Adaptive adversarial pattern contrast algorithm for black-box model and domain attack

doi:10.1016/j.engappai.2025.111980

Adaptive adversarial pattern contrast algorithm for black-box model and domain attack

2025-08-16

0

PRE

AI

DOI：10.1016/j.engappai.2025.111980

Original

Original request for help

Save

Abstract

En 中文

• Adaptive Pattern Contrast crafts adversarial examples cross models and domains. • The similarity contrast loss aligns noise patterns and separates clean data. • It merges clean low-frequency data with high-frequency adversarial features. • It significantly boosts success rates across various models and domains.

Keywords:

Adaptive Pattern Contrast

Similarity Contrast Loss

Cross-model Adversarial Examples

Domain Adaptation

Feature Separation

Adaptive adversarial pattern contrast algorithm for black-box model and domain attack

Abstract

Journal

Organization