Probabilistic Contrastive Test-Time Adaptation

Test-time adaptation (TTA) enhances generalization against out-of-distribution data during inference. Recent advances in TTA leverage some techniques such as contrastive learning and entropy minimization to improve the discriminability and robustness of models in target domains. However, existing me...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Proceedings of the ... IEEE International Conference on Acoustics, Speech and Signal Processing (1998) s. 1 - 5
Hlavní autori: You, Linjing, Lu, Jiabao, Huang, Xiayuan
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 06.04.2025
Predmet:
ISSN:2379-190X
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract Test-time adaptation (TTA) enhances generalization against out-of-distribution data during inference. Recent advances in TTA leverage some techniques such as contrastive learning and entropy minimization to improve the discriminability and robustness of models in target domains. However, existing methods often overlook simultaneous distribution shifts of sample and label, such as long-tail distributions, and contrastive learning approaches may require substantial storage for sample pairs. In this paper, we propose a novel Probabilistic Contrastive Test-time Adaptation (PCTA) method based on Expectation Maximization (EM), which is used to estimate the von Mises Fisher (vMF) distribution of test samples to capture both sample distribution and class proportions. The estimated distributions are used for probabilistic contrastive learning to adapt feature representations and optimize classification through class-weighted entropy minimization. Experimental results show that PCTA significantly enhances the performance across various distribution shifts and outperforms state-of-the-art methods in different scenarios involving both sample and label shifts. Code is available at https://github.com/youlj109/PCTA.
AbstractList Test-time adaptation (TTA) enhances generalization against out-of-distribution data during inference. Recent advances in TTA leverage some techniques such as contrastive learning and entropy minimization to improve the discriminability and robustness of models in target domains. However, existing methods often overlook simultaneous distribution shifts of sample and label, such as long-tail distributions, and contrastive learning approaches may require substantial storage for sample pairs. In this paper, we propose a novel Probabilistic Contrastive Test-time Adaptation (PCTA) method based on Expectation Maximization (EM), which is used to estimate the von Mises Fisher (vMF) distribution of test samples to capture both sample distribution and class proportions. The estimated distributions are used for probabilistic contrastive learning to adapt feature representations and optimize classification through class-weighted entropy minimization. Experimental results show that PCTA significantly enhances the performance across various distribution shifts and outperforms state-of-the-art methods in different scenarios involving both sample and label shifts. Code is available at https://github.com/youlj109/PCTA.
Author Lu, Jiabao
Huang, Xiayuan
You, Linjing
Author_xml – sequence: 1
  givenname: Linjing
  surname: You
  fullname: You, Linjing
  email: youlinjing2023@ia.ac.cn
  organization: Institute of Automation Chinese Academy of Sciences,Beijing,China
– sequence: 2
  givenname: Jiabao
  surname: Lu
  fullname: Lu, Jiabao
  email: lujb9921@mails.jlu.edu.cn
  organization: School of Communication Engineering Jilin University,Changchun,China
– sequence: 3
  givenname: Xiayuan
  surname: Huang
  fullname: Huang, Xiayuan
  email: xiayuan.huang@ia.ac.cn
