{"url":"/task/imitation-learning","name":"Imitation Learning","slug":"imitation-learning","description_markdown":"**Imitation Learning** is a framework for learning a behavior policy from demonstrations. Usually, demonstrations are presented in the form of state-action trajectories, with each pair indicating the action to take at the state being visited. In order to learn the behavior policy, the demonstrated actions are usually utilized in two ways. The first, known as Behavior Cloning (BC), treats the action as the target label for each state, and then learns a generalized mapping from states to actions in a supervised manner. Another way, known as Inverse Reinforcement Learning (IRL), views the demonstrated actions as a sequence of decisions, and aims at finding a reward/cost function under which the demonstrated decisions are optimal. \r\n\r\nFinally, a newer methodology, Inverse Q-Learning aims at directly learning Q-functions from expert data, implicitly representing rewards, under which the optimal policy can be given as a Boltzmann distribution similar to soft Q-learning\r\n\r\n<span class=\"description-source\">Source: [Learning to Imitate ](https://ai.stanford.edu/blog/learning-to-imitate)</span>","categories":[{"name":"Methodology","url":"/area/methodology"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":2122,"papers_with_code":691,"benchmarks":0,"benchmark_tables_in_archive":0,"benchmark_tables_shown":0,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":21,"subtasks":1,"parent_tasks":1},"benchmarks":[],"datasets":[{"url":"/dataset/carla","name":"CARLA","full_name":"Car Learning to Act","num_papers_in_archive":1345},{"url":"/dataset/airsim","name":"AirSim","full_name":"","num_papers_in_archive":285},{"url":"/dataset/ai2-thor","name":"AI2-THOR","full_name":"AI2-THOR","num_papers_in_archive":243},{"url":"/dataset/rlbench","name":"RLBench","full_name":"","num_papers_in_archive":167},{"url":"/dataset/torcs","name":"TORCS","full_name":"The Open Racing Car Simulator","num_papers_in_archive":96},{"url":"/dataset/interaction-dataset","name":"INTERACTION Dataset","full_name":"","num_papers_in_archive":81},{"url":"/dataset/maniskill2","name":"ManiSkill2","full_name":"","num_papers_in_archive":40},{"url":"/dataset/codraw","name":"CoDraw","full_name":null,"num_papers_in_archive":14},{"url":"/dataset/chalet","name":"CHALET","full_name":"Cornell House Agent Learning Environment","num_papers_in_archive":10},{"url":"/dataset/cubicasa5k","name":"CubiCasa5K","full_name":null,"num_papers_in_archive":10},{"url":"/dataset/atari-head","name":"Atari-HEAD","full_name":null,"num_papers_in_archive":9},{"url":"/dataset/atari-grand-challenge","name":"Atari Grand Challenge","full_name":"","num_papers_in_archive":7},{"url":"/dataset/simitate","name":"Simitate","full_name":null,"num_papers_in_archive":5},{"url":"/dataset/minerl","name":"MineRL","full_name":"MineRL","num_papers_in_archive":3},{"url":"/dataset/motion-policy-networks","name":"Motion Policy Networks","full_name":"","num_papers_in_archive":3},{"url":"/dataset/il-datasets","name":"IL-Datasets","full_name":"Imitation Datasets","num_papers_in_archive":2},{"url":"/dataset/stardata","name":"StarData","full_name":"StarData","num_papers_in_archive":2},{"url":"/dataset/deformpam-dataset","name":"DeformPAM-Dataset","full_name":"Dataset of DeformPAM","num_papers_in_archive":1},{"url":"/dataset/reactive-diffusion-policy-dataset","name":"Reactive Diffusion Policy-Dataset","full_name":"Dataset of Reactive Diffusion Policy","num_papers_in_archive":1},{"url":"/dataset/rllab-framework","name":"RLLab Framework","full_name":"","num_papers_in_archive":1},{"url":"/dataset/texygen-platform","name":"Texygen Platform","full_name":"","num_papers_in_archive":1}],"subtasks":[{"url":"/task/behavioural-cloning","name":"Behavioural cloning"}],"parent_tasks":[{"url":"/task/decision-making","name":"Decision Making"}],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":691,"tagged_in_all":2122,"items":[{"url":"/paper/generative-adversarial-imitation-learning","title":"Generative