arXiv AI September 10, 2026 By Jian-Ping Mei, Weibin Zhang, Jie Chen, Xuyun Zhang, Tiantian Zhu

D-ADD: An Effective Plug-In for Defending Against Model Stealing

The paper introduces D-ADD, a plug‑in defense for image classification models that protects against model‑stealing attacks. It uses a non‑parametric detector called Account‑aware Distribution Discrepancy (ADD) to identify malicious queries by modeling each class as a multivariate normal distribution and computing weighted distribution discrepancies. With an enhanced version ADD$^+$ that handles domain shifts and combined with random‑based prediction poisoning, D-ADD offers strong protection while minimally affecting benign users in both soft‑ and hard‑label settings.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

computer-vision

arXiv Machine Learning

Aug 6

One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP

arXiv:2505. 19840v3 Announce Type: replace-cross Abstract: Deep Neural Networks (DNNs) have achieved widespread success yet remain prone to adversarial attacks.

By Binyan Xu, Xilin Dai, Di Tang, Kehuan Zhang

llms multimodal safety

arXiv Machine Learning

Jun 8

ADAGE: Active Defenses Against GNN Extraction

arXiv:2503. 00065v4 Announce Type: replace-cross Abstract: Graph Neural Networks (GNNs) achieve high performance in various real-world applications, such as drug discovery, traffic states prediction, and recommendation systems.

By Jing Xu, Franziska Boenisch, Adam Dziedzic

rag benchmarks

arXiv Computer Vision

6d ago

Region-Level Black-Box Defense Against Stealthy Embedding-Space Backdoors in CLIP

arXiv:2609.31558v1 Announce Type: new Abstract: Contrastive Language--Image Pretraining (CLIP) has emerged as a dominant vision backbone due to its strong transferability and zero-shot capabilities....

By Ahmed Abdelnaby, Mohamed Elmahallawy

rag safety

arXiv Machine Learning

Aug 3

Low-Cost Hard-Label Adversarial Attack with Theoretical Foundations

arXiv:2601. 14300v4 Announce Type: replace Abstract: Hard-label black-box attacks, relying solely on top-1 predictions, represent one of the most challenging yet practically threat models.

By Jun Liu, Leo Yu Zhang, Fengpeng Li, Isao Echizen, Jiantao Zhou

computer-vision benchmarks safety

Hugging Face Trending Papers

Aug 17

DSPrompt: Dynamic Soft Prompt Defense Against M-RAG Corruption

Multimodal Retrieval Augmented Generation (M-RAG) is increasingly vulnerable to adversarial attacks where malicious data are crafted to produce embeddings that align with benign entries in the vector...

rag multimodal benchmarks safety

arXiv Machine Learning

Jul 7

VLMGuard: Bootstrapping Malicious Prompt Detectors from Unlabeled Vision-Language Prompts in the Wild

arXiv:2410. 00296v2 Announce Type: replace Abstract: Vision-language Models (VLMs) are essential for contextual understanding of both visual and textual information.

By Junlin Fang, Wenyu Chen, Reshmi Ghosh, Robert Sim, Ahmed Salem, Vitor R. Carvalho, Emily Lawton, Sharon Li, Jack W. Stokes, Sean Du

llms multimodal benchmarks

One email a morning, machine-written

Related stories

One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP

ADAGE: Active Defenses Against GNN Extraction

Region-Level Black-Box Defense Against Stealthy Embedding-Space Backdoors in CLIP

Low-Cost Hard-Label Adversarial Attack with Theoretical Foundations

DSPrompt: Dynamic Soft Prompt Defense Against M-RAG Corruption

VLMGuard: Bootstrapping Malicious Prompt Detectors from Unlabeled Vision-Language Prompts in the Wild