NESTER: An Adaptive Neurosymbolic Method for Treatment Effect Estimation

Treatment effect estimation from observational data is a central problem in causal inference. Methods based on potential outcomes framework solve this problem by exploiting inductive biases and heuristics from causal inference. Each existing technique addresses a specific aspect of treatment effect estimation, such as controlling propensity score, enforcing randomization, etc., by designing neural network architectures and regularizers. In this paper, we propose an adaptive method called Neurosymbolic Treatment Effect Estimator (NESTER), a generalized method for treatment effect estimation. NESTER brings together all the desiderata for treatment effect estimation into one framework. For this purpose, we design a Domain Specific Language (DSL) for the treatment effect estimation based on inductive biases used in literature. We also theoretically study NESTER's capability for the treatment effect estimation task. Our comprehensive empirical results show that NESTER performs better on benchmark datasets than state-of-the-art methods without compromising run time requirements.