.png){kind=icon}
Return
Task-driven attention attack: An enhanced adversarial framework for long text
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.128834.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Existing attacks fail on long texts due to inefficient scoring mechanism. • TDAA: Novel task-driven attack framework for long texts. • Integrates contextual semantics & task-specific attention. • Adaptive length threshold + statistically non-essential sentence elimination. • 34.7 % ↑ ASR & 40 % ↑ efficiency vs baselines.
Journal
IF:
7.5
Papers:
3.0W
Citations:
10.2W
Organization
No organization information available
Cited Papers
ShareSaveShareSaveWordChange: Adversarial Examples Generation Approach for Chinese Text Classification.png)
IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSave no more
no more