AI Paper Atlas

Research map / Probabilistic and scientific modeling / Treatment effects

Efficient and Robust Semi-supervised Estimation of Average Treatment Effect with Partially Annotated Treatment and Response

Jue Hou, Rajarshi Mukherjee, Tianxi Cai

JMLR 2025 · Probabilistic and scientific modeling › Treatment effects · approach: Benchmarking treatment effect models

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
improve surrogate quality for both treatment and outcome to AUC 0.95 (ablation)SMMAL · Simulation, all settingsrelative efficiency vs supervised: relative efficiency about 1.32-1.64 across all settingsImproved
improve surrogate quality for both treatment and outcome to AUC 0.99 (ablation)SMMAL · Simulation, all settingsrelative efficiency vs supervised: relative efficiency about 2.23-2.89 across all settingsImproved
use the unlabeled data with surrogates (semi-supervised) instead of labeled data onlydoubly robust ATE estimator · Low-dimensional smooth simulation (AUC 0.80, 0.80)relative efficiency: 1.02No clear effect
use the unlabeled data with surrogates instead of labeled data only, with AUC 0.999 for A and 0.95 for YSMMAL · Low-dimensional smooth simulationrelative efficiency: SD 4.08 -> 2.69, RE 2.29Improved
use dichotomized surrogates as if they were true treatment and outcome (unsupervised)ATE estimation · Simulation, 95% CI coveragecoverageHurt
apply SMMAL instead of supervised DML on labeled datacolorectal cancer EHR ATE · Real-world EHR (N=4147)relative efficiency: 1.88 relative efficiency vs SL DMLImproved
apply SMMAL instead of supervised calibrated estimatorcolorectal cancer EHR ATE · Real-world EHR (N=4147)relative efficiency: 1.35 relative efficiency vs SL calImproved
adjust for confounders instead of crude analysis (ablation)SMMAL · Real-world EHR (N=4147)estimated treatment effect directionMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy