Through Wind and Rain. Building robust, interpretable AI systems that support human learning, communication, and collaboration.
School of Computer Science · University of Oklahoma · Norman, OK
Directed by Dr. Jie Cao, Assistant Professor
🌐 Website · 🔬 Research · 📚 Publications · 👥 Team · 📰 News · ✉️ Contact
We work toward reliable human–AI language systems — from multi-party dialogue and LLM agents to structured prediction and trustworthy deployment in education, healthcare, and science.
| Theme | |
|---|---|
| 🗣️ Dialogue & Discourse | Multi-party, multi-modal dialogue systems and discourse-aware modeling. |
| 🧩 Structured Prediction | Discourse-aware representation learning and structured prediction over language. |
| 🤖 LLM Alignment & Agents | LLM alignment and agentic systems that act and collaborate. |
| 🎛️ Multimodal Interaction | Grounded communication across speech, vision, and text. |
| 🔭 AI for Science | ML for discovery across materials, chemistry, and the physical sciences. |
| 🛡️ Trustworthy AI | Reliable, interpretable AI for education and healthcare. |
Browse publications by theme →
- A Multimodal Large Language Model for Materials Science — Nature Machine Intelligence (2026) · doi
- HydroAgent: Closing the Gap Between Frontier LLMs and Human Experts in Hydrologic Model Calibration via Simulator-Grounded RL — arXiv (2026) · arXiv:2605.17792
| Project | What it is |
|---|---|
| Multi-Round-Text-Simplifier | Iterative, multi-round text simplification |
| dialogue-act | Dialogue-act modeling |
| lab-website | Source of ounlp.org |
We welcome students and collaborators excited about NLP/ML for dialogue, agents, and trustworthy AI. See ounlp.org or email jie.cao@ou.edu.
🌪️ Inspired by Oklahoma's powerful skies — staying focused through uncertainty, and advancing discovery.