Bootstrap Generalization Ability from Loss Landscape Perspective

Huanran Chen, Shitong Shao, Ziyi Wang, Zirui Shang*, Jin Chen, Xiaofeng Ji, Xinxiao Wu

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

4 引用 (Scopus)

摘要

Domain generalization aims to learn a model that can generalize well on the unseen test dataset, i.e., out-of-distribution data, which has different distribution from the training dataset. To address domain generalization in computer vision, we introduce the loss landscape theory into this field. Specifically, we bootstrap the generalization ability of the deep learning model from the loss landscape perspective in four aspects, including backbone, regularization, training paradigm, and learning rate. We verify the proposed theory on the NICO++, PACS, and VLCS datasets by doing extensive ablation studies as well as visualizations. In addition, we apply this theory in the ECCV 2022 NICO Challenge1 and achieve the 3rd place without using any domain invariant methods.

源语言英语
主期刊名Computer Vision – ECCV 2022 Workshops, Proceedings
编辑Leonid Karlinsky, Tomer Michaeli, Ko Nishino
出版商Springer Science and Business Media Deutschland GmbH
500-517
页数18
ISBN(印刷版)9783031250743
DOI
出版状态已出版 - 2023
活动17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, 以色列
期限: 23 10月 202227 10月 2022

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
13806 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议17th European Conference on Computer Vision, ECCV 2022
国家/地区以色列
Tel Aviv
时期23/10/2227/10/22

指纹

探究 'Bootstrap Generalization Ability from Loss Landscape Perspective' 的科研主题。它们共同构成独一无二的指纹。

引用此