Dual Instrumental Variable Regression

We present a novel algorithm for instrumental variable (IV) regression, DualIV, which simplifies traditional two-stage methods via a dual formulation. Inspired by problems in stochastic programming, we show that the two-stage procedure for nonlinear IV regression can be reformulated as a convex-concave saddle-point problem. Our formulation circumvents the first-stage regression which is a potential bottleneck in real-world applications. Based on this new approach, we develop a simple kernel-based algorithm with a closed-form solution. Empirical results show that we are competitive to existing, more complicated algorithms for instrumental variable regression.