On counterfactual inference with unobserved confounding
This work addresses causal inference challenges in observational studies with unobserved confounding, offering a method for consistent imputation of missing covariates, but it is incremental as it builds on existing exponential family and pooling techniques.
The paper tackles the problem of learning counterfactual distributions for heterogeneous units with unobserved confounding using only one sample per unit, by modeling outcomes as an exponential family and introducing a convex pooling objective, achieving an error bound of O(s log k / p) for sparse parameters.
Given an observational study with $n$ independent but heterogeneous units, our goal is to learn the counterfactual distribution for each unit using only one $p$-dimensional sample per unit containing covariates, interventions, and outcomes. Specifically, we allow for unobserved confounding that introduces statistical biases between interventions and outcomes as well as exacerbates the heterogeneity across units. Modeling the conditional distribution of the outcomes as an exponential family, we reduce learning the unit-level counterfactual distributions to learning $n$ exponential family distributions with heterogeneous parameters and only one sample per distribution. We introduce a convex objective that pools all $n$ samples to jointly learn all $n$ parameter vectors, and provide a unit-wise mean squared error bound that scales linearly with the metric entropy of the parameter space. For example, when the parameters are $s$-sparse linear combination of $k$ known vectors, the error is $O(s\log k/p)$. En route, we derive sufficient conditions for compactly supported distributions to satisfy the logarithmic Sobolev inequality. As an application of the framework, our results enable consistent imputation of sparsely missing covariates.