Trustworthy Data Science and AI Lab

Rigorous Science for
Trustworthy AI

We study the privacy, robustness, and reliability of AI, spanning large language models, autonomous agents, federated learning, and modern data systems. Our work aims to combine rigorous foundations with practical algorithms and deployable systems.

Meet the Team → Publications →
LLM Security Differential Privacy Federated Learning Trustworthy Agents Secure Vector Database Data Markets
LLM Security Federated Learning Data Infra Agent AI RAG / Security Diff. Privacy
Lab News

Latest Updates

View all news →

Research Focus

What We Work On

We tackle trustworthy AI at three levels — foundational theory, algorithm design, and real-world system deployment.

Privacy Tech Foundations
Building the theoretical and algorithmic foundations of privacy-enhancing technologies — differential privacy, MPC, TEE, federated learning, and machine unlearning — from rigorous guarantees to deployable systems.
AI Security & Safety
Studying how LLMs and autonomous agents can be attacked and defended — covering jailbreak, prompt injection, backdoor attacks, and adversarial risks across the full AI stack.
Trustworthy Data Systems
Designing data pipelines and AI systems that are auditable and reliable — through secure vector database, data markets, multimodal safety evaluation, and trustworthy AI for DB.
Latest Work

Recent Publications

Selected recent publications — full list on Publications or Google Scholar.

Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space Defects
Hanxi Li, Jianan Zhou, Jiale Lao, Yibo Wang, Zhengmao Ye, Yang Cao, Junfen Wang, Mingjie Tang
VLDB 2027
arXiv
Joint Certification for Attributed Graphs: Beyond Topology-Only Robustness
Blaise Delattre, Hengyu Wu, Wei Yang Bryan Lim, Yang Cao
NeurIPS 2026
Rank-Aware Differentially Private Release of Listwise Preferences for LLM Alignment
Junwei Chen, Manjiang Yu, Pengpeng Qiao, Yang Cao
NeurIPS 2026
Align Before Aggregation: Basis-Consistent Federated LoRA under Heterogeneous Ranks
Pengpeng Qiao, Yang Cao, Lingling Zhang, Guo Cheng, Junwei Chen, Manjiang Yu, Wei Yang Bryan Lim, Masatoshi Yoshikawa
NeurIPS 2026

View all publications →

Lab at a Glance

Research, Collaboration, and Impact

An international research team advancing trustworthy AI through rigorous research, competitive funding, and collaborations across academia and industry.

100+
Peer-reviewed Publications
10+
Research Grants
20+
Researchers & Students

Academic and industry partnerships across Asia, Europe, and North America.

Advancing Trustworthy AI
Bridging rigorous research with real-world AI and data systems.
Publication Record
40+ papers in venues including CCS, VLDB, ICML, NeurIPS, and SIGMOD.
International Collaborations
Joint research and grants with overseas partners.
Meet the Team → Join Us →