Saved in:
Bibliographic Details
Main Authors: Wang, Haoliang, Zhao, Chen, Chen, Feng
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.07392
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912115550322688
author Wang, Haoliang
Zhao, Chen
Chen, Feng
author_facet Wang, Haoliang
Zhao, Chen
Chen, Feng
contents Open-set domain generalization addresses a real-world challenge: training a model to generalize across unseen domains (domain generalization) while also detecting samples from unknown classes not encountered during training (open-set recognition). However, most existing approaches tackle these issues separately, limiting their practical applicability. To overcome this limitation, we propose a unified framework for open-set domain generalization by introducing Feature-space Semantic Invariance (FSI). FSI maintains semantic consistency across different domains within the feature space, enabling more accurate detection of OOD instances in unseen domains. Additionally, we adopt a generative model to produce synthetic data with novel domain styles or class labels, enhancing model robustness. Initial experiments show that our method improves AUROC by 9.1% to 18.9% on ColoredMNIST, while also significantly increasing in-distribution classification accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2411_07392
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Feature-Space Semantic Invariance: Enhanced OOD Detection for Open-Set Domain Generalization
Wang, Haoliang
Zhao, Chen
Chen, Feng
Computer Vision and Pattern Recognition
Artificial Intelligence
Open-set domain generalization addresses a real-world challenge: training a model to generalize across unseen domains (domain generalization) while also detecting samples from unknown classes not encountered during training (open-set recognition). However, most existing approaches tackle these issues separately, limiting their practical applicability. To overcome this limitation, we propose a unified framework for open-set domain generalization by introducing Feature-space Semantic Invariance (FSI). FSI maintains semantic consistency across different domains within the feature space, enabling more accurate detection of OOD instances in unseen domains. Additionally, we adopt a generative model to produce synthetic data with novel domain styles or class labels, enhancing model robustness. Initial experiments show that our method improves AUROC by 9.1% to 18.9% on ColoredMNIST, while also significantly increasing in-distribution classification accuracy.
title Feature-Space Semantic Invariance: Enhanced OOD Detection for Open-Set Domain Generalization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2411.07392