Search papers, labs, and topics across Lattice.
This paper investigates the adversarial robustness of open-source vision-language models (VLMs), LLaVA-v1.5-7B and Qwen2.5-VL-7B, in a simulated e-commerce environment. The study employs three gradient-based adversarial attacks (BIM, PGD, and a CLIP-based spectral attack) to assess the vulnerability of these agents. Results show LLaVA-v1.5-7B is highly susceptible to these attacks, while Qwen2.5-VL-7B demonstrates significantly greater robustness, highlighting crucial differences in adversarial resilience among VLM architectures.
Open-source VLMs can be easily fooled by simple gradient-based attacks, but the degree of vulnerability varies drastically across architectures.
We study adversarial robustness of open-source vision-language model (VLM) agents deployed in a self-contained e-commerce environment built to simulate realistic pre-deployment conditions. We evaluate two agents, LLaVA-v1.5-7B and Qwen2.5-VL-7B, under three gradient-based attacks: the Basic Iterative Method (BIM), Projected Gradient Descent (PGD), and a CLIP-based spectral attack. Against LLaVA, all three attacks achieve substantial attack success rates (52.6%, 53.8%, and 66.9% respectively), demonstrating that simple gradient-based methods pose a practical threat to open-source VLM agents. Qwen2.5-VL proves significantly more robust across all attacks (6.5%, 7.7%, and 15.5%), suggesting meaningful architectural differences in adversarial resilience between open-source VLM families. These findings have direct implications for the security evaluation of VLM agents prior to commercial deployment.
