Vision Token Manipulation Attacks on Cloud-Edge Inference of Large Vision-Language Models
cs.CR, cs.AI, cs.CV
Submitted: 2026-07-02
Updated: 2026-07-02
Code: https://github.com/superkevingit/Vision-Token-Manipulation-Attack
Terminology
Sources
- Qwen2.5-VL Technical Report
- On the Robustness of Split Learning against Adversarial Attacks
- Explaining and Harnessing Adversarial Examples
- Hyperion: Low-Latency Ultra-HD Video Analytics via Collaborative Vision Transformer Inference
- BackdoorVLM: A Benchmark for Backdoor Attacks and Defenses on Vision-Language Models
- MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts
- edgeVLM: Cloud-edge Collaborative Real-time VLM based on Context Transfer
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills
- Token Composition: A Graph Based on EVM Logs