서브메뉴
검색
Robust Machine Learning for Biomedical Data: Efficiency, Reliability, and Generalizability
Robust Machine Learning for Biomedical Data: Efficiency, Reliability, and Generalizability
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211150950
- ISBN
- 9798383566725
- DDC
- 616
- 저자명
- You, Chenyu.
- 서명/저자
- Robust Machine Learning for Biomedical Data: Efficiency, Reliability, and Generalizability
- 발행사항
- [Sl] : Yale University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 273 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-01, Section: B.
- 주기사항
- Advisor: Duncan, James S.
- 학위논문주기
- Thesis (Ph.D.)--Yale University, 2024.
- 초록/해제
- 요약In the rapidly growing area of machine learning (ML), there is profound promise in crafting intelligent, data-driven methods for diverse real-world applications. Yet, in safety-critical domains like healthcare, some fundamental challenges remain. First, the insufficiency of raw biomedical data emphasizes the need for data-efficient and robust learning approaches.Second, the imperative of safety and stability necessitates a cohesive framework that unifies learning with theoretical guarantees.Third, the inherent heterogeneity and distribution shifts in real-world clinical data call for robust and generalizable learning methods.Amid the challenges we face, my research vision is to establish a solid foundation for healthcare and machine learning that enables the sustained and trustworthy deployments of artificial intelligence (AI) systems within the complex landscape of real-world clinical settings. Fundamentally, my research agenda is centered around building robust, reliable and equitable biomedical AI systems, that not only efficiently harness the large amount of raw biomedical data, but also accurately reconstruct anatomical structures (e.g., organs, bones, tumors) from sensory signals, intelligently process imperfect biomedical data (e.g., images, text, signals), and are more crucially anchored with theoretical grounding.To this end, I systematically studies representation learning for medical computing and analysis, standing out in three key aspects. First, I formulate the problem from practical and theoretical perspectives, ensuring a comprehensive understanding. Second, I seamlessly integrate the theoretical foundations into state-of-the-art algorithms, not only improving benchmark performance but also providing certifiable theoretical guarantees, where unified medical- and learning- theoretic analysis is just beginning to take shape. Third, my work equips medical AI agents with the capability to generalize across diverse scenarios, significantly broadening their applicability and effectiveness.I have incorporated ideas from infant development into designing mechanisms that enable medical AI systems to explore and understand their environment through experimentation. The findings indicate that my approach to learning empowers the AI agent to seamlessly adapt to unseen clinical settings and leverage its accumulated knowledge to efficiently tackle new clinical challenges.
- 일반주제명
- Medical imaging
- 일반주제명
- Computer engineering
- 키워드
- Machine Learning
- 기타저자
- Yale University Electrical Engineering
- 기본자료저록
- Dissertations Abstracts International. 86-01B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017160286
■00520250211150950
■006m o d
■007cr#unu||||||||
■020 ▼a9798383566725
■035 ▼a(MiAaPQ)AAI30993047
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a616
■1001 ▼aYou, Chenyu.
■24510▼aRobust Machine Learning for Biomedical Data: Efficiency, Reliability, and Generalizability
■260 ▼a[Sl]▼bYale University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a273 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-01, Section: B.
■500 ▼aAdvisor: Duncan, James S.
■5021 ▼aThesis (Ph.D.)--Yale University, 2024.
■520 ▼aIn the rapidly growing area of machine learning (ML), there is profound promise in crafting intelligent, data-driven methods for diverse real-world applications. Yet, in safety-critical domains like healthcare, some fundamental challenges remain. First, the insufficiency of raw biomedical data emphasizes the need for data-efficient and robust learning approaches.Second, the imperative of safety and stability necessitates a cohesive framework that unifies learning with theoretical guarantees.Third, the inherent heterogeneity and distribution shifts in real-world clinical data call for robust and generalizable learning methods.Amid the challenges we face, my research vision is to establish a solid foundation for healthcare and machine learning that enables the sustained and trustworthy deployments of artificial intelligence (AI) systems within the complex landscape of real-world clinical settings. Fundamentally, my research agenda is centered around building robust, reliable and equitable biomedical AI systems, that not only efficiently harness the large amount of raw biomedical data, but also accurately reconstruct anatomical structures (e.g., organs, bones, tumors) from sensory signals, intelligently process imperfect biomedical data (e.g., images, text, signals), and are more crucially anchored with theoretical grounding.To this end, I systematically studies representation learning for medical computing and analysis, standing out in three key aspects. First, I formulate the problem from practical and theoretical perspectives, ensuring a comprehensive understanding. Second, I seamlessly integrate the theoretical foundations into state-of-the-art algorithms, not only improving benchmark performance but also providing certifiable theoretical guarantees, where unified medical- and learning- theoretic analysis is just beginning to take shape. Third, my work equips medical AI agents with the capability to generalize across diverse scenarios, significantly broadening their applicability and effectiveness.I have incorporated ideas from infant development into designing mechanisms that enable medical AI systems to explore and understand their environment through experimentation. The findings indicate that my approach to learning empowers the AI agent to seamlessly adapt to unseen clinical settings and leverage its accumulated knowledge to efficiently tackle new clinical challenges.
■590 ▼aSchool code: 0265.
■650 4▼aMedical imaging
■650 4▼aComputer engineering
■653 ▼aMachine Learning
■653 ▼aMedical Image Analysis
■690 ▼a0574
■690 ▼a0800
■690 ▼a0464
■71020▼aYale University▼bElectrical Engineering.
■7730 ▼tDissertations Abstracts International▼g86-01B.
■790 ▼a0265
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17160286▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


