Research map / Probabilistic and scientific modeling / Treatment effects
Distributional treatment effects
Estimates distributional and kernel treatment effects with doubly robust tests.
51 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Treatment effects, keyword cluster “treatment effects · heterogeneous treatment · average treatment”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
16 papers in 2023–24 and 26 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 12.2% to 15.9%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · Learning spatiotemporal piecewise-geodesic trajectories from longitudinal manifold-valued data (NeurIPS 2017 · 12 citations)
- 2018 · Orthogonal Machine Learning: Power and Limitations (ICML 2018 · 8 citations)
- 2019 · Orthogonal Random Forest for Causal Inference (ICML 2019 · 31 citations)
- 2020 · A New Framework for Online Testing of Heterogeneous Treatment Effect (AAAI 2020 · 5 citations)
- 2021 · A/B Testing for Recommender Systems in a Two-sided Marketplace (NeurIPS 2021 · 9 citations)
- 2023 · Counterfactuals for the Future (AAAI 2023 · 6 citations)
- 2024 · Probabilities of Causation with Nonbinary Treatment and Effect (AAAI 2024 · 6 citations)
- 2025 · On Efficient Estimation of Distributional Treatment Effects under Covariate-Adaptive Randomization (ICML 2025 · 2 citations)
- 2026 · Design-Based Anytime-Valid Inference for Randomized Experiments with Delayed Outcomes and Staggered Entry (ICML 2026 · 0 citations)
Earliest papers here
- Learning spatiotemporal piecewise-geodesic trajectories from longitudinal manifold-valued data (NeurIPS 2017 · 12 citations)
- Orthogonal Machine Learning: Power and Limitations (ICML 2018 · 8 citations)
- Orthogonal Random Forest for Causal Inference (ICML 2019 · 31 citations)
What has been tried
34 papers here report 294 measured changes (a component added, removed or swapped, compared with a baseline): helped 69%, no clear effect 12%, hurt 14%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Replace the RBF kernel with the Sine kernel | IHDP, ACIC, Boston · PEHE | Hurt | ABC3: Active Bayesian Causal Inference with Cohn Criteria in Randomized Experiments (AAAI 2025 · 1 citation) | |
| Aggregate permuted WTAB p-values with Cauchy combination rather than using WTAB without permutation aggregation | Real-world-derived synthetic data, LightGBM · power | 0.738 -> 0.758 | Helped | Strategic A/B testing via Maximum Probability-driven Two-armed Bandit (ICML 2025 · 0 citations) |
| deliberately misspecify the δ(X) estimator | Synthetic one-sided noncompliance · consistency | No clear effect | Efficient Adaptive Experimentation with Noncompliance (NeurIPS 2025 · 0 citations) | |
| Set T1 = 0, removing the uniform-exploration stage | 101-unit network simulation · cumulative regret | No clear effect | Online Experimental Design With Estimation-Regret Trade-off Under Network Interference (NeurIPS 2025 · 0 citations) | |
| Remove adversarial balancing | TG, τ=6 (γ=0,2,4) · RMSE | −0.1%; +1.6%; +4.3% | Mixed | A Dual-module Framework for Counterfactual Estimation over Time (ICML 2024) |
| Add the selection indicator generator Gs | Synthetic, S = 1 samples · √PEHE | 1.453±0.649 -> 0.984±0.074 | Helped | A Generative Approach for Treatment Effect Estimation under Collider Bias: From an Out-of-Distribution Perspective (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 20 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- IHDP · 2 of 20 papers
Baselines
- DIM · 2 of 20 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 42 papers; from OpenAlex and the papers' affiliations).
- Tsinghua University · 4 papers
- Amazon · 3 papers
- CyberAgent · 3 papers
- The University of Tokyo · 2 papers
- University of California, Berkeley · 2 papers
- Google · 2 papers
- Massachusetts Institute of Technology · 2 papers
- Carnegie Mellon University · 2 papers
Use it
Download BibTeX of 33 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Meta-learners and partial identification (78 papers)
- Benchmarking treatment effect models (77 papers)
Elsewhere in the topic
- Proximal and deep causal effect estimation (105 papers)
- Latent-variable causal inference (91 papers)
- Instrumental variable estimation (19 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).