LADDER: Multi-Objective Backdoor Attack via Evolutionary Algorithm
Dazhuang Liu
Network and Distributed System Security (NDSS) Symposium 2025 · Day 3 · ML Backdoors
This talk introduces LADDER, a novel approach to crafting **backdoor attacks** in computer vision tasks that simultaneously optimizes multiple, often conflicting, attack objectives. Presented by Dazhuang Liu from Delft University of Technology, the research addresses a critical gap in existing backdoor methodologies: the inability of traditional gradient-based optimization techniques to robustly and stably achieve triggers that are both highly effective and exceptionally stealthy, while also maintaining the integrity of the benign model’s performance.
AI review
Competent ML security research that applies multi-objective evolutionary optimization to backdoor trigger generation — a genuinely useful framing that sidesteps the instability of Lagrange/SGD approaches. The contribution is real but incremental: frequency-domain backdoor triggers aren't new, EA-based optimization isn't new, and the combination, while technically sound, lands as a methodological refinement rather than a field-shifting result.