QR-Code

AdvSCOD: Bayesian-Based Out-Of-Distribution Detection via Curvature Sketching and Adversarial Sample Enrichment

Detecting out-of-distribution (OOD) samples is critical for the deployment of deep neural networks (DNN) in real-world scenarios. An appealing direction in which to conduct OOD detection is to measure the epistemic uncertainty in DNNs using the Bayesian model, since it is much more explainable. SCOD...

Ausführliche Beschreibung

Gespeichert in:
Hauptverfasser: Jiacheng Qiao, Chengzhi Zhong, Peican Zhu, Keke Tang
Format: Artigo
Sprache:Inglês
Veröffentlicht: MDPI AG 2023-01-01
Schriftenreihe:Mathematics
Schlagworte:
OOD detection
Bayesian posterior
uncertainty
adversarial perturbation
sample enrichment
Online-Zugang:https://www.mdpi.com/2227-7390/11/3/692
Tags: Tag hinzufügen
Keine Tags, Fügen Sie das erste Tag hinzu!

Ähnliche Einträge