Not All Adversarial Examples Require a Complex Defense: Identifying Over-optimized Adversarial Examples with IQR-based Logit Thresholding

Poster (Scientific congresses and symposiums)

Ozbulak, Utku; Van Messem, Arnout; De Neve, Wesley

2019 • 2019 International Joint Conference on Neural Networks (IJCNN)

Permalink
https://hdl.handle.net/2268/265033

Utku_Ozbulak_19_IJCNN_Poster r1.pptx

Publisher postprint (2.38 MB)

All documents in ORBi are protected by a user license.

Disciplines :

Mathematics
Computer science

Author, co-author :

Ozbulak, Utku

Van Messem, Arnout ; Université de Liège - ULiège > Département de mathématique > Statistique applquée aux sciences

De Neve, Wesley

Language :

English

Title :

Not All Adversarial Examples Require a Complex Defense: Identifying Over-optimized Adversarial Examples with IQR-based Logit Thresholding

Publication date :

2019

Event name :

2019 International Joint Conference on Neural Networks (IJCNN)

Event place :

Budapest, Hungary

Event date :

14-19 July 2019

Audience :

International

Available on ORBi :

since 15 November 2021

Number of views

48 (3 by ULiège)

Number of downloads

2 (0 by ULiège)