Certain and consistent domain adaptation

Bhadrinath Nagabandi, Andrew Dudley, Hemanth Demakethepalli Venkateswara, Sethuraman Panchanathan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Unsupervised domain adaptation algorithms seek to transfer knowledge from labeled source datasets in order to predict the labels for target datasets in the presence of domain-shift. In this paper we propose the Certain and Consistent Domain Adaptation (CCDA) model for unsupervised domain adaptation. The CCDA aligns the source and target domains using adversarial training and reduces the domain adaptation problem to a semi supervised learning (SSL) problem. We estimate the target labels using consistency regularization and entropy minimization on the domain-aligned target samples whose predictions are consistent across multiple stochastic perturbations. We evaluate the CCDA on benchmark datasets and demonstrate that it outperforms competitive baselines from domain adaptation literature.

Original languageEnglish (US)
Title of host publicationSmart Multimedia - 2nd International Conference, ICSM 2019, Revised Selected Papers
EditorsTroy McDaniel, Stefano Berretti, Igor D.D. Curcio, Anup Basu
PublisherSpringer
Pages341-356
Number of pages16
ISBN (Print)9783030544065
DOIs
StatePublished - 2020
Event2nd International Conference on Smart Multimedia, ICSM 2019 - San Diego, United States
Duration: Dec 16 2019Dec 18 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12015 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Smart Multimedia, ICSM 2019
CountryUnited States
CitySan Diego
Period12/16/1912/18/19

Keywords

  • Consistency regularization
  • Domain adaptation
  • Entropy regularization
  • Semi supervised learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Certain and consistent domain adaptation'. Together they form a unique fingerprint.

  • Cite this

    Nagabandi, B., Dudley, A., Demakethepalli Venkateswara, H., & Panchanathan, S. (2020). Certain and consistent domain adaptation. In T. McDaniel, S. Berretti, I. D. D. Curcio, & A. Basu (Eds.), Smart Multimedia - 2nd International Conference, ICSM 2019, Revised Selected Papers (pp. 341-356). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12015 LNCS). Springer. https://doi.org/10.1007/978-3-030-54407-2_29