Outcome-Aware Spectral Feature Learning for Instrumental Variable Regression
This work addresses causal inference challenges for researchers in statistics and machine learning, but it is incremental as it builds on existing spectral feature methods.
The paper tackles the problem of causal effect estimation with hidden confounders in nonparametric instrumental variable regression by introducing Augmented Spectral Feature Learning, which makes feature learning outcome-aware to address failures when the true causal function is poorly represented by dominant spectral features, and validates the approach on benchmarks.
We address the problem of causal effect estimation in the presence of hidden confounders using nonparametric instrumental variable (IV) regression. An established approach is to use estimators based on learned spectral features, that is, features spanning the top singular subspaces of the operator linking treatments to instruments. While powerful, such features are agnostic to the outcome variable. Consequently, the method can fail when the true causal function is poorly represented by these dominant singular functions. To mitigate, we introduce Augmented Spectral Feature Learning, a framework that makes the feature learning process outcome-aware. Our method learns features by minimizing a novel contrastive loss derived from an augmented operator that incorporates information from the outcome. By learning these task-specific features, our approach remains effective even under spectral misalignment. We provide a theoretical analysis of this framework and validate our approach on challenging benchmarks.