We develop computational cognitive models for AI that can explain, learn, and collaborate. Our research bridges cognitive science, linguistics, and robotics to create trustworthy, human-like intelligent systems.
The cognitive architecture for intelligent agents with language, embodiment, and agency
OntoAgent provides a comprehensive framework for building agents capable of interpreting multiple input modalities, reasoning about goals and plans, and acting in real or simulated environments—all while maintaining transparency and trustworthiness.
Deep semantic analysis producing ontologically-grounded meaning representations.
Natural language generation from semantic representations to fluent text.
Agents that know what they know and can explain their own reasoning.
Episodic memory, knowledge bases, and goal-driven agenda management.
Human-AI Robotic Team Member Operating with Natural Intelligence and Communication
Our flagship cognitive-robotic architecture integrates OntoAgent with general-purpose robot control systems. HARMONIC enables robots to reason about joint goals, explain their actions, and recognize when to defer to human teammates.
Meaningful, intentional communication grounded in semantic understanding.
Transparent causal reasoning that humans can inspect and verify.
Coordinated action with mutual trust and shared goal understanding.
Integration of high-level reasoning with low-level robot control.