AI RESEARCH

Riesz Regression As Direct Density Ratio Estimation

arXiv CS.LG

ArXi:2511.04568v2 Announce Type: replace-cross This study clarifies the relationship between Riesz regression [Chernozhuko, 2021] and density ratio estimation (DRE) in causal inference problems, such as average treatment effect estimation. We first show that the Riesz representer can be written as a signed density ratio and then nstrate that the Riesz regression objective coincides with the least-squares importance fitting criterion [Kanamori, 2009