Research map / Reinforcement learning / Imitation learning
Theory of inverse RL
Give theoretical guarantees and exploration strategies for offline and active inverse reinforcement learning.
44 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Imitation learning, keyword cluster “inverse reinforcement · irl · reward function”
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.
18 papers in 2023–24 and 8 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 15.1% to 6.9%.
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 · Cooperative Inverse Reinforcement Learning (NeurIPS 2016 · 387 citations)
- 2017 · Compatible Reward Inverse Reinforcement Learning (NeurIPS 2017 · 17 citations)
- 2019 · Meta-Inverse Reinforcement Learning with Probabilistic Context Variables (NeurIPS 2019 · 34 citations)
- 2020 · Maximum Likelihood Constraint Inference for Inverse Reinforcement Learning (ICLR 2020 · 22 citations)
- 2021 · Inverse Constrained Reinforcement Learning (ICML 2021 · 6 citations)
- 2022 · Active Exploration for Inverse Reinforcement Learning (NeurIPS 2022 · 6 citations)
- 2023 · Misspecification in Inverse Reinforcement Learning (AAAI 2023 · 5 citations)
- 2024 · Meta-Inverse Reinforcement Learning for Mean Field Games via Probabilistic Context Variables (AAAI 2024 · 5 citations)
- 2025 · Partial Identifiability in Inverse Reinforcement Learning for Agents with Non-Exponential Discounting (AAAI 2025 · 1 citation)
- 2026 · Trust Region Inverse Reinforcement Learning: Explicit Dual Ascent using Local Policy Updates (ICML 2026 · 0 citations)
Earliest papers here
- Cooperative Inverse Reinforcement Learning (NeurIPS 2016 · 387 citations)
- Inverse Reinforcement Learning through Policy Gradient Minimization (AAAI 2016 · 28 citations)
- Compatible Reward Inverse Reinforcement Learning (NeurIPS 2017 · 17 citations)
What has been tried
37 papers here report 297 measured changes (a component added, removed or swapped, compared with a baseline): helped 60%, no clear effect 8%, hurt 26%, 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 |
|---|---|---|---|---|
| add meta-regularization | HalfCheetah · cumulative reward | -386.78 -> -214.63 | Helped | In-Trajectory Inverse Reinforcement Learning: Learn Incrementally Before an Ongoing Trajectory Terminates (NeurIPS 2024 · 3 citations) |
| use TD3 instead of TD7 as policy optimizer | DMControl suite · return | No clear effect | Non-Adversarial Inverse Reinforcement Learning via Successor Feature Matching (ICLR 2025 · 0 citations) | |
| Increase confidence from 30% to 90% | Blocked Walker · constraint violation rate | Helped | Confidence Aware Inverse Constrained Reinforcement Learning (ICML 2024 · 0 citations) | |
| replace discounted visitation distribution with empirical distribution | Ant(LS) · AUC | 0.94±0.01 -> 0.98±0.01 | Helped | Learning Constraints from Offline Demonstrations via Superior Distribution Correction Estimation (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 7 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- IQ-Learn · 2 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 26 papers; from OpenAlex and the papers' affiliations).
- The Chinese University of Hong Kong · 8 papers
- Politecnico di Milano · 4 papers
- University of Waterloo · 4 papers
- Huawei · 4 papers
- University of Auckland · 2 papers
- University of Oxford · 2 papers
- Shanghai Jiao Tong University · 2 papers
- Mohamed bin Zayed University of Artificial Intelligence · 2 papers
Use it
Download BibTeX of 37 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
- Reward learning from demonstrations (70 papers)
Elsewhere in the topic
- Offline imitation from imperfect data (108 papers)
- Imitation combined with RL (108 papers)
- Behavior cloning theory and models (68 papers)
- Adversarial imitation learning (33 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).