Research map / Reinforcement learning / Policy gradient and actor-critic
Confounding-robust policy evaluation
Estimate policy value under confounding or distribution shift via counterfactual and density-ratio methods.
47 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Policy gradient and actor-critic, keyword cluster “ope · estimators · importance”
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.
11 papers in 2023–24 and 14 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 3.2% to 3.0%.
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 · Data-Efficient Policy Evaluation Through Behavior Policy Search (ICML 2017 · 13 citations)
- 2018 · Balanced Policy Evaluation and Learning (NeurIPS 2018 · 69 citations)
- 2019 · DualDICE: Behavior-Agnostic Estimation of Discounted Stationary Distribution Corrections (NeurIPS 2019 · 111 citations)
- 2020 · Efficient Policy Learning from Surrogate-Loss Classification Reductions (ICML 2020 · 7 citations)
- 2021 · Action Candidate Based Clipped Double Q-learning for Discrete and Continuous Action Tasks (AAAI 2021 · 14 citations)
- 2022 · Markovian Interference in Experiments (NeurIPS 2022 · 18 citations)
- 2023 · Integrating Reward Maximization and Population Estimation: Sequential Decision-Making for Internal Revenue Service Audit Selection (AAAI 2023 · 5 citations)
- 2024 · Policy Learning for Balancing Short-Term and Long-Term Rewards (ICML 2024 · 0 citations)
- 2025 · Distributionally Robust Policy Evaluation and Learning for Continuous Treatment with Observational Data (AAAI 2025 · 0 citations)
- 2026 · DIMM: Decoupled Multi-hierarchy Kalman Filter via Reinforcement Learning (AAAI 2026 · 3 citations)
Earliest papers here
- Data-Efficient Policy Evaluation Through Behavior Policy Search (ICML 2017 · 13 citations)
- Importance Sampling with Unequal Support (AAAI 2017 · 8 citations)
- Balanced Policy Evaluation and Learning (NeurIPS 2018 · 69 citations)
What has been tried
38 papers here report 287 measured changes (a component added, removed or swapped, compared with a baseline): helped 65%, no clear effect 9%, hurt 21%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
What reviewers will expect
Datasets and baselines named in the 9 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- on-policy Monte Carlo · 2 of 9 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 25 papers; from OpenAlex and the papers' affiliations).
- Carnegie Mellon University · 3 papers
- London School of Economics · 3 papers
- University of Chicago · 2 papers
- Stanford University · 2 papers
- Beijing Technology and Business University · 2 papers
- Yunnan University · 2 papers
- University of Pennsylvania · 2 papers
Use it
Download BibTeX of 38 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
- Importance-weighted OPE estimators (57 papers)
- Confidence intervals and OPE benchmarks (28 papers)
Elsewhere in the topic
- Model-based RL with guarantees (245 papers)
- Regularized scalable deep value learning (216 papers)
- Temporal-difference and discounting theory (143 papers)
- Tree search and policy optimization (109 papers)
- Safety layers and shielding (85 papers)
- Variance-reduced policy gradients (79 papers)
- Actor-critic theory (76 papers)
- Off-policy actor-critic with expressive policies (71 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).