Ulric B. and Evelyn L. Bray Social Sciences Seminar

Abstract: This talk will discuss two related papers that use machine learning as a complement to theoretical modeling, rather than a substitute for it. The key concepts are completeness, which compares how well a model predicts to the "best possible" predictions, and restrictiveness, which measures a theory's ability to match arbitrary hypothetical data: A very unrestrictive theory will be complete on almost any data.