The paper introduces TrustFlip, an attack that exploits consistency‑based defenses in vehicular collaborative perception by deploying physical adversarial objects to create inconsistent observations among benign vehicles. This misattribution lowers the trust score of a targeted vehicle, leading to its exclusion from the collaboration and a degradation of perception performance. The authors evaluate the attack across multiple architectures, showing it can remove a benign vehicle in up to 87.7% of scenarios and reduce Average Precision by up to 13%, and propose a mitigation called TrustReflect that reduces the attack success rate by 35–100%.
By Yutong Liu, Chenyi Wang, Ming F. Li, Qingzhao Zhang
Camera-based object detectors are vulnerable to physical adversarial attacks designed to suppress detections. While adversarial training and input purification offer some protection, they often overfit to specific attack distributions and fail on adaptive adversaries.
Gradient Mirage is a defense for large language model split learning that disrupts the assumption that exposed gradients faithfully reflect the client’s full-label training objective, thereby thwarting gradient matching attacks. It introduces inconsistencies across objective, direction, and scale by using selective autoregressive supervision, scale blinding, and directional privatization with a von Mises-Fisher mechanism. The method preserves utility through dual-track backpropagation and bottom-gradient recovery, achieving a stronger privacy‑utility trade‑off in experiments.
By Shiyu Miao, Yunlong Mao, Zirui Huang, Liang Yao, Tianshuo Zheng, Yanhui Gu, Fan Liu, Sheng Zhong
arXiv:2608. 16031v1 Announce Type: new Abstract: Camera-based object detectors are vulnerable to physical adversarial attacks designed to suppress detections.
By Yuting Wu, Dongfang Guo, Xiangzhong Luo, Qun Song, Rui Tan
arXiv:2607. 11063v1 Announce Type: new Abstract: Despite progress in Embodied AI, Vision-and-Language Navigation systems remain vulnerable to adversarial visual disturbances.
By Chenyang Li, Kaige Li, Zeyu Jiang, Changhao Chen
FSPGD introduces a feature-space black-box attack for semantic segmentation that targets intermediate representations rather than just output logits. The method uses a dual loss: an external loss to disrupt cross-model feature alignment and an internal loss to reduce consistency among same-class instances. Experiments on Pascal VOC 2012 and Cityscapes show that FSPGD outperforms existing logit-level and segmentation-specific attacks across CNN and Transformer backbones, and its adversarial examples improve robustness when used for training.
By Eun-Sol Park, MiSo Park, Yong-Goo Shin
arXiv:2608. 07750v1 Announce Type: cross Abstract: Deep Neural Networks (DNNs) have found successful deployment in numerous vision perception systems.
By Cong Chen, Jean-Philippe Monteuuis, Jonathan Petit
arXiv:2610.08331v1 Announce Type: cross
Abstract: The rapid integration of Vision Language Models (VLMs) into sensitive systems introduces critical safety vulnerabilities that remain unexplored in ex...
By Heyam Bin Jahlan Areej Alhothali Abeer Alhothali
arXiv:2606. 02267v1 Announce Type: new Abstract: The vulnerability of deep neural networks to adversarial examples poses a significant challenge for real-world deployment.
By Nicolas Stalder, Benjamin F. Grewe, Matteo Saponati, Pau Vilimelis Aceituno
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.
By Ridma Jayasundara, Shaheer Mohamed, Tharindu Fernando, Harshala Gammulle, Basura Fernando, Sanka Rasnayake, A V Subramanyam, Sridha Sridharan, Clinton Fookes
The paper investigates how Vision Language Models (VLMs) can be fooled by small, human‑imperceptible changes to images. It introduces a gradient‑based attack that targets only the vision encoder, reducing computational cost while still effectively disrupting both untargeted and targeted multimodal alignment. Experiments on open‑source VLMs such as Qwen2.5‑VL, Granite‑Vision, FastVLM, and Phi‑3.5‑Vision demonstrate that these perturbations can significantly alter the models’ textual outputs.
By Ilan Zini, Boussad Addad, Katarzyna Kapusta
The paper presents a mechanistic defense for Vision‑Language‑Action (VLA) models against adversarial patches. By using a sparse autoencoder, the authors identify a feature whose activation correlates strongly with the presence of an adversarial patch and suppress this feature only when a linear probe detects an attack. This conditional intervention improves robustness on the LIBERO‑10 benchmark while avoiding the performance degradation that occurs with continuous suppression.
By Yukiya Horiba, Koshiro Aoki, Shunsuke Yasuki, Bum Jun Kim, Taiki Miyanishi