 Double machine learning

Double machine learning uses orthogonal estimating equations and cross-fitting so that sufficiently accurate machine-learning nuisance estimates yield asymptotically normal target estimates.