Advancing Multi-Agent Decision Intelligence
MDRL is a research lab focused on the science and engineering of intelligent systems that reason, plan, and act across complex, multi-step decision problems.
Research Areas
Our work spans foundational theory and applied systems across the full spectrum of multi-decision AI.
AI Safety
Developing principled methods to ensure AI systems behave reliably, transparently, and in alignment with human values under all conditions.
Reinforcement Learning
Developing algorithms that enable agents to learn optimal policies through interaction with complex, high-dimensional environments.
Multi-Agent Systems
Studying how multiple AI agents coordinate, compete, and cooperate to solve large-scale sequential decision problems.
Inverse Reinforcement Learning
Inferring the underlying goals and reward structures of agents from observed behavior to enable better alignment and imitation.