Shuang Wu

Shuang Wu

Ph.D. Student at UConn · LLM Security & Reliability

I am a Ph.D. student in Computer Engineering at the University of Connecticut (UConn), advised by Prof. Yuan Hong. I joined UConn in January 2026.

My research focuses on LLM security and reliability, including jailbreak attacks and defenses, unified evaluation, and knowledge boundaries. I have also worked on cross-modal knowledge distillation and multimodal emotion recognition.

Previously, I received my B.Eng. degree in Computer Science and Technology from South China University of Technology (SCUT) in July 2025.

Seeking a Summer 2027 research internship. I am available May 10–August 22, 2027. Please reach out at shuang.wu@uconn.edu.

🔥 News

  • 2026.10:  🎉 Our SoK on LLM prompt security was accepted to IEEE S&P 2027!
  • 2026.01:  🎉🎉 I arrived in Storrs, CT and started my Ph.D. journey at UConn!
  • 2025.07:  🎓 I graduated from South China University of Technology (SCUT).
  • 2025.04:  🎉🎉 One paper accepted to IJCAI 2025!

📝 Publications

IEEE S&P 2027
PromptSecurity connects taxonomies, datasets, and unified evaluation of LLM attacks and defenses

SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and Defenses

Hanbin Hong*, Shuang Wu*, Shuya Feng, Nima Naderloui, Shenao Yan, Jingyu Zhang, Ali Arastehfard, Heqing Huang, Yuan Hong

IEEE Symposium on Security and Privacy (S&P), 2027 · Co-first author

Paper / Code / Dataset / Leaderboard

  • Introduces PromptSecurity, an open-source evaluation platform, alongside JailbreakDB and PromptSecurity-Eval for reproducible benchmarking.
  • Evaluates 11 target models, 20 attacks, nine defenses, and three automated judges to study configuration sensitivity and safety–utility trade-offs.
IJCAI 2025
Overview of the cross-modal knowledge distillation framework for emotion recognition

A Cross-Modal Densely Guided Knowledge Distillation Based on Modality Rebalancing Strategy for Enhanced Unimodal Emotion Recognition

Shuang Wu, Heng Liang, Yong Zhang, Yanlin Chen, Ziyu Jia

International Joint Conference on Artificial Intelligence (IJCAI), 2025 · First author

PDF

  • This paper proposes a cross-modal densely guided knowledge distillation framework to enhance unimodal emotion recognition performance by leveraging multimodal knowledge.

📖 Education

  • 2026.01–Present, Ph.D. Student in Computer Engineering, University of Connecticut, USA. Advisor: Prof. Yuan Hong.
  • 2021.09–2025.07, B.Eng. in Computer Science and Technology, South China University of Technology (SCUT), China.

💻 Internships

  • 2025.03–2025.06, Algorithm Intern, Baidu International Technology (Shenzhen) Co., Ltd., Shenzhen, China. Worked on pre-training data pipelines, distributed text processing, and automatic data labeling.
  • 2024.06–2025.01, Research Intern, Institute of Automation, Chinese Academy of Sciences, China. Worked on cross-modal deep learning, knowledge distillation, and multimodal fusion.