English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Estimating heterogeneous treatment effects from observational data is acrucial task across many fields, helping policy and decision-makers take betteractions. There has been recent progress on robust and efficient methods forestimating the conditional average treatment effect (CATE) function, but thesemethods often do not take into account the risk of hidden confounding, whichcould arbitrarily and unknowingly bias any causal estimate based onobservational data. We propose a meta-learner called the B-Learner, which canefficiently learn sharp bounds on the CATE function under limits on the levelof hidden confounding. We derive the B-Learner by adapting recent results forsharp and valid bounds of the average treatment effect (Dorn et al., 2021) intothe framework given by Kallus & Oprescu (2023) for robust and model-agnosticlearning of conditional distributional treatment effects. The B-Learner can useany function estimator such as random forests and deep neural networks, and weprove its estimates are valid, sharp, efficient, and have a quasi-oracleproperty with respect to the constituent estimators under more generalconditions than existing methods. Semi-synthetic experimental comparisonsvalidate the theoretical findings, and we use real-world data to demonstratehow the method might be used in practice.

B-Learner: Quasi-Oracle Bounds on Heterogeneous Causal Effects Under Hidden Confounding