Inspired by Oklahoma's powerful skies, the tornado mark represents resilience and momentum. It reflects the OUNLP Lab's spirit: taming chaos into robust, reliable systems, staying focused through uncertainty, and advancing discovery.
Building trustworthy AI systems that support human learning, communication, and collaboration.
Multi-party and multi-modal dialogue systems and discourse-aware modeling.
Discourse-aware representation learning and structured prediction over language.
Large-language-model alignment and agentic systems that act and collaborate.
Grounded, multi-modal communication across speech, vision, and text.
Machine-learning models for scientific discovery across materials, chemistry, and the physical sciences.
Reliable, interpretable AI for education and healthcare contexts.
From multi-party dialogue and LLM agents to structured prediction and trustworthy deployment, our projects span the path from research models to real-world use in education and healthcare.
See our publications