Research

The theory underneath the practice

Before I built AI systems for a living, I studied why intelligent actors make the choices they make. That question, in mathematics, history, and machine learning, still drives how I design systems today. Full texts are available below.

Master's Thesis201781 pages

Rational Minds of the Persian Gulf War (1990)

University of Illinois, Springfield. M.A. in History

Most histories of the Persian Gulf War explain what happened. This thesis tests whether the decisions were rational. Applying game theory, mathematical structures, and formal logic to the choices made by Saddam Hussein, the Bush administration, and coalition leadership, it builds an objective framework for testing rationality in wartime decision-making, one that strips away the subjectivity and hindsight bias that conventional military history struggles to escape. The work bridges behavioral economics and history, and it is the intellectual foundation for how I still evaluate high-stakes decisions: identify the actor's goals, model the payoffs, and judge the choice against the information available at the time.

Game TheoryBehavioral EconomicsDecision AnalysisMilitary History
Short Review Essay20202 pages

An Abstract View of Behavioral Predictions in Artificial Intelligence

Independent review

This review examines whether machine learning can predict human choices, and what that would imply for free will, through reinforcement learning and inverse reinforcement learning: RL infers choices from an actor's goals, IRL infers goals from an actor's choices. It interrogates the two-system model of impulsive and patient decision-making, challenges the assumption that impulsive choices are irrational, and connects the theory to applications from targeted treatment planning to understanding behavior in mental illness. Written at the intersection of the philosophy of mind and applied ML, it reflects a conviction that has shaped my production work since: a prediction system needs a model of the actor's goals, because the observed choices alone underdetermine them.

Reinforcement LearningInverse RLBehavioral PredictionPhilosophy of Mind

The applied side

The production systems this way of thinking produced are on the Impact page.

View the Impact