  organization: Institute of Automation Chinese Academy of Sciences,Beijing,China
BookMark eNo1j1FLwzAUhaNMcJv7Bz7UH9B6k9wmuY-j6BQGDtaHvY2kvYPI1o42CP57C-rTd-DA4TsLMev6joV4klBICfT8Xq33-x2SMVAoUGUhwREoAzdiRZacLkEbZ1HeirnSlnJJcLgXi3H8BICpcHNR7IY--BDPcUyxyaq-S4Of4hdnNY8pr-OFs3Xrr8mn2HcP4u7kzyOv_rgU9etLXb3l24_NpLPNI-mUO4dSIVlLCKFErULL2pclGi_lRAbdBMtoDEnkYCfRlpsACLJFyye9FI-_s5GZj9chXvzwffy_p38Aok9F9A
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/ICASSP49660.2025.10890260
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISBN 9798350368741
EISSN 2379-190X
EndPage 5
ExternalDocumentID 10890260
Genre orig-research
GrantInformation_xml – fundername: National Natural Science Foundation of China
  funderid: 10.13039/501100001809
– fundername: Beijing Nova Program
  funderid: 10.13039/501100005090
GroupedDBID 23M
6IE
6IF
6IH
6IK
6IL
6IM
6IN
AAJGR
AAWTH
ABLEC
ACGFS
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IJVOP
IPLJI
M43
OCL
RIE
RIL
RIO
RNS
ID FETCH-LOGICAL-i93t-884124977940b5432bde3a5546a11a55e03cb7e466914eb7350decb0401d47ef3
IEDL.DBID RIE
IngestDate Wed Nov 19 08:26:49 EST 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i93t-884124977940b5432bde3a5546a11a55e03cb7e466914eb7350decb0401d47ef3
PageCount 5
ParticipantIDs ieee_primary_10890260
PublicationCentury 2000
PublicationDate 2025-April-6
PublicationDateYYYYMMDD 2025-04-06
PublicationDate_xml – month: 04
  year: 2025
  text: 2025-April-6
  day: 06
PublicationDecade 2020
PublicationTitle Proceedings of the ... IEEE International Conference on Acoustics, Speech and Signal Processing (1998)
PublicationTitleAbbrev ICASSP
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0008748
Score 2.2878242
Snippet Test-time adaptation (TTA) enhances generalization against out-of-distribution data during inference. Recent advances in TTA leverage some techniques such as...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Adaptation models
Contrastive learning
Distribution Shifts
Entropy
Heavily-tailed distribution
Label Shifts
Minimization
Probabilistic Contrastive Learning
Probabilistic logic
Robustness
Signal processing
Signal processing algorithms
Speech processing
Test-Time Adaptation
Title Probabilistic Contrastive Test-Time Adaptation
URI https://ieeexplore.ieee.org/document/10890260
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1La8MwDBZbGWO77JWxNxns6iyOnTg5lrKywSiF9tBb8UODXtKSpvv9k93HtsMOOznYDsEWkiNL3yeAJ4cy1ZqXDHmqmbSZY_QTa1hmycHVriCxB0m_q8GgnEyq4QasHrAwiBiSzzDxjyGW7-Z25a_KSMN9VKwgD31fqWIN1tqZ3VLJ8hAeNySaz2-97mg0lJ58krzALE-2L_8qoxJOkf7JP79_CtE3Hi8e7k6aM9jD-hyOf1AJXkBC4yaw5Xri5diTTjV66W1ZPCbDzzzUI-46vViH3iMY91_GvVe2qYXAZpVoWVmGKtGKtCc1uRSZcSi0zzDTnFOLqbBGoSyKiks0SuSpQ2tIQ7mTCj_EJXTqeY1XEBvBtfIBV46VtDQ5R8FNVmmFxpJzcQ2RX_l0sWa7mG4XffNH_y0c-f0N2SzFHXTaZoX3cGA_29myeQgy-gIyZpDj
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT8MwDLbQQDwuvDbxpkhcM5ombdrjNDFtYkyT1sNuUx5G2mWb9uD3E2cP4MCBU6oklZpYdurY32eAZ4cy1prnDHmsmbSJY_4n1rDEegdXu8yLPUi6q3q9fDgs-huwesDCIGJIPsM6PYZYvpvaFV2VeQ2nqFjmPfR9Kp1VrOFaO8ObK5kfwtOGRvOl02wMBn1J9JPeD0zS-vb1X4VUwjnSOv3nF5xB9RuRF_V3Z8057OHkAk5-kAleQt2Pm8CXS9TLEdFOzfWCrFlUetPPCOwRNZyerYPvVShbr2WzzTbVENi4EEuW56FOtPL6E5tUisQ4FJpyzDTnvsVYWKNQZlnBJRol0tihNV5HuZMKP0QNKpPpBK8gMoJrRSFXjoW0fnKKgpuk0AqN9e7FNVRp5aPZmu9itF30zR_9j3DULt-7o26n93YLx7TXIbclu4PKcr7Ceziwn8vxYv4Q5PUF_fSUMg
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=Proceedings+of+the+...+IEEE+International+Conference+on+Acoustics%2C+Speech+and+Signal+Processing+%281998%29&rft.atitle=Probabilistic+Contrastive+Test-Time+Adaptation&rft.au=You%2C+Linjing&rft.au=Lu%2C+Jiabao&rft.au=Huang%2C+Xiayuan&rft.date=2025-04-06&rft.pub=IEEE&rft.eissn=2379-190X&rft.spage=1&rft.epage=5&rft_id=info:doi/10.1109%2FICASSP49660.2025.10890260&rft.externalDocID=10890260