Research map / Reinforcement learning / MDPs and planning
POMDP policy optimization
Solve POMDPs with Monte Carlo, transformers, and policy-based methods.
68 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in MDPs and planning, keyword cluster “markov decision · processes · observable”
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 18 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 7.9% to 10.3%.
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
- 2016 · RAO*: An Algorithm for Chance-Constrained POMDP's (AAAI 2016 · 69 citations)
- 2017 · Accelerated Vector Pruning for Optimal POMDP Solvers (AAAI 2017 · 41 citations)
- 2018 · rho-POMDPs have Lipschitz-Continuous epsilon-Optimal Value Functions (NeurIPS 2018 · 7 citations)
- 2019 · Personalized Robot Tutoring Using the Assistive Tutor POMDP (AT-POMDP) (AAAI 2019 · 50 citations)
- 2020 · Off-Policy Evaluation in Partially Observable Environments (AAAI 2020 · 30 citations)
- 2021 · Reinforcement Learning with State Observation Costs in Action-Contingent Noiselessly Observable Markov Decision Processes (NeurIPS 2021 · 7 citations)
- 2022 · Blockwise Sequential Model Learning for Partially Observable Reinforcement Learning (AAAI 2022 · 4 citations)
- 2023 · Safe Policy Improvement for POMDPs via Finite-State Controllers (AAAI 2023 · 10 citations)
- 2024 · Mitigating Partial Observability in Sequential Decision Processes via the Lambda Discrepancy (NeurIPS 2024 · 2 citations)
- 2025 · Learning Logic Specifications for Policy Guidance in POMDPs: an Inductive Logic Programming Approach (AAAI 2025 · 5 citations)
- 2026 · The Value Function Semi-Algebraic Set in Partially Observable Markov Decision Processes (ICML 2026 · 0 citations)
Earliest papers here
- RAO*: An Algorithm for Chance-Constrained POMDP's (AAAI 2016 · 69 citations)
- Learning for Decentralized Control of Multiagent Systems in Large, Partially-Observable Stochastic Environments (AAAI 2016 · 29 citations)
- A Symbolic SAT-Based Algorithm for Almost-Sure Reachability with Small Strategies in POMDPs (AAAI 2016 · 29 citations)
What has been tried
48 papers here report 319 measured changes (a component added, removed or swapped, compared with a baseline): helped 61%, no clear effect 15%, hurt 18%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use widely separated λ values, with values close to 0 or 1 | large partially observable environments · performance | Helped | Mitigating Partial Observability in Sequential Decision Processes via the Lambda Discrepancy (NeurIPS 2024 · 2 citations) | |
| Use lazy rather than basic policy sampling | SensorPlacement-12, policy-tree depths M=1 to M=5 · CPU time for 50 CE-iterations | 2.65 -> 1.09; 12.61 -> 1.58; 137.56 -> 2.2; 900.97 -> 3.07; 3928.45 -> 5.13 | Helped | A Surprisingly Simple Continuous-Action POMDP Solver: Lazy Cross-Entropy Search Over Policy Trees (AAAI 2024 · 1 citation) |
| Full observability instead of partial observability | PyBullet fully observable tasks · normalized score | No clear effect | Rethinking Transformers in Solving POMDPs (ICML 2024 · 1 citation) | |
| Replace Twisted Imagination with Nested Latent Imagination | SafetyPointGoal2 · Reward | 13.59 -> 10.79 | Hurt | PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning (ICML 2025 · 0 citations) |
| increase the number of Feynman–Kac and belief-filter particles | POMDP benchmarks · learning performance | Helped | Sequential Monte Carlo for Policy Optimization in Continuous POMDPs (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 14 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- POMDP Coder · 2 of 14 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 34 papers; from OpenAlex and the papers' affiliations).
- Radboud University · 3 papers
- George Washington University · 3 papers
- Massachusetts Institute of Technology · 3 papers
- University of Illinois, Urbana Champaign · 3 papers
- Australian National University · 2 papers
- Johns Hopkins University · 2 papers
- Tsinghua University · 2 papers
- University of Texas, Austin · 2 papers
- Google · 2 papers
- Amazon · 2 papers
Use it
Download BibTeX of 40 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
- Average-reward and regularized MDPs (135 papers)
- POMDPs and confounded RL (99 papers)
- Robust MDP policy gradient (87 papers)
Elsewhere in the topic
- Function approximation in RL theory (93 papers)
- Offline RL sample complexity (86 papers)
- Policy optimization theory (61 papers)
- Rich-observation exploration (53 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).