Research map / Probabilistic and scientific modeling / Treatment effects
Localized Debiased Machine Learning: Efficient Inference on Quantile Treatment Effects and Beyond
Nathan Kallus, Xiaojie Mao, Masatoshi Uehara
Read the paperCheck your idea against 137,217 papers
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use LDML instead of IPW, DML-D and DML-F | LDML estimator · Simulation: second tertile of Y(1) from incomplete data | mean-squared error | Improved |
| use LDML instead of IPW, DML-D and DML-F | LDML estimator · Simulation: second tertile of Y(1) | 95% CI coverage | Improved |
| adjust for covariates with LDML instead of raw quantile difference | LDML with LASSO · 401(k) eligibility QTE, 50% quantile | QTE (thousand dollars): 8.98 -> 4.68 | Mixed |
| adjust for covariates with LDML instead of raw difference | LDML with LASSO · 401(k) eligibility QTE, 75% quantile | QTE (thousand dollars): 29.67 -> 13.94 | Mixed |
| increase number of cross-fitting folds K from 5 to 25 | LDML with LASSO · 401(k) eligibility QTE, 75% quantile | QTE (thousand dollars): 14.00 -> 13.93 | No clear effect |
| increase K from 5 to 25 | LDML with Neural Net · 401(k) eligibility QTE, 50% quantile | QTE (thousand dollars): 5.56 -> 5.55 | No clear effect |
| use forest nuisance estimator instead of LASSO | LDML · 401(k) eligibility QTE, 50% quantile | standard error: 0.68 -> 1.85 | Hurt |
| use neural net instead of LASSO | LDML · 401(k) participation LQTE, 75% quantile | LQTE (thousand dollars): 21.78 -> 32.73 | Mixed |
| increase K from 5 to 15 | LDML with LASSO · 401(k) participation LQTE, 50% quantile | LQTE (thousand dollars): 8.64 -> 8.55 | No clear effect |
Closest papers by meaning
- It’s Hard to Be Normal: The Impact of Noise on Structure-agnostic Estimation (NeurIPS 2025 · similarity 0.88)
- Local Learning for Covariate Selection in Nonparametric Causal Effect Estimation with Latent Variables (NeurIPS 2025 · similarity 0.87)
- Doubly Robust Proximal Causal Learning for Continuous Treatments (ICLR 2024 · similarity 0.86)
- Learning Representations of Instruments for Partial Identification of Treatment Effects (ICML 2025 · similarity 0.86)
- Mitigating Spurious Correlations via Disagreement Probability (NeurIPS 2024 · similarity 0.86)
- Double Machine Learning for Causal Inference under Shared-State Interference (ICML 2025 · similarity 0.86)
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.
