End-to-End Balancing for Causal Continuous Treatment-Effect Estimation

Mohammad Taha Bahadori, Eric Tchetgen Tchetgen, David Heckerman

Keywords: End-to-end learning, Entropy Balancing, Continuous Treatments

Abstract:

We study the problem of observational causal inference with continuous treatment. We focus on the challenge of estimating the causal response curve for infrequently-observed treatment values.
We design a new algorithm based on the framework of entropy balancing which learns weights that directly maximize causal inference accuracy using end-to-end optimization. Our weights can be customized for different datasets and causal inference algorithms. We propose a new theory for consistency of entropy balancing for continuous treatments. Using synthetic and real-world data, we show that our proposed algorithm outperforms the entropy balancing in terms of causal inference accuracy.

One-sentence Summary: We argue that entropy balancing is suboptimal. We propose the end-to-end balancing framework that directly improves the accuracy of causal inference.