Research map / Reinforcement learning / Visual and representation RL
RL with tree search for discovery
Search molecules and plans with RL and Monte Carlo tree search.
121 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.
41 papers in 2023–24 and 35 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 7.2% to 5.0%.
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
- 2016 · Hierarchical Decision Making In Electricity Grid Management (ICML 2016 · 26 citations)
- 2017 · Fake News Mitigation via Point Process Based Intervention (ICML 2017 · 114 citations)
- 2018 · A Deep Reinforced Model for Abstractive Summarization (ICLR 2018 · 1,255 citations)
- 2019 · Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules (ICML 2019 · 142 citations)
- 2020 · Adaptive Quantitative Trading: An Imitative Deep Reinforcement Learning Approach (AAAI 2020 · 137 citations)
- 2021 · Reinforced Imitative Graph Representation Learning for Mobile User Profiling: An Adversarial Training Perspective (AAAI 2021 · 24 citations)
- 2022 · Fully Adaptive Framework: Neural Computerized Adaptive Testing for Online Education (AAAI 2022 · 60 citations)
- 2023 · Predicate Invention for Bilevel Planning (AAAI 2023 · 16 citations)
- 2024 · EarnHFT: Efficient Hierarchical Reinforcement Learning for High Frequency Trading (AAAI 2024 · 21 citations)
- 2025 · Optimizing Gene-Based Testing for Antibiotic Resistance Prediction (AAAI 2025 · 1 citation)
- 2026 · Brownian Bridge Augmented Surrogate Simulation and Injection Planning for Geological CO2 Storage (AAAI 2026 · 1 citation)
Earliest papers here
- Hierarchical Decision Making In Electricity Grid Management (ICML 2016 · 26 citations)
- Implementing Troubleshooting with Batch Repair (AAAI 2016 · 7 citations)
- Energy- and Cost-Efficient Pumping Station Control (AAAI 2016 · 4 citations)
What has been tried
97 papers here report 900 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 10%, hurt 18%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add optimal value supervisor | GALA · Converged reward sum | -0.01 -> 4.43 | Helped | EarnHFT: Efficient Hierarchical Reinforcement Learning for High Frequency Trading (AAAI 2024 · 21 citations) |
| Replace a GRU-based DQN agent with a GRU-based PPO agent | 9 victim websites · TFA | 224 -> 197 | Helped | BertRLFuzzer: A BERT and Reinforcement Learning Based Fuzzer (Student Abstract) (AAAI 2024 · 7 citations) |
| add the leader's incentive | EDRP · ORR | 0.76±0.00 -> 0.79±0.02 | Helped | Transition-Informed Reinforcement Learning for Large-Scale Stackelberg Mean-Field Games (AAAI 2024 · 7 citations) |
| Add prefix-enhanced training with prefix-to-full samples | MuST-C En-De tst-COMMON · BLEU | >2 BLEU points around the 1000ms mark | Helped | Divergence-Guided Simultaneous Speech Translation (AAAI 2024 · 6 citations) |
| Use local and global sequence features in LGFEN rather than an MLP predictor | DNA, Top10 mean · mean binding activity | 0.981 ± 0.008 -> 0.989 ± 0.006 | Helped | Designing Biological Sequences without Prior Knowledge Using Evolutionary Reinforcement Learning (AAAI 2024 · 4 citations) |
| Use SMILES instead of SELFIES | PyTDC tasks · reward | Helped | Searching for High-Value Molecules Using Reinforcement Learning and Transformers (ICLR 2024 · 3 citations) | |
| add metadata to GenoAR | E. coli · accuracy | 89.57 -> 94.04 | Helped | Optimizing Gene-Based Testing for Antibiotic Resistance Prediction (AAAI 2025 · 1 citation) |
| add genetic search to the GFlowNet policy training | PMO 23 oracles · AUC Top-10 | 15.738 -> 16.213 | Helped | Genetic-guided GFlowNets for Sample Efficient Molecular Optimization (NeurIPS 2024 · 1 citation) |
| add visual features to the tabular-only model | Small-area search, 2.5%, 5%, and 10% positive rates · ANT | 1.008 -> 1.052; 1.972 -> 2.240; 4.216 -> 4.372 | Helped | Active Geospatial Search for Efficient Tenant Eviction Outreach (AAAI 2025 · 0 citations) |
| Use codebook size k=1 instead of k=4 | CraftWorld Standard test · Expansions | 103.23 -> 263.92 | Hurt | Subgoal-Guided Policy Heuristic Search with Learned Subgoals (ICML 2025 · 0 citations) |
| Apply contrastive training to question–KC and solution step–KC pairs | KC-cluster question retrieval · micro-F1 | 0.2305 -> 0.8865 | Helped | Personalized Exercise Recommendation with Semantically-Grounded Knowledge Tracing (NeurIPS 2025 · 0 citations) |
| Add abstraction learning to evolutionary program synthesis | arithmetic_base · ground-truth entities per episodic roll-out | 23.98 | Helped | Learning Interestingness in Automated Mathematical Theory Formation (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 23 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- Random · 2 of 23 papers
- random policy · 2 of 23 papers
- ReAct · 2 of 23 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 76 papers; from OpenAlex and the papers' affiliations).
- Nanyang Technological University · 6 papers
- Meta · 4 papers
- Huawei · 3 papers
- University of Cambridge · 3 papers
- Shanghai Jiao Tong University · 2 papers
- Tianjin University · 2 papers
- Zhejiang University · 2 papers
- Hong Kong University of Science and Technology · 2 papers
- Stanford University · 2 papers
- Google DeepMind · 2 papers
Use it
Download BibTeX of 39 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)
- Inferring decision-making from behavior (93 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).