Zhaoyi Li

MS student, ISE, University of Chinese Academy of Sciences

Joined

October 2024

Names

Zhaoyi Li (Preferred)

Emails

****@mbzuai.ac.ae (Confirmed)

****@gmail.com (Confirmed)

Personal Links

Google Scholar

Semantic Scholar

Career & Education History

MS student

ISE, University of Chinese Academy of Sciences (ucas.edu)

2022 – 2026

Undergrad student

ISE, Shandong University (sdu.edu.cn)

2018 – 2022

Advisors, Relations & Conflicts

Coauthor

Xiaohan Zhao

2025 – Present

Coauthor

Yaxin Luo

2025 – Present

PhD Advisor

Zhiqiang Shen

2024 – Present

Coauthor

Jiacheng Cui

2024 – Present

Expertise

Dataset distillation

2024 – Present

Multimodal Large Language Models

2024 – Present

Publications

Dataset Distillation via Committee Voting
Jiacheng Cui, Zhaoyi Li, Xiaochen Ma, Xinyue Bi, Yaxin Luo, Zhiqiang Shen
- Submitted to ICLR 2026
- Readers: Everyone
$\texttt{M-Attack-V2}$: Pushing the Frontier of Black-Box LVLM Attacks via Fine-Grained Detail Targeting
Xiaohan Zhao, Zhaoyi Li, Yaxin Luo, Jiacheng Cui, Zhiqiang Shen
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1
Zhaoyi Li, Xiaohan Zhao, Dong-Dong Wu, Jiacheng Cui, Zhiqiang Shen
- ICML 2025 R2-FM Workshop Poster
- Readers: Everyone
A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1
Zhaoyi Li, Xiaohan Zhao, Dong-Dong Wu, Jiacheng Cui, Zhiqiang Shen
- NeurIPS 2025 poster
- Readers: Everyone
M-Attack-V2: Pushing the Frontier of Black-Box LVLM Attacks via Fine-Grained Detail Targeting
Xiaohan Zhao, Zhaoyi Li, Yaxin Luo, Zhiqiang Shen
- Submitted to NeurIPS 2025
- Readers: Everyone
Open CaptchaWorld: A Comprehensive Web-based Platform for Testing and Benchmarking Multimodal LLM Agents
Yaxin Luo, Zhaoyi Li, Jiacheng Liu, Jiacheng Cui, Xiaohan Zhao, Zhiqiang Shen
- NeurIPS 2025 Datasets and Benchmarks Track poster
- Readers: Everyone
A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1
Zhaoyi Li, Xiaohan Zhao, Dong-Dong Wu, Jiacheng Cui, Zhiqiang Shen
- CoRR 2025
- Readers: Everyone
Dataset Distillation via Committee Voting
Jiacheng Cui, Zhaoyi Li, Xiaochen Ma, Xinyue Bi, Yaxin Luo, Zhiqiang Shen
- CoRR 2025
- Readers: Everyone

Zhaoyi Li

Names

Emails

Personal Links

Career & Education History

Advisors, Relations & Conflicts

Expertise

Publications

Co-Authors