서브메뉴
검색
Contextures: The Mechanism of Representation Learning
Contextures: The Mechanism of Representation Learning
Detailed Information
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202103130
- ISBN
- 9798286448548
- DDC
- 004
- 저자명
- Zhai, Runtian.
- 서명/저자
- Contextures: The Mechanism of Representation Learning
- 발행사항
- [Sl] : Carnegie Mellon University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 166 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-01, Section: B.
- 주기사항
- Advisor: Ravikumar, Pradeep;Kolter, Zico.
- 학위논문주기
- Thesis (Ph.D.)--Carnegie Mellon University, 2025.
- 초록/해제
- 요약This dissertation establishes the contexture theory to mathematically characterize the mechanism of representation learning, also known as pretraining. Despite the remarkable empirical success of foundation models, it is not very clear what representations they learn, and why these representations are useful for various disparate downstream tasks. A scientific understanding of representation learning is critical, especially at this point when scaling up the model size is producing diminishing returns, and designing new pretraining methods is imperative for further progress.Prior work treated different representation learning methods quite differently, whereas the contexture theory provides a unified framework for delineating the representations these methods learn. The central argument is that a representation is learned from the association between the input X and a context variable A. We prove that if an encoder captures the maximum information of this association, in which case we say that the encoder learns the contexture, then it will be optimal on the class of tasks that are compatible with the context. We also show that a context is the most useful when the association between X and A is neither too strong nor too weak. The important implication of the contexture theory is that increasing the model size alone will achieve diminishing returns, and further advancements require better contexts.We demonstrate that lots of existing pretraining objectives can learn the contexture, including supervised learning, self-supervised learning, generative models, etc. Based on that, we introduce two general objectives-SVME and KISE, for learning the contexture. We also show how to mix multiple contexts together, which is an effortless way to create better contexts from existing ones. Then, we prove statistical learning bounds for representation learning, and extend the framework to spectrally transformed kernel regression for semi-supervised learning. Finally, we discuss the effect of the data distribution shift from pretraining to the downstream task.
- 일반주제명
- Computer science
- 일반주제명
- Computer engineering
- 키워드
- Learning theory
- 키워드
- Machine learning
- 키워드
- Downstream task
- 기타저자
- Carnegie Mellon University Computer Science
- 기본자료저록
- Dissertations Abstracts International. 87-01B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017357097
■00520260202103130
■006m o d
■007cr#unu||||||||
■020 ▼a9798286448548
■035 ▼a(MiAaPQ)AAI31939492
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a004
■1001 ▼aZhai, Runtian.▼0(orcid)0000-0003-3332-3466
■24510▼aContextures: The Mechanism of Representation Learning
■260 ▼a[Sl]▼bCarnegie Mellon University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a166 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-01, Section: B.
■500 ▼aAdvisor: Ravikumar, Pradeep;Kolter, Zico.
■5021 ▼aThesis (Ph.D.)--Carnegie Mellon University, 2025.
■520 ▼aThis dissertation establishes the contexture theory to mathematically characterize the mechanism of representation learning, also known as pretraining. Despite the remarkable empirical success of foundation models, it is not very clear what representations they learn, and why these representations are useful for various disparate downstream tasks. A scientific understanding of representation learning is critical, especially at this point when scaling up the model size is producing diminishing returns, and designing new pretraining methods is imperative for further progress.Prior work treated different representation learning methods quite differently, whereas the contexture theory provides a unified framework for delineating the representations these methods learn. The central argument is that a representation is learned from the association between the input X and a context variable A. We prove that if an encoder captures the maximum information of this association, in which case we say that the encoder learns the contexture, then it will be optimal on the class of tasks that are compatible with the context. We also show that a context is the most useful when the association between X and A is neither too strong nor too weak. The important implication of the contexture theory is that increasing the model size alone will achieve diminishing returns, and further advancements require better contexts.We demonstrate that lots of existing pretraining objectives can learn the contexture, including supervised learning, self-supervised learning, generative models, etc. Based on that, we introduce two general objectives-SVME and KISE, for learning the contexture. We also show how to mix multiple contexts together, which is an effortless way to create better contexts from existing ones. Then, we prove statistical learning bounds for representation learning, and extend the framework to spectrally transformed kernel regression for semi-supervised learning. Finally, we discuss the effect of the data distribution shift from pretraining to the downstream task.
■590 ▼aSchool code: 0041.
■650 4▼aComputer science
■650 4▼aComputer engineering
■653 ▼aFoundation models
■653 ▼aLearning theory
■653 ▼aMachine learning
■653 ▼aRepresentation learning
■653 ▼aDownstream task
■690 ▼a0984
■690 ▼a0800
■690 ▼a0464
■71020▼aCarnegie Mellon University▼bComputer Science.
■7730 ▼tDissertations Abstracts International▼g87-01B.
■790 ▼a0041
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357097▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
Preview
Export
ChatGPT Discussion
AI Recommended Related Books
פרט מידע
- הזמנה
- לא קיים
- התיקיה שלי
- צפה הראשון בקשה
- Non-Book Loan Application
- Nighttime Book Loan Application
Available after logging in.


