2023
DOI: 10.1007/s10994-023-06307-y
|View full text |Cite
|
Sign up to set email alerts
|

Reducing classifier overconfidence against adversaries through graph algorithms

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
1
0

Year Published

2023
2023
2023
2023

Publication Types

Select...
1

Relationship

0
1

Authors

Journals

citations
Cited by 1 publication
(1 citation statement)
references
References 27 publications
0
1
0
Order By: Relevance
“…While there are no prior works in the context of link prediction, a few works have tackled the issue for GNNbased node classification. Teixeira et al [25] first pointed to the issue of miscalibration of GNNs. Wang et al [7] highlighted an overall underconfident behavior and proposed CaGCN, a method that leverages node embeddings from auxiliary GNN temperature scales, outperforming off-the-shelf calibration strategies.…”
Section: Related Workmentioning
confidence: 99%
“…While there are no prior works in the context of link prediction, a few works have tackled the issue for GNNbased node classification. Teixeira et al [25] first pointed to the issue of miscalibration of GNNs. Wang et al [7] highlighted an overall underconfident behavior and proposed CaGCN, a method that leverages node embeddings from auxiliary GNN temperature scales, outperforming off-the-shelf calibration strategies.…”
Section: Related Workmentioning
confidence: 99%