“It’s not that life has dreams, but that dreams create life.”
「人生に夢があるのではなく、夢が人生をつくるのです」
— Taeko Uzuki

🚀 About the Protagonist of 「青春との戦い」

Eddy Luo (罗威迪 & 一ノ瀬 エイジ), an incoming PhD student at University of Georgia advised by Prof.Xiang Zhen, his research interests:

• Trustworthy AI & AI Safety: Using interpretability methods, discover security vulnerabilities in AI systems, including foundation models and AI agents, and develop corresponding defense and detection algorithms, including safety alignment strategies.
• AI in Security: Leverage AI to drive defense and attack strategies on systems, including web system and operating system.
• Lifelong AI Algorithms: Develop lifelong learning AI frameworks and defense systems by utilizing reinforcement learning, cognitive science, bio-inspired algorithms, active learning, and so on.

Before his PhD journey, he had a research internship on AI safety at Prof.Chaowei Xiao’s Lab at University of Wisconsin-Madison. In pursuit of academic excellence, he has served as a research assistant at OSU NLP Group and ICICLE Institute under the mentorship of Prof.Yu Su. This valuable experience has significantly contributed to his understanding of natural language processing and multimodalities.

His CV is here: Eddy’s Curriculum Vitae. Eddy welcomes collaboration opportunities. Please feel free to contact him at Email: luo.1455[shift+2]uga[dot]edu. どうぞよろしくお願いします!

📰 Eddy’s News

• 2025.05.15🎉, Two of our works, AGrail: A Lifelong Agent Guardrail with Effective and Adaptive Safety Detection and Disentangling Memory and Reasoning Ability in Large Language Models have been accepted by ACL’2025 main conference. Thanks to my collaborators.
• 2025.04.15🎉, Our work JailBreakV-28K: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks wins $20,000 SafeBench Prize for Advancing MultiModal Large Language Model Security Benchmarking from Center for AI Safety.
• 2025.04.09🎉, I will join the University of Georgia as a PhD student in August 2025.
• 2025.01.22🎉, Our paper Dynamic Guided and Domain Applicable Safeguards for Enhanced Security in Large Language Models has been accepted by NAACL’2025 findings. I served as the first author. Thanks to my collaborators.
• 2024.07.16🎉, Our paper Bringing Back the Context: Camera Trap Species Identification as Link Prediction on Multimodal Knowledge Graphs has been accepted by CIKM’2024. Thanks to my collaborators.
• 2024.07.09🎉, Our paper JailBreakV-28K: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks has been accepted by COLM’2024. I served as the first author. Thanks to my collaborators.

🎓 Eddy’s Story

• 2025.08 - Present, P.hD. in Computer Science, School of Computing, University of Georgia
• 2022.09 - 2025.06, B.A. in Computer and Information Science, Art and Science College, The Ohio State University (Dean’s List of Art and Science College in 22 Fall and 24 SP)

📝 Eddy’s Pre-print

Arxiv

Weidi Luo*, Qiming Zhang*, Tianyu Lu*, Xiaogeng Liu, Yue Zhao, Zhen Xiang, Chaowei Xiao, Doxing via the Lens: Revealing Privacy Leakage in Image Geolocation for Agentic Multi-Modal Large Reasoning Model, [Paper]

Arxiv

Siyuan Ma*, Weidi Luo*, Yu Wang, Xiaogeng Liu, Muhao Chen, Bo Li, Chaowei Xiao, Visual-RolePlay: Universal Jailbreak Attack on MultiModal Large Language Models via Role-playing Image Character, [Paper] [Code].

Arxiv

Zeru Shi, Zhenting Wang, Yongye Su, Weidi Luo, Hang Gao, Fan Yang, Ruixiang Tang, Yongfeng Zhang, Auto-Prompt Generation Is Not Robustness: Prompt Optimization Driven by Adversarial Training, [Paper] [Code].

📸 Eddy’s Publication

ACL'2025

Weidi Luo, Shenghong Dai, Xiaogeng Liu, Suman Banerjee, Huan Sun, Muhao Chen, Chaowei Xiao, AGrail: A Lifelong Agent Guardrail with Effective and Adaptive Safety Detection, [Paper] [Project Page] [Code]

NAACL'2025

Weidi Luo*, He Cao*, Yu Wang, Zijing Liu, Aidan Wong, Bin Feng, Yuan Yao, Yu Li, Dynamic Guided and Domain Applicable Safeguards for Enhanced Security in Large Language Models, [Paper] [Code].

COLM'2024

Weidi Luo*, Siyuan Ma*, Xiaogeng Liu*, Xiaoyu Guo, Chaowei Xiao, JailBreakV-28K: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks, [Paper] [Project Page] [Code].

ACL'2025

Mingyu Jin, Weidi Luo, Sitao Cheng, Xinyi Wang, Wenyue Hua, Ruixiang Tang, William Yang Wang, Yongfeng Zhang, Disentangling Memory and Reasoning Ability in Large Language Models, [Paper] [Code]

CIKM'2024

Vardaan Pahuja, Weidi Luo, Yu Gu, Cheng-Hao Tu, Hong-You Chen, Tanya Berger-Wolf, Charles Stewart, Song Gao, Wei-Lun Chao, Yu Su, Bringing Back the Context: Camera Trap Species Identification as Link Prediction on Multimodal Knowledge Graphs, [Paper] [Project Page] [Code].

🎖 Eddy’s Award

• 2025.04, SafeBench Award from Center for AI Safety

🔍 Eddy’s Reviewer Experience

• ACL’2025, EMNLP’2025

💻 Eddy’s Internship Experience

• 2023.12 - Present, SaFoLab, Madison, Wisconsin, United States
• 2024.07 - 2024.10, International Digital Economy Academy(IDEA), Shenzhen, Guangdong, China
• 2022.08 - 2023.12, ICICLE Institute｜OSU NLP Group, Columbus, Ohio, United States