Characterizing the Distinguishability of Product Distributions through Multicalibration

Learn about the distinguishability of product distributions through the lens of multicalibration in this 25-minute conference talk presented by Cassandra Marcussen at the Fields Institute. Explore joint research conducted with Aaron Putterman and Salil Vadhan that examines how multicalibration techniques can be used to characterize when product distributions can be effectively distinguished from one another. Discover the theoretical foundations and practical implications of this work in computational complexity and machine learning theory, gaining insights into the intersection of statistical learning and distribution testing.