Research map / Reinforcement learning / Visual and representation RL
Vision-based robot RL
Learn transferable visual control policies and world models for embodied agents.
44 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2026) · an approach in Visual and representation RL, keyword cluster “robot · manipulation · simulation”
State History Tried 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.
10 papers in 2023–24 and 14 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 1.7% to 2.0%.
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
- 2018 · Sim2Real Viewpoint Invariant Visual Servoing by Recurrent Control (CVPR 2018 · 110 citations)
- 2019 · A Model-Free Affective Reinforcement Learning Approach to Personalization of an Autonomous Social Robot Companion for Early Literacy Education (AAAI 2019 · 143 citations)
- 2020 · SAPIEN: A SimulAted Part-Based Interactive ENvironment (CVPR 2020 · 402 citations)
- 2021 · Habitat 2.0: Training Home Assistants to Rearrange their Habitat (NeurIPS 2021 · 24 citations)
- 2022 · Coarse-To-Fine Q-Attention: Efficient Learning for Visual Robotic Manipulation via Discretisation (CVPR 2022 · 74 citations)
- 2023 · Affordances From Human Videos as a Versatile Representation for Robotics (CVPR 2023 · 112 citations)
- 2024 · A Simulation Benchmark for Autonomous Racing with Large-Scale Human Data (NeurIPS 2024 · 12 citations)
- 2025 · Optimal Viewpoint Selection for Autonomous Photography Using Reinforcement Learning (AAAI 2025 · 0 citations)
- 2026 · Plan in Sandbox, Navigate in Open Worlds: Learning Physics-Grounded Abstracted Experience for Embodied Navigation (ICML 2026 · 0 citations)
Earliest papers here
- Sim2Real Viewpoint Invariant Visual Servoing by Recurrent Control (CVPR 2018 · 110 citations)
- A Model-Free Affective Reinforcement Learning Approach to Personalization of an Autonomous Social Robot Companion for Early Literacy Education (AAAI 2019 · 143 citations)
- Sim-Real Joint Reinforcement Transfer for 3D Indoor Navigation (CVPR 2019 · 29 citations)
What has been tried
33 papers here report 314 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 7%, hurt 22%, 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 |
|---|---|---|---|---|
| remove the reference path | F317 and GT3 · lap completion | 97.67 -> Fail (F317); 109.51 -> Fail (GT3) | Hurt | A Simulation Benchmark for Autonomous Racing with Large-Scale Human Data (NeurIPS 2024 · 12 citations) |
| Add height information to the 2D trajectory | Unseen tasks (overall) · success rate | 50% -> 67% | Helped | RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches (ICLR 2024 · 5 citations) |
| Use stage termination checks instead of a 25-step timeout | RS-Can · success rate | 69% -> 100% | Helped | Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Tasks (ICLR 2024 · 4 citations) |
| Add the rigid-body dynamics inductive bias to the predictor | Hopper-hop · prediction performance | Helped | Neural Motion Simulator Pushing the Limit of World Models in Reinforcement Learning (CVPR 2025 · 0 citations) | |
| use dense rewards instead of sparse rewards | CALVIN D · Avg. Length | 1.74 -> 1.90 | Helped | ReinboT: Amplifying Robot Visual-Language Manipulation with Reinforcement Learning (ICML 2025 · 0 citations) |
| add ELO ranking instead of greedy replacement of the top-rated goal | Drawer Open · final success rate | 0.20 ± 0.35 -> 0.97 ± 0.11 | Helped | GoalLadder: Incremental Goal Discovery with Vision-Language Models (NeurIPS 2025 · 0 citations) |
| Add a scaffolded exploration policy that collects data alongside the target policy | Blind Pick · task performance | Helped | Privileged Sensing Scaffolds Reinforcement Learning (ICLR 2024 · 0 citations) | |
| Use sparse reward instead of the shaped reward | AllegroHand · Reward | 38777 -> 0 | Hurt | Position: Automatic Environment Shaping is the Next Frontier in RL (ICML 2024 · 0 citations) |
| combine observation and action warnings with AND rather than OR | All environments · TWA | 0.51^{± 0.01} -> 0.65^{± 0.01} | Helped | Failure Prediction at Runtime for Generative Robot Policies (NeurIPS 2025) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 24 papers; from OpenAlex and the papers' affiliations).
- University of California, Berkeley · 3 papers
- University of California, San Diego · 3 papers
- Carnegie Mellon University · 2 papers
- Meta · 2 papers
- New York University · 2 papers
- University of Pennsylvania · 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
- Skill learning and diffusion policies (86 papers)
- Dexterous manipulation transfer (141 papers)
Elsewhere in the topic
- 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)
- Learned state and action representations (148 papers)
- RL with tree search for discovery (121 papers)
- Inferring decision-making from behavior (93 papers)
- Robust visual encoders for RL (86 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).