본문

서브메뉴

Learning Representations Toward the Understanding of Out-Of-Distribution for Neural Networks
Learning Representations Toward the Understanding of Out-Of-Distribution for Neural Networ...
Learning Representations Toward the Understanding of Out-Of-Distribution for Neural Networks

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105541
ISBN  
9798263395490
DDC  
388.312
저자명  
Kwon, Gukyeong.
서명/저자  
Learning Representations Toward the Understanding of Out-Of-Distribution for Neural Networks
발행사항  
[Sl] : Georgia Institute of Technology, 2021
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2021
형태사항  
125 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: A.
주기사항  
Advisor: Alregib, Ghassan.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2021.
초록/해제  
요약Data-driven representations achieve powerful generalization performance in diverse information processing tasks. However, the generalization is often limited to test data from the same distribution as training data (in-distribution (ID)). In addition, the neural networks often make overconfident and incorrect predictions for data outside training distribution, called out-of-distribution (OOD). In this dissertation, we develop representations that can characterize OOD for the neural networks and utilize the characterization to efficiently generalize to OOD.We categorize the data-driven representations based on information flow in neural networks and develop novel gradient-based representations. In particular, we utilize the backpropagated gradients to represent what the neural networks has not learned in the data. The capability of gradient-based representations for OOD characterization is comprehensively analyzed in comparison with standard activation-based representations. We also develop a regularization technique for the gradient-based representations to better characterize OOD. We develop an anomaly detection algorithm named GradCon using the gradient constraint and achieve state-of-the-art performance in OOD detection.We also propose activation-based representations learned with auxiliary information to efficiently generalize to data from OOD. We use an unsupervised learning framework to learn the aligned representations of visual and attribute data. This aligned representation are utilized to calibrate the overconfident prediction toward ID classes. The generalization performance of the aligned representations is validated in the application of generalized zero-shot learning. Our developed GZSL method, GatingAE, achieves state-of-the-art performance in generalizing to OOD without using labeled OOD data. Also, balanced performance for both ID and OOD is achieved by mitigating the prediction bias presented in the network. Finally, GatingAE requires significantly less number of model parameters compared to other state-of-the-art methods.
일반주제명  
Automobile parking
일반주제명  
Wavelet transforms
일반주제명  
Back propagation
일반주제명  
Fourier transforms
일반주제명  
Street signs
일반주제명  
Neural networks
일반주제명  
Traffic control
일반주제명  
Information processing
일반주제명  
Data compression
일반주제명  
Computer science
일반주제명  
Mathematics
일반주제명  
Transportation
일반주제명  
Urban planning
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05A.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2021        us                              c    eng  d
■001000017360524
■00520260202105541
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263395490
■035    ▼a(MiAaPQ)AAI32315129
■035    ▼a(MiAaPQ)GeorgiaTech65045
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a388.312
■1001  ▼aKwon,  Gukyeong.
■24510▼aLearning  Representations  Toward  the  Understanding  of  Out-Of-Distribution  for  Neural  Networks
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2021
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2021
■300    ▼a125  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  A.
■500    ▼aAdvisor:  Alregib,  Ghassan.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2021.
■520    ▼aData-driven  representations  achieve  powerful  generalization  performance  in  diverse  information  processing  tasks.  However,  the  generalization  is  often  limited  to  test  data  from  the  same  distribution  as  training  data  (in-distribution  (ID)).  In  addition,  the  neural  networks  often  make  overconfident  and  incorrect  predictions  for  data  outside  training  distribution,  called  out-of-distribution  (OOD).  In  this  dissertation,  we  develop  representations  that  can  characterize  OOD  for  the  neural  networks  and  utilize  the  characterization  to  efficiently  generalize  to  OOD.We  categorize  the  data-driven  representations  based  on  information  flow  in  neural  networks  and  develop  novel  gradient-based  representations.  In  particular,  we  utilize  the  backpropagated  gradients  to  represent  what  the  neural  networks  has  not  learned  in  the  data.  The  capability  of  gradient-based  representations  for  OOD  characterization  is  comprehensively  analyzed  in  comparison  with  standard  activation-based  representations.  We  also  develop  a  regularization  technique  for  the  gradient-based  representations  to  better  characterize  OOD.  We  develop  an  anomaly  detection  algorithm  named  GradCon  using  the  gradient  constraint  and  achieve  state-of-the-art  performance  in  OOD  detection.We  also  propose  activation-based  representations  learned  with  auxiliary  information  to  efficiently  generalize  to  data  from  OOD.  We  use  an  unsupervised  learning  framework  to  learn  the  aligned  representations  of  visual  and  attribute  data.  This  aligned  representation  are  utilized  to  calibrate  the  overconfident  prediction  toward  ID  classes.  The  generalization  performance  of  the  aligned  representations  is  validated  in  the  application  of  generalized  zero-shot  learning.  Our  developed  GZSL  method,  GatingAE,  achieves  state-of-the-art  performance  in  generalizing  to  OOD  without  using  labeled  OOD  data.  Also,  balanced  performance  for  both  ID  and  OOD  is  achieved  by  mitigating  the  prediction  bias  presented  in  the  network.  Finally,  GatingAE  requires  significantly  less  number  of  model  parameters  compared  to  other  state-of-the-art  methods.
■590    ▼aSchool  code:  0078.
■650  4▼aAutomobile  parking
■650  4▼aWavelet  transforms
■650  4▼aBack  propagation
■650  4▼aFourier  transforms
■650  4▼aStreet  signs
■650  4▼aNeural  networks
■650  4▼aTraffic  control
■650  4▼aInformation  processing
■650  4▼aData  compression
■650  4▼aComputer  science
■650  4▼aMathematics
■650  4▼aTransportation
■650  4▼aUrban  planning
■690    ▼a0800
■690    ▼a0984
■690    ▼a0405
■690    ▼a0709
■690    ▼a0999
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05A.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2021
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360524▼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
    TF14779 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.