Faculty Profile: Prof. Jiawei Han — UIUC Data Mining Group
Faculty Profile: Prof. Jiawei Han — UIUC Data Mining Group
Position: Michael Aiken Chair Professor (Full) Institution: UIUC, Siebel School of Computing and Data Science Lab: Data Mining Research Group — dm1.cs.uiuc.edu Report date: 2026-06-12
Research Focus
Data mining, text mining, knowledge graphs, knowledge-enhanced LLMs, RL for search and reasoning agents, weakly supervised learning.
Academic Profile
- ACM Fellow, IEEE Fellow
- KDD Innovation Award, IEEE Computer Society Technical Achievement Award
- Citation count: 200,000+ (Google Scholar)
- 30+ years at UIUC; graduated 60+ PhD students
- Senior author on recent LLM+RL work (Search-R1, s3)
Key Publications (2025)
| Paper | Venue | Relevance |
|---|---|---|
| Search-R1: Training LLMs to reason and leverage search with RL | COLM’25 | RL training for search agent — directly relevant to post-training |
| s3: Train a search agent via RL with minimal data | EMNLP’25 | Low-data RL for agents |
| Reasoning-Enhanced Healthcare with KG Community Retrieval | ICLR’25 | KG-augmented reasoning |
| TELEClass: Taxonomy+LLM hierarchical classification | WWW’25 | LLM for text classification |
| PlugMem: plug-and-play memory for LLM agents (w/ ChengXiang Zhai) | 2025 | LLM agent memory |
Alumni Placement
One of the largest PhD labs at UIUC; well-established alumni network spanning:
- Google (Research), Microsoft Research, Amazon, Meta, Baidu
- Multiple faculty positions at top-50 CS departments
- Placement record is among the strongest at UIUC overall
Fit with Weijia Zhang
| Dimension | Assessment |
|---|---|
| RL for search / reasoning agents | ✅ Search-R1, s3 directly relevant |
| Knowledge graphs + LLM | ✅ Strong, KG-augmented reasoning |
| LLM memory for agents | ✅ PlugMem |
| Data pipeline / text mining | ✅ Long track record |
| General NLP agents | ⚠️ Adjacent (text mining focus, not conversation/GUI) |
| SFT / post-training pipeline | ⚠️ Indirect involvement |
| GUI / VLM / multimodal agents | ❌ Not his area |
Verdict
Strong placement record and meaningful recent work in RL for agents (Search-R1) and KG-augmented reasoning. Core identity is still data mining / text mining / knowledge graphs rather than pure agentic AI. Good PhD option if Weijia is interested in the knowledge + search + agent intersection. Lower priority than Ji or Hakkani-Tür for Weijia’s current focus on GUI/VLM/agentic systems.