Sheila A. McIlraith

Sheila A. McIlraith is a Professor at the University of Toronto, specializing in Artificial Intelligence, with a focus on Reinforcement Learning, Epistemic Planning, and Temporal Logic. Her work bridges theoretical AI with practical applications, including

• Multi-agent verification
• Reward machines for deep RL
• Non-Markovian fairness
• Constraint-based neural sequence generation

.

Her research emphasizes interpretable models, ethical AI integration, and robust planning. Recent publications highlight collaborations with students like Andrew C. Li and Toryn Q. Klassen. She has contributed to conferences such as AAAI, ICAPS, and NeurIPS, advancing methods in partially observable RL and LTL-based task specification.

She has supervised numerous students and co-authored studies on epistemic planning, reward shaping, and ethical curriculum design. Her work often employs

• SAT solvers
• Logical filtering
• Policy decomposition

to address complex AI challenges.