Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Dian Shao

Dian Shao

3 indexed papers

Recent (6 mo)
3
With code
0
Influential cites
0
Benchmarked
0

Publications per year

3
26

Top categories

Vision×2Crypto×2AI×2ML×2

Frequent co-authors

Zedian Shao2×
Like Liu1×
Zhengzheng Xu1×
Haitao He1×
Hongzhe Li1×
Shuchang Zhang1×

Research Timeline

2026
Leave My Images Alone: Preventing Multi-Modal Large Language Models from Analyzing Images via Visual Prompt Injection

The paper introduces ImageProtector, a user-side method that embeds an imperceptible perturbation into images to prevent Multi-modal Large Language Models (MLLMs) from analyzing and extracting sensitive information from them.

Cordyceps: Covert Control Attacks on LLMs via Data Poisoning

The paper introduces 'covert control attacks,' a novel and stealthy data poisoning method that teaches LLMs an information hiding scheme, allowing malicious instructions to be encoded and decoded and bypassing existing defenses.

Knowing the Self, Understanding the World: A Dual-Cognition Benchmark for UAV Spatio-temporal Reasoning with MLLMs

The paper introduces UAV-DualCog, a benchmark for evaluating multimodal large language models in UAV scenarios for joint self-state and environment-state reasoning.

Highlighted terms show continued research focus across papers

Papers

cs.CVEmpiricalRecentJul 17, 2026

Knowing the Self, Understanding the World: A Dual-Cognition Benchmark for UAV Spatio-temporal Reasoning with MLLMs

Like Liu, Zhengzheng Xu, Haitao He, Hongzhe Li +2 more

The paper introduces UAV-DualCog, a benchmark for evaluating multimodal large language models in UAV scenarios for joint self-state and environment-state reasoning.

View →
cs.CRcs.AIcs.LGRecent
May 26, 2026

Cordyceps: Covert Control Attacks on LLMs via Data Poisoning

Zedian Shao, Charles Fleming, Teodora Baluta

The paper introduces 'covert control attacks,' a novel and stealthy data poisoning method that teaches LLMs an information hiding scheme, allowing malicious instructions to be encoded and decoded and…

View →
cs.CVcs.AIcs.CRRecentApr 10, 2026

Leave My Images Alone: Preventing Multi-Modal Large Language Models from Analyzing Images via Visual Prompt Injection

Zedian Shao, Hongbin Liu, Yuepeng Hu, Neil Zhenqiang Gong

The paper introduces ImageProtector, a user-side method that embeds an imperceptible perturbation into images to prevent Multi-modal Large Language Models (MLLMs) from analyzing and extracting sensiti…

View →