New method enhances treatment effect estimation using unlabeled data

Researchers have developed new methods for treatment effect estimation using semi-supervised learning and auxiliary unlabeled covariates. Their framework, termed prediction-powered causal inference, introduces efficiency bounds and estimators that achieve lower asymptotic variance by incorporating this additional data. The study considers both one-sample (censoring) and two-sample (case-control) settings, demonstrating improved estimation accuracy in both scenarios. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a new framework for causal inference that could improve the accuracy of AI models in domains requiring understanding of treatment effects.

RANK_REASON Academic paper on a novel statistical method for causal inference. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

• Masahiro Kato
• arXiv

• paper
• other