Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a novel method for estimating density ratios by parametrizing them as a displacement field, where the log-ratio is derived from the Stein operator of the base distribution applied to this field. This unified approach simultaneously provides statistical and dynamical insights into distribution shifts through a single convex optimization problem. The proposed inference algorithms, push-forward and pull-back, enable efficient model correction and layered transformation fitting, respectively.
Unifying density ratio estimation and distribution transport, this work reveals how a single displacement field can simultaneously capture statistical shifts and dynamical transformations between probability distributions.
Density ratios quantify distribution shift from a probability-mass point of view, whereas displacement fields describe, from a dynamical point of view, how one distribution is transported onto another. Although both offer complementary insights, they are usually estimated separately, and converting one into the other requires post-processing. In this paper, we estimate the density ratio between a target and a base distribution by parametrizing it through a displacement field acting on the base: the log-ratio is modeled as minus the Stein operator of the base applied to the field, up to a normalizing constant. This gives both statistical and dynamical descriptions of the distribution shift through a single convex optimization problem. Iterating this estimate-and-move step gives two inference algorithms: push-forward moves the model and corrects a pretrained sampler without retraining it, whereas pull-back moves the data closer to the base and fits a transformation model one layer at a time. Applications to distribution shift in simulation-based inference and to nonlinear independent component analysis illustrate the benefits and limitations of the approach.