Build truthful, safe, and secure AI systems with theoretical guarantees.

Conformal Abstention Conformal Prediction Uncertainty Quantification Reinforcement Learning Large Language Models Agentic AI Physical AI

Align AI agents to be truthful, safe, and secure.

Agentic AI AI Alignment Reinforcement Learning

Live with safe robots.

Vision-Language-Action Models Sim-to-Real Adaptation Reinforcement Learning Conformal Abstention Conformal Prediction Uncertainty Quantification

Make Agentic AI better by Red Teaming.

Agentic AI Reinforcement Learning