Episodic Training for Domain Generalization

Da Li, Jianshu Zhang, Yongxin Yang, Cong Liu, Yi-Zhe Song, Timothy Hospedales

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

Abstract

Domain generalization (DG) is the challenging and topical problem of learning models that generalize to novel testing domains with different statistics than a set of known training domains. The simple approach of aggregating data from all source domains and training a single deep neural network end-to-end on all the data provides a surprisingly strong baseline that surpasses many prior published methods. In this paper we build on this strong baseline by designing an episodic training procedure that trains a single deep network in a way that exposes it to the domain shift that characterises a novel domain at runtime. Specifically, we decompose a deep network into feature extractor and classifier components, and then train each component by simulating it interacting with a partner who is badly tuned for the current domain. This makes both components more robust, ultimately leading to our networks producing state-of-the-art performance on three DG benchmarks. Furthermore, we consider the pervasive workflow of using an ImageNet trained CNN as a fixed feature extractor for downstream recognition tasks. Using the Visual Decathlon benchmark, we demonstrate that our episodic-DG training improves the performance of such a general purpose feature extractor by explicitly training a feature for robustness to novel problems. This shows that DG training can benefit standard practice in computer vision.
Original languageEnglish
Title of host publication2019 IEEE/CVF International Conference on Computer Vision (ICCV)
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1446-1445
Number of pages10
ISBN (Electronic)978-1-7281-4804-5
ISBN (Print)978-1-7281-4803-8
DOIs
Publication statusPublished - 27 Feb 2020
EventInternational Conference on Computer Vision 2019 - Seoul, Korea, Republic of
Duration: 27 Oct 20192 Nov 2019
http://iccv2019.thecvf.com/

Publication series

Name
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
ISSN (Print)1550-5499
ISSN (Electronic)2380-7504

Conference

ConferenceInternational Conference on Computer Vision 2019
Abbreviated titleICCV 2019
Country/TerritoryKorea, Republic of
CitySeoul
Period27/10/192/11/19
Internet address

Fingerprint

Dive into the research topics of 'Episodic Training for Domain Generalization'. Together they form a unique fingerprint.

Cite this