Skip to main content

Table 5 Performance of GPBEAM with different perturbation generation algorithms

From: Perturbing BEAMs: EEG adversarial attack to deep learning models for epilepsy diagnosing

 

I-FGSM

MI-FGSM

DII-FGSM

PGD

C&W

\(\epsilon\)

0.1

0.3

0.5

0.1

0.3

0.5

0.1

0.3

0.5

0.1

0.3

0.5

0.1

0.3

0.5

Acc

0.75

0.58

0.50

0.57

0.47

0.43

0.69

0.54

0.46

0.52

0.43

0.41

0.49

0.40

0.34

SR

0.18

0.35

0.43

0.37

0.46

0.52

0.25

0.39

0.47

0.41

0.51

0.53

0.45

0.55

0.59

\({\mathrm{DL}}_{B}\)

0.09

0.22

0.33

0.10

0.22

0.35

0.09

0.21

0.34

0.08

0.23

0.37

0.10

0.26

0.45

Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us