I study how representation choices constrain inference, prediction, and decision-making under regime change in complex systems.
I focus on the modeling components themselves, not just their assembly. Different building blocks induce different geometries, expressive limits, and failure modes.
Most modeling approaches start by choosing a model and then trying to make it fit the data. My work focuses on an earlier and often overlooked question:
How do we describe a system in the first place?
Different ways of describing a system highlight different aspects of how it behaves. Some descriptions work well in one situation but quietly break when conditions change. I study how these choices affect what models can reliably infer or predict, and how to recognize when assumptions no longer hold.
Many systems operate across multiple regimes. Understanding where assumptions hold, and where they break, is essential for robust modeling.
Andrew Garcia
Andrew Garcia—Submitted (double-blind review)
Andrew Garcia & Marco Vega—BCRP Working Paper (forthcoming)