EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

W-DOE: Wasserstein Distribution-Agnostic Outlier Exposure

Qizhou Wang, Bo Han, Yang Liu, Chen Gong, Tongliang Liu, Jiming Liu

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2025-05 · 47 (5) · 3530-3545

DOI: 10.1109/tpami.2025.3531000 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025-05. Verfasst von Qizhou Wang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NSFC General Program, RGC Young Collaborative Research, Basic and Applied Basic Research Foundation of Guangdong Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In open-world environments, classification models should be adept at identifying out-of-distribution (OOD) data whose semantics differ from in-distribution (ID) data, leading to the emerging research in OOD detection. As a promising learning scheme, outlier exposure (OE) enables the models to learn from auxiliary OOD data, enhancing model representations in discerning between ID and OOD patterns. However, these auxiliary OOD data often do not fully represent real OOD scenarios, potentially biasing our models in practical OOD detection. Hence, we propose a novel OE-based learning method termed Wasserstein Distribution-agnostic Outlier Exposure (W-DOE), which is both theoretically sound and experimentally superior to previous works. The intuition is that by expanding the coverage of training-time OOD data, the models will encounter fewer unseen OOD cases upon deployment. In W-DOE, we achieve additional OOD data to enlarge the OOD coverage, based on a new data synthesis approach called implicit data synthesis (IDS). It is driven by our new insight that perturbing model parameters can lead to implicit data transformation, which is simple to implement yet effective to realize. Furthermore, we suggest a general learning framework to search for the synthesized OOD data that can benefit the models most, ensuring the OOD performance for the enlarged OOD coverage measured by the Wasserstein metric. Our approach comes with provable guarantees for open-world settings, demonstrating that broader OOD coverage ensures reduced estimation errors and thereby improved generalization for real OOD cases. We conduct extensive experiments across a series of representative OOD detection setups, further validating the superiority of W-DOE against state-of-the-art counterparts in the field.

Förderung

  • NSFC General Program
  • RGC Young Collaborative Research
  • Basic and Applied Basic Research Foundation of Guangdong Province
  • HKBU Faculty Niche Research Areas
  • HKBU CSD Departmental Incentive Scheme
  • National Natural Science Foundation of China
  • NSF for Distinguished Young Scholar of Jiangsu Province

Zitieren

APA

Wang, Q., Han, B., Liu, Y., Gong, C., Liu, T., & Liu, J. (2025). W-DOE: Wasserstein Distribution-Agnostic Outlier Exposure. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(5), 3530-3545. https://doi.org/10.1109/tpami.2025.3531000

Vancouver

Wang Q, Han B, Liu Y, Gong C, Liu T, Liu J. W-DOE: Wasserstein Distribution-Agnostic Outlier Exposure. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025;47(5):3530-3545. doi:10.1109/tpami.2025.3531000

BibTeX
@article{Wang2025,
  author = {Wang, Qizhou and Han, Bo and Liu, Yang and Gong, Chen and Liu, Tongliang and Liu, Jiming},
  title = {W-DOE: Wasserstein Distribution-Agnostic Outlier Exposure},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2025},
  volume = {47},
  number = {5},
  pages = {3530--3545},
  doi = {10.1109/tpami.2025.3531000}
}
Kostenlos starten →