Supplementary Results for Table 2 (b) and Fig 3. di Chen Lin (95910)

Soggetti:

Attack success rates (%) of adversarial attacks against six models. ^* indicates the white-box attacks. di Chen Lin (95910)

Soggetti:

Logarithm of loss and attack success rates (%) of Diff-AdaNAG, Diff-PGD, and PGD on various numbers of iterations. di Chen Lin (95910)

Soggetti:

Attack success rates (%) of adversarial attacks against twelve models. ^* indicates the white-box attacks. di Chen Lin (95910)

Soggetti:

Comparison of SI-NI-FGSM and SI+Ours. di Chen Lin (95910)

Soggetti:

Supplementary results for Table 3 (b). di Chen Lin (95910)

Soggetti:

Visualization of adversarial samples. di Chen Lin (95910)

Soggetti:

Attack success rates (%) of adversarial attacks against twelve models. ^* indicates the white-box attacks. di Chen Lin (95910)

Soggetti:

Attack success rates (%) of adversarial attacks against twelve models. ^* indicates the white-box attacks. di Chen Lin (95910)

Soggetti:

Supplementary 4esults for Table 3 (a). di Chen Lin (95910)

Soggetti:

Attack success rates (%) of adversarial attacks against twelve models. ^* indicates the white-box attacks. di Chen Lin (95910)

Soggetti:

Supplementary Results for Table 2 (a). di Chen Lin (95910)

Soggetti:

Supplementary results for Table 3 (c). di Chen Lin (95910)

Soggetti:

Our proposed Diff-AdaNAG framework. di Chen Lin (95910)

Soggetti:

Supplementary results for Table 4. di Chen Lin (95910)

Soggetti:

Supplementary Results for Table 2 (c). di Chen Lin (95910)

Soggetti: