Flow Models for Arbitrary Conditional Likelihoods

by   Yang Li, et al.

Understanding the dependencies among features of a dataset is at the core of most unsupervised learning tasks. However, a majority of generative modeling approaches are focused solely on the joint distribution p(x) and utilize models where it is intractable to obtain the conditional distribution of some arbitrary subset of features x_u given the rest of the observed covariates x_o: p(x_u | x_o). Traditional conditional approaches provide a model for a fixed set of covariates conditioned on another fixed set of observed covariates. Instead, in this work we develop a model that is capable of yielding all conditional distributions p(x_u | x_o) (for arbitrary x_u) via tractable conditional likelihoods. We propose a novel extension of (change of variables based) flow generative models, arbitrary conditioning flow models (AC-Flow), that can be conditioned on arbitrary subsets of observed covariates, which was previously infeasible. We apply AC-Flow to the imputation of features, and also develop a unified platform for both multiple and single imputation by introducing an auxiliary objective that provides a principled single "best guess" for flow models. Extensive empirical evaluations show that our models achieve state-of-the-art performance in both single and multiple imputation across image inpainting and feature imputation in synthetic and real-world datasets. Code is available at https://github.com/lupalab/ACFlow.


page 5

page 6

page 7

page 10

page 11

page 12


Arbitrary Conditional Distributions with Energy

Modeling distributions of covariates, or density estimation, is a core c...

PriSTI: A Conditional Diffusion Framework for Spatiotemporal Imputation

Spatiotemporal data mining plays an important role in air quality monito...

Universal Conditional Machine

We propose a single neural probabilistic model based on variational auto...

Missing Features Reconstruction Using a Wasserstein Generative Adversarial Imputation Network

Missing data is one of the most common preprocessing problems. In this p...

Missingness Augmentation: A General Approach for Improving Generative Imputation Models

Despite tremendous progress in missing data imputation task, designing n...

Training and Inference on Any-Order Autoregressive Models the Right Way

Conditional inference on arbitrary subsets of variables is a core proble...

Correcting Conditional Mean Imputation for Censored Covariates and Improving Usability

Analysts are often confronted with censoring, wherein some variables are...

Please sign up or login with your details

Forgot password? Click here to reset