Conditional Independence Testing using Generative Adversarial Networks
This work addresses the challenge of detecting conditional dependence for researchers in statistics and machine learning, particularly in high-dimensional settings like genetics, though it appears incremental as it builds on existing GAN methods for hypothesis testing.
The paper tackles the problem of conditional independence testing in high-dimensional feature spaces by introducing a new test statistic based on generative adversarial networks, which approximates a conditional distribution to maximize power while controlling type I error without distributional assumptions, and demonstrates significant power gains in synthetic simulations and application to genetic data.
We consider the hypothesis testing problem of detecting conditional dependence, with a focus on high-dimensional feature spaces. Our contribution is a new test statistic based on samples from a generative adversarial network designed to approximate directly a conditional distribution that encodes the null hypothesis, in a manner that maximizes power (the rate of true negatives). We show that such an approach requires only that density approximation be viable in order to ensure that we control type I error (the rate of false positives); in particular, no assumptions need to be made on the form of the distributions or feature dependencies. Using synthetic simulations with high-dimensional data we demonstrate significant gains in power over competing methods. In addition, we illustrate the use of our test to discover causal markers of disease in genetic data.