Research map / Reinforcement learning / Visual and representation RL
Inferring decision-making from behavior
Interpret human and animal decisions via inverse RL and policy recovery.
93 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Visual and representation RL, keyword cluster “search · decision · systems”
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.
25 papers in 2023–24 and 33 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 4.4% to 4.7%.
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 · Adaptive optimal training of animal behavior (NeurIPS 2016 · 25 citations)
- 2017 · Query-Efficient Imitation Learning for End-to-End Simulated Driving (AAAI 2017 · 106 citations)
- 2018 · Context-Aware Symptom Checking for Disease Diagnosis Using Hierarchical Reinforcement Learning (AAAI 2018 · 101 citations)
- 2019 · Correlation Priors for Reinforcement Learning (NeurIPS 2019 · 3 citations)
- 2020 · Learning Situational Driving (CVPR 2020 · 68 citations)
- 2021 · Dialog Policy Learning for Joint Clarification and Active Learning Queries (AAAI 2021 · 9 citations)
- 2022 · Action-modulated midbrain dopamine activity arises from distributed control policies (NeurIPS 2022 · 9 citations)
- 2023 · Understanding Plasticity in Neural Networks (ICML 2023 · 7 citations)
- 2024 · Learning to Make Adherence-aware Advice (ICLR 2024 · 2 citations)
- 2025 · Learning Personalized Decision Support Policies (AAAI 2025 · 4 citations)
- 2026 · Delphi: A Neuro-Symbolic Framework for Individualized, Safe and Interpretable Treatment Recommendation (AAAI 2026 · 2 citations)
Earliest papers here
- Adaptive optimal training of animal behavior (NeurIPS 2016 · 25 citations)
- Catching heuristics are optimal control policies (NeurIPS 2016)
- Query-Efficient Imitation Learning for End-to-End Simulated Driving (AAAI 2017 · 106 citations)
What has been tried
75 papers here report 661 measured changes (a component added, removed or swapped, compared with a baseline): helped 59%, no clear effect 13%, hurt 22%, mixed 7%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use full pooling instead of no pooling | Oralytics simulation environment · Mean OSCB | 69.375 (0.047) -> 69.724 (0.047) | Helped | A Deployed Online Reinforcement Learning Algorithm in an Oral Health Clinical Trial (AAAI 2025 · 3 citations) |
| Add Stage 2 policy distillation | ExTES · SR | 0.865 -> 0.903 | Helped | Simulation-Free Hierarchical Latent Policy Planning for Proactive Dialogues (AAAI 2025 · 3 citations) |
| Increase the policy-fix horizon from k=1 to k=6 (r=5) | Gardener 100×100, 2500 walls, 250 plants, 250 frogs, 1-Step · kills (plants / frogs) | 1.90 / 1.12 -> 1.02 / 0.66 | Helped | ASP-Driven Emergency Planning for Norm Violations in Reinforcement Learning (AAAI 2025 · 1 citation) |
| Remove eccentricity-dependent layers from the visual feature extractor | Three in-domain conditions · saccade size | Hurt | Gazing at Rewards: Eye Movements as a Lens into Human and AI Decision-Making in Hybrid Visual Foraging (CVPR 2025 · 1 citation) | |
| use n-step rather than 1-step Q-learning for the action-value function | Atari 2600 games · average reward | The paper reports a 30.2% improvement (from 1.39 to 1.81). | Helped | Learning Uncertainty-Aware Temporally-Extended Actions (AAAI 2024 · 1 citation) |
| Add the soft lower bound (β ≠ 0) | GranTurismo 7 · competitive policy diversity | Helped | Discovering Creative Behaviors through DUPLEX: Diverse Universal Features for Policy Exploration (NeurIPS 2024 · 1 citation) | |
| add the approximated state-entropy term, increasing β | Energy constraint · average standard deviation of activity jumps | Helped | Controlled maximal variability along with reliable performance in recurrent neural networks (NeurIPS 2024 · 1 citation) | |
| Remove distance weighting from the connection-cost regulariser | DO, 3D-DO, G2K · module isolation | Hurt | Inducing, Detecting and Characterising Neural Modules: A Pipeline for Functional Interpretability in Reinforcement Learning (ICML 2025 · 0 citations) | |
| Use a very small epsilon threshold | RoboMimic Square ph · score | No clear effect | Falcon: Fast Visuomotor Policies via Partial Denoising (ICML 2025 · 0 citations) | |
| Set interpolation weight β to 1.0 (pure look-ahead) | Synthetic experiment · total welfare | Helped | Policy Design for Two-sided Platforms with Participation Dynamics (ICML 2025 · 0 citations) | |
| Use joint rather than disjoint position-evidence grid coding | Accumulating tower task · learning efficiency | Helped | A Multi-Region Brain Model to Elucidate the Role of Hippocampus in Spatially Embedded Decision-Making (ICML 2025 · 0 citations) | |
| Add state-dependent hidden-mode transitions and history-dependent rewards and policies | 5×5 Gridworld · test log-likelihood | Helped | Inverse Reinforcement Learning with Switching Rewards and History Dependency for Characterizing Animal Behaviors (ICML 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 24 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- PPO · 4 of 24 papers
- DQN · 2 of 24 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 58 papers; from OpenAlex and the papers' affiliations).
- Imperial College London · 4 papers
- Carnegie Mellon University · 4 papers
- Tsinghua University · 3 papers
- Massachusetts Institute of Technology · 3 papers
- Google DeepMind · 2 papers
- University of California, Los Angeles · 2 papers
- Harvard University · 2 papers
- University of Cambridge · 2 papers
- Fudan University · 2 papers
- Harbin Institute of Technology · 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
- In-context and adaptive RL (202 papers)
- Scalable RL systems and environments (179 papers)
- Probabilistic and autonomous RL (177 papers)
- RL for combinatorial routing (158 papers)
- RL with tree search for discovery (121 papers)
- Curriculum and environment design (81 papers)
Elsewhere in the topic
- Learned state and action representations (148 papers)
- Dexterous manipulation transfer (141 papers)
- Skill learning and diffusion policies (86 papers)
- Robust visual encoders for RL (86 papers)
- Novelty and impact-driven exploration (85 papers)
- Hierarchical skills and priors (84 papers)
- Latent-space world models (83 papers)
- Goal-conditioned policy learning (66 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).