본문

서브메뉴

Understanding Generalization in Deep Learning Through Occam's Razor
Understanding Generalization in Deep Learning Through Occam's Razor
Understanding Generalization in Deep Learning Through Occam's Razor

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202103053
ISBN  
9798286424351
DDC  
519
저자명  
Lotfi, Sanae.
서명/저자  
Understanding Generalization in Deep Learning Through Occams Razor
발행사항  
[Sl] : New York University, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
260 p
주기사항  
Source: Dissertations Abstracts International, Volume: 86-12, Section: A.
주기사항  
Advisor: Gordon Wilson, Andrew.
학위논문주기  
Thesis (Ph.D.)--New York University, 2025.
초록/해제  
요약Gaining insight into the mechanisms behind the generalization of deep learning models is crucial to build on their strengths, address their limitations, and deploy them in safety-critical applications. As state-of-the-art models for various data modalities become increasingly large and are trained on internet-scale data, the notion of generalization becomes more challenging to define. In this thesis, I study generalization through the lens of Occam's razor: among models that can fit the training data, the simplest is most likely to perform well on unseen data. Compression bounds provide a principled way to capture this intuition through a trade-off between the model's training performance and its compressed size.First, I present our work on deriving state-of-the-art generalization bounds for image classification models, providing key insights into why these models generalize effectively in practice. I then explore the challenges of extending these bounds to pretrained large language models (LLMs), establishing the first non-vacuous bounds for LLMs. Our findings reveal that larger LLMs not only yield better bounds but also find simpler representations of the data. Furthermore, we demonstrate that LLMs retain their understanding of patterns but forget highly unstructured data more rapidly as we compress them more aggressively. Finally, I discuss the connection between generalization bounds and the marginal likelihood, a Bayesian tool used for model selection and hyperparameter tuning. Specifically, I demonstrate how generalization bounds can predict practical issues like overfitting and underfitting when using the marginal likelihood for model selection, and propose a remedy that is more aligned with generalization.
일반주제명  
Applied mathematics
일반주제명  
Statistics
일반주제명  
Computer science
일반주제명  
Information science
키워드  
Occam's razor
키워드  
Generalization
키워드  
Deep learning
키워드  
Large language models
키워드  
Compression bounds
기타저자  
New York University Center for Data Science
기본자료저록  
Dissertations Abstracts International. 86-12A.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017356871
■00520260202103053
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798286424351
■035    ▼a(MiAaPQ)AAI31932110
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a519
■1001  ▼aLotfi,  Sanae.
■24510▼aUnderstanding  Generalization  in  Deep  Learning  Through  Occam's  Razor
■260    ▼a[Sl]▼bNew  York  University▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a260  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  86-12,  Section:  A.
■500    ▼aAdvisor:  Gordon  Wilson,  Andrew.
■5021  ▼aThesis  (Ph.D.)--New  York  University,  2025.
■520    ▼aGaining  insight  into  the  mechanisms  behind  the  generalization  of  deep  learning  models  is  crucial  to  build  on  their  strengths,  address  their  limitations,  and  deploy  them  in  safety-critical  applications.  As  state-of-the-art  models  for  various  data  modalities  become  increasingly  large  and  are  trained  on  internet-scale  data,  the  notion  of  generalization  becomes  more  challenging  to  define.  In  this  thesis,  I  study  generalization  through  the  lens  of  Occam's  razor:  among  models  that  can  fit  the  training  data,  the  simplest  is  most  likely  to  perform  well  on  unseen  data.  Compression  bounds  provide  a  principled  way  to  capture  this  intuition  through  a  trade-off  between  the  model's  training  performance  and  its  compressed  size.First,  I  present  our  work  on  deriving  state-of-the-art  generalization  bounds  for  image  classification  models,  providing  key  insights  into  why  these  models  generalize  effectively  in  practice.  I  then  explore  the  challenges  of  extending  these  bounds  to  pretrained  large  language  models  (LLMs),  establishing  the  first  non-vacuous  bounds  for  LLMs.  Our  findings  reveal  that  larger  LLMs  not  only  yield  better  bounds  but  also  find  simpler  representations  of  the  data.  Furthermore,  we  demonstrate  that  LLMs  retain  their  understanding  of  patterns  but  forget  highly  unstructured  data  more  rapidly  as  we  compress  them  more  aggressively.  Finally,  I  discuss  the  connection  between  generalization  bounds  and  the  marginal  likelihood,  a  Bayesian  tool  used  for  model  selection  and  hyperparameter  tuning.  Specifically,  I  demonstrate  how  generalization  bounds  can  predict  practical  issues  like  overfitting  and  underfitting  when  using  the  marginal  likelihood  for  model  selection,  and  propose  a  remedy  that  is  more  aligned  with  generalization.
■590    ▼aSchool  code:  0146.
■650  4▼aApplied  mathematics
■650  4▼aStatistics
■650  4▼aComputer  science
■650  4▼aInformation  science
■653    ▼aOccam's  razor
■653    ▼aGeneralization
■653    ▼aDeep  learning
■653    ▼aLarge  language  models
■653    ▼aCompression  bounds
■690    ▼a0364
■690    ▼a0984
■690    ▼a0723
■690    ▼a0800
■690    ▼a0463
■71020▼aNew  York  University▼bCenter  for  Data  Science.
■7730  ▼tDissertations  Abstracts  International▼g86-12A.
■790    ▼a0146
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17356871▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF18630 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.