Arbitrary Conditional Distributions with Energy
This work addresses a more general and useful problem of arbitrary conditional density estimation for practitioners needing inference based on prior knowledge, offering a simpler and higher-performing approach than prior methods.
This paper tackles the problem of arbitrary conditional density estimation, which involves modeling any possible conditional distribution over a set of covariates. The proposed method, Arbitrary Conditioning with Energy (ACE), achieves state-of-the-art results for arbitrary conditional likelihood estimation and data imputation on standard benchmarks.
Modeling distributions of covariates, or density estimation, is a core challenge in unsupervised learning. However, the majority of work only considers the joint distribution, which has limited utility in practical situations. A more general and useful problem is arbitrary conditional density estimation, which aims to model any possible conditional distribution over a set of covariates, reflecting the more realistic setting of inference based on prior knowledge. We propose a novel method, Arbitrary Conditioning with Energy (ACE), that can simultaneously estimate the distribution $p(\mathbf{x}_u \mid \mathbf{x}_o)$ for all possible subsets of unobserved features $\mathbf{x}_u$ and observed features $\mathbf{x}_o$. ACE is designed to avoid unnecessary bias and complexity -- we specify densities with a highly expressive energy function and reduce the problem to only learning one-dimensional conditionals (from which more complex distributions can be recovered during inference). This results in an approach that is both simpler and higher-performing than prior methods. We show that ACE achieves state-of-the-art for arbitrary conditional likelihood estimation and data imputation on standard benchmarks.