arXiv:2509. 23689v2 Announce Type: replace Abstract: Model Merging (MM) has proven to be an effective alternative to multi-task learning, where several fine-tuned models are merged, without access to the tasks' training data, into one model that retains performance across different tasks.
The paper introduces Inverse Knowledge Distillation (IKD), an attack‑agnostic technique that enhances adversarial transferability by maximizing the discrepancy between benign and adversarial prediction distributions on a surrogate model. IKD employs a CE/KL‑equivalent soft‑label objective to push adversarial predictions away from a fixed benign anchor, leveraging Fisher‑sensitive surrogate directions. The authors provide theoretical analysis showing CE and KL induce identical gradients, derive a lower bound on Fisher‑subspace overlap, and demonstrate through extensive ImageNet experiments that IKD consistently improves black‑box attack performance across CNN, ViT, and defended models.
arXiv:2609.10002v1 Announce Type: new Abstract: Deepfake detectors remain vulnerable to transfer-based black-box attacks, in which adversarial examples are generated on a source surrogate model and t...
The paper investigates whether pixels alone can determine an image’s origin—human, AI class, or specific generator—under adversarial edits. It establishes a minimax limit: the best possible robust acceptance gap equals the minimum total‑variation distance between the target distribution and attacked source distributions, independent of verifier design. The study also shows that practical public verifiers can fail before reaching this theoretical ceiling, highlighting the need to evaluate both statistical limits and deployed verifier behavior separately.
arXiv:2605.25663v2 Announce Type: replace-cross Abstract: Black-box adversarial attacks that minimize only the ground-truth confidence suffer from class drift: perturbations wander through the featur...
The paper introduces Fast Preemptive Robustification (FPR), a lightweight defense that enhances the robustness of deep neural networks against transferable adversarial attacks. By sharpening Laplacian responses through a single 3×3 channel‑wise convolution, FPR eliminates the need for surrogate models, iterative optimization, or specialized training. Experiments show that FPR lowers untargeted attack success rates by 12.7% and reduces targeted attack success from 10.7% to 4.1%.
The paper introduces Meta-Adaptive Multimodal Jailbreaking (MAMJ), a method that jointly optimizes an attack strategy prompt and attacker weights to generate more effective jailbreaks against vision‑language models. Using an LLM‑based critique to refine the strategy and group‑level success‑rate rewards to update the weights, MAMJ achieves high attack success rates on MM‑SafetyBench, outperforming existing baselines by up to 24.1 percentage points. The learned attacker also transfers to unseen models and remains robust against typical defenses, highlighting a systemic vulnerability in current VLMs.
arXiv:2608. 06674v1 Announce Type: cross Abstract: Adversarial vulnerabilities remain a major concern for the safe deployment of neural networks, particularly in object detection, a core task embedded in many safety-critical systems.
arXiv:2608. 11815v1 Announce Type: new Abstract: Transfer-based adversarial attacks craft adversarial examples using surrogate models to mislead black-box victim models.
arXiv:2606. 15308v1 Announce Type: new Abstract: While multimodal large language models (MLLMs) have shown strong visual reasoning abilities, serving a large model for every query is computationally expensive.
arXiv:2607. 26574v2 Announce Type: replace-cross Abstract: Safety classifiers ("guards") are the dominant black-box defense for vision-language models, yet a guard judges an input's surface form, not its meaning: a harmful request re-encoded as set theory, formal logic, a classical language, code, or text rendered inside an image slips past a guard that would block it in plain language - the decode gap.
arXiv:2606. 15441v1 Announce Type: cross Abstract: Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution.