Adaptive Joint Distribution Learning
Damir Filipović, Michael Multerer, Paul Georg Schneider · SIAM Journal on Mathematics of Data Science · 2025
Abstract. We develop a new framework for estimating joint probability distributions using tensor product reproducing kernel Hilbert spaces (RKHS). Our framework accommodates a low-dimensional, normalized, and positive model of a Radon–Nikodym derivative, which we estimate from sample sizes of up to several millions, alleviating the inherent limitations of RKHS modeling. Well-defined normalized and positive conditional distributions are natural by-products to our approach. Our proposal is fast to compute and accommodates learning problems ranging from prediction to classification. Our theoretical findings are supplemented by favorable numerical results.