Denoising Multiple Outcomes to Improve Causal Policy Learning and Optimization

Leveraging dimensionality reduction via reduced-rank regression to denoise multiple noisy outcomes can improve the estimation and optimization of causal policies.