Information-Geometric Inverse Distillation for Enhancing Adversarial Transferability
Quick summary
arXiv:2502.17003v2 Announce Type: replace-cross Abstract: Transfer-based adversarial attacks rely on surrogate models to craft perturbations, yet often overfit the surrogate's decision boundary. To address this problem, we propose Inverse Knowledge Distillation (IKD), a simple and attack-agnostic mechanism that maximizes the prediction-distribution discrepancy between benign and adversarial samples on the surrogate model. IKD uses a CE/KL-equivalent soft-label objective to push adversarial predictions away from a fixed benign prediction anchor and enrich the attack with Fisher-sensitive surrog
Key takeaways
- arXiv:2502.17003v2 Announce Type: replace-cross Abstract: Transfer-based adversarial attacks rely on surrogate models to craft perturbations, yet often overfit the surrogate's decision boundary.
- To address this problem, we propose Inverse Knowledge Distillation (IKD), a simple and attack-agnostic mechanism that maximizes the prediction-distribution discrepancy between benign and adversarial samples on the surrogate model.
- IKD uses a CE/KL-equivalent soft-label objective to push adversarial predictions away from a fixed benign prediction anchor and enrich the attack with Fisher-sensitive surrog
Why it matters
This model development creates a new option for users and a new testing obligation for developers. A fixed evaluation set comparing quality, cost and failure behavior is more useful than launch claims.
Member comments