Data-centric AI research group at the University of Science and Technology of China.
中国科学技术大学 · 认知智能全国重点实验室 · 数据驱动智能研究组
Official Homepage · PI Homepage · GitHub Organization
Star-Team is a research group led by Prof. Hao Wang at USTC. We study how high-quality data, behavior signals, knowledge, and model systems can work together to build more reliable, scalable, and useful intelligent systems.
Our work is rooted in data-centric AI and spans recommender systems, data mining, large language models, AI4Science, and trustworthy data systems. The group is based at the National Key Laboratory of Cognitive Intelligence and works in close academic collaboration with Prof. Enhong Chen's team, who is a recipient of the National Science Fund for Distinguished Young Scholars.
- Data-centric AI: dataset regeneration, dataset distillation, feature generation, data quality, and data-efficient learning.
- Recommender systems: sequential recommendation, CTR prediction, long-term user behavior modeling, and large recommendation models.
- Data mining and user intelligence: behavior modeling, interaction mining, graph learning, and decision-oriented intelligence.
- Large language models: retrieval-augmented generation, agentic recommendation, long-text generation, and knowledge reasoning.
- AI4Science and trustworthy AI: scientific evaluation, domain knowledge modeling, robust learning, and interpretable systems.
| Repository | Focus | Signal |
|---|---|---|
| DR4SR | Data-centric sequential recommendation | KDD 2024 Best Student Paper |
| Awesome-Large-Recommendation-Models | Curated resources for large recommendation models | Community resource |
| GE4Rec | Generative feature modeling for CTR prediction | ICML 2025 |
| FuXi-alpha | Feature-interaction enhanced recommendation models | WWW 2025 |
| MIRRN | Long-term user behavior modeling for CTR prediction | KDD 2025 |
| RaPID | Retrieval-augmented long-text generation | ACL 2025 |
- Visit the official homepage for the full research profile, publications, projects, awards, and recruitment information.
- Browse DR4SR, GE4Rec, and FuXi-alpha for representative data-centric recommendation work.
- Follow Awesome-Large-Recommendation-Models for recent advances in large recommendation models.
- Check RaPID and related repositories for LLM and retrieval-augmented generation research.
- PI: Prof. Hao Wang, School of Computer Science and Technology, USTC
- Homepage: ustc-starteam.github.io
- GitHub: USTC-StarTeam