End-to-End Balancing for Causal Continuous Treatment-Effect Estimation
This work addresses the challenge of stable weight estimation in causal inference for continuous treatments, offering an incremental improvement over existing entropy balancing methods.
The paper tackles the problem of observational causal inference with continuous treatments by introducing an end-to-end optimization algorithm for entropy balancing, which learns weights to maximize causal inference accuracy. The result is a more accurate estimation of causal effects compared to baseline entropy balancing, as demonstrated on synthetic and real-world data.
We study the problem of observational causal inference with continuous treatments in the framework of inverse propensity-score weighting. To obtain stable weights, we design a new algorithm based on entropy balancing that learns weights to directly maximize causal inference accuracy using end-to-end optimization. In the process of optimization, these weights are automatically tuned to the specific dataset and causal inference algorithm being used. We provide a theoretical analysis demonstrating consistency of our approach. Using synthetic and real-world data, we show that our algorithm estimates causal effect more accurately than baseline entropy balancing.