Target robust discriminant analysis

Wouter M. Kouw, Marco Loog

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

Abstract

Often, in practice, the data distribution at test time differs, to a smaller or larger extent, from that of the original training data. Consequentially, the so-called source classifier, trained on the labeled data, deteriorates on the test, or target data. Domain adaptive classifiers aim to alleviate this problem, but typically assume a particular type of domain shift. Most are not robust to violations of domain shift assumptions and may perform even worse than the non-adaptive source classifier. We construct robust parameter estimators for discriminant analysis that guarantee performance improvements of the adaptive classifier over the source classifier.
Original languageEnglish
Title of host publicationIAPR International Workshop on Statistical Pattern Recognition
PublisherSpringer
Number of pages10
Publication statusAccepted/In press - 21 Dec 2020

Keywords

  • Domain adaptation
  • Robust estimators

Fingerprint Dive into the research topics of 'Target robust discriminant analysis'. Together they form a unique fingerprint.

Cite this