Estimating Causal Effects with the Neural Autoregressive Density Estimator

Estimation of causal effects is fundamental in situations were the underlying\nsystem will be subject to active interventions. Part of building a causal\ninference engine is defining how variables relate to each other, that is,\ndefining the functional relationship between variables given conditional\ndependencies. In this paper, we deviate from the common assumption of linear\nrelationships in causal models by making use of neural autoregressive density\nestimators and use them to estimate causal effects within the Pearl's\ndo-calculus framework. Using synthetic data, we show that the approach can\nretrieve causal effects from non-linear systems without explicitly modeling the\ninteractions between the variables.\n