all AI news
InstructTA: Instruction-Tuned Targeted Attack for Large Vision-Language Models
June 27, 2024, 4:47 a.m. | Xunguang Wang, Zhenlan Ji, Pingchuan Ma, Zongjie Li, Shuai Wang
cs.CV updates on arXiv.org arxiv.org
Abstract: Large vision-language models (LVLMs) have demonstrated their incredible capability in image understanding and response generation. However, this rich visual interaction also makes LVLMs vulnerable to adversarial examples. In this paper, we formulate a novel and practical targeted attack scenario that the adversary can only know the vision encoder of the victim LVLM, without the knowledge of its prompts (which are often proprietary for service providers and not publicly available) and its underlying large language model …
arxiv cs.cv instruction-tuned language language models replace type vision vision-language vision-language models
More from arxiv.org / cs.CV updates on arXiv.org
PlaNet-S: Automatic Semantic Segmentation of Placenta
1 day, 23 hours ago |
arxiv.org
Continuous 3D Myocardial Motion Tracking via Echocardiography
1 day, 23 hours ago |
arxiv.org
Optimal Transport Aggregation for Visual Place Recognition
1 day, 23 hours ago |
arxiv.org
Jobs in AI, ML, Big Data
VP, Enterprise Applications
@ Blue Yonder | Scottsdale
Data Scientist - Moloco Commerce Media
@ Moloco | Redwood City, California, United States
Senior Backend Engineer (New York)
@ Kalepa | New York City. Hybrid
Senior Backend Engineer (USA)
@ Kalepa | New York City. Remote US.
Senior Full Stack Engineer (USA)
@ Kalepa | New York City. Remote US.
Senior Full Stack Engineer (New York)
@ Kalepa | New York City., Hybrid