a model's property. Epistemic limits are information-theoretic converse results on the hypothesis test that checks a model's property. First, we prove a limit on how much information personalized models can use while ensuring reliable test for performance gains across all users -- epistemic limits of personalization. We leverage this limit to develop a tool to help with feature selection. Second, we show a limit for reliably testing if model performance is equitable across multiple demographic groups --epistemic limit of fairness testing. We exploit this limit to design a metric for efficient algorithmic bias detection. Third, we prove a limit for testing if one model outperforms another on average -- epistemic limit of model selection. We use this result to delineate the set of indistinguishably good models --Rashomon set. Finally, we argue that the epistemic limits in model selection imply that explaining the predictions of ML models is necessary. Then, we develop efficient methods for explaining the content produced by large language models.
■590 ▼aSchool code: 0084.
■650 4▼aApplied mathematics
■650 4▼aStatistics
■653 ▼aExplainability
■653 ▼aFairness
■653 ▼aHypothesis testing
■653 ▼aInformation theory
■653 ▼aPredictive multiplicity
■690 ▼a0364
■690 ▼a0463
■690 ▼a0800
■71020▼aHarvard University▼bEngineering and Applied Sciences - Applied Math.
■7730 ▼tDissertations Abstracts International▼g87-05B.
■790 ▼a0084
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17358612▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