Adversarial Imitation Learning","date":"2016-06-10","arxiv_id":"1606.03476","repositories_listed":18,"syntology":{"n":36,"n_ran":12,"n_unverified":24,"n_pointer_only":11}},{"url":"/paper/end-to-end-driving-via-conditional-imitation","title":"End-to-end Driving via Conditional Imitation Learning","date":"2017-10-06","arxiv_id":"1710.02410","repositories_listed":7,"syntology":{"n":3,"n_ran":0,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/exact-combinatorial-optimization-with-graph","title":"Exact Combinatorial Optimization with Graph Convolutional Neural Networks","date":"2019-06-04","arxiv_id":"1906.01629","repositories_listed":6,"syntology":{"n":2,"n_ran":0,"n_unverified":2,"n_pointer_only":1}},{"url":"/paper/behavioral-cloning-from-observation","title":"Behavioral Cloning from Observation","date":"2018-05-04","arxiv_id":"1805.01954","repositories_listed":6,"syntology":{"n":7,"n_ran":4,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/deep-q-learning-from-demonstrations","title":"Deep Q-learning from Demonstrations","date":"2017-04-12","arxiv_id":"1704.03732","repositories_listed":6,"syntology":null},{"url":"/paper/iq-learn-inverse-soft-q-learning-for","title":"IQ-Learn: Inverse soft-Q Learning for Imitation","date":"2021-06-23","arxiv_id":"2106.12142","repositories_listed":5,"syntology":{"n":3,"n_ran":2,"n_unverified":1,"n_pointer_only":3}},{"url":"/paper/sqil-imitation-learning-via-regularized","title":"SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards","date":"2019-05-27","arxiv_id":"1905.11108","repositories_listed":5,"syntology":null},{"url":"/paper/variational-discriminator-bottleneck","title":"Variational Discriminator Bottleneck: Improving Imitation Learning, Inverse RL, and GANs by Constraining Information Flow","date":"2018-10-01","arxiv_id":"1810.00821","repositories_listed":5,"syntology":{"n":13,"n_ran":0,"n_unverified":13,"n_pointer_only":0}},{"url":"/paper/comparing-the-efficacy-of-fine-tuning-and","title":"Comparing the Efficacy of Fine-Tuning and Meta-Learning for Few-Shot Policy Imitation","date":"2023-06-23","arxiv_id":"2306.13554","repositories_listed":4,"syntology":null},{"url":"/paper/orca-progressive-learning-from-complex","title":"Orca: Progressive Learning from Complex Explanation Traces of GPT-4","date":"2023-06-05","arxiv_id":"2306.02707","repositories_listed":4,"syntology":null},{"url":"/paper/a-gan-like-approach-for-physics-based","title":"A GAN-Like Approach for Physics-Based Imitation Learning and Interactive Character Control","date":"2021-05-21","arxiv_id":"2105.10066","repositories_listed":4,"syntology":null},{"url":"/paper/amp-adversarial-motion-priors-for-stylized","title":"AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control","date":"2021-04-05","arxiv_id":"2104.02180","repositories_listed":4,"syntology":null},{"url":"/paper/proof-artifact-co-training-for-theorem","title":"Proof Artifact Co-training for Theorem Proving with Language Models","date":"2021-02-11","arxiv_id":"2102.06203","repositories_listed":4,"syntology":{"n":4,"n_ran":3,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/chauffeurnet-learning-to-drive-by-imitating","title":"ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst","date":"2018-12-07","arxiv_id":"1812.03079","repositories_listed":4,"syntology":null},{"url":"/paper/self-imitation-learning","title":"Self-Imitation Learning","date":"2018-06-14","arxiv_id":"1806.05635","repositories_listed":4,"syntology":{"n":6,"n_ran":1,"n_unverified":5,"n_pointer_only":3}},{"url":"/paper/infogail-interpretable-imitation-learning","title":"InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations","date":"2017-03-26","arxiv_id":"1703.08840","repositories_listed":4,"syntology":null},{"url":"/paper/openvla-an-open-source-vision-language-action","title":"OpenVLA: An Open-Source Vision-Language-Action Model","date":"2024-06-13","arxiv_id":"2406.09246","repositories_listed":3,"syntology":{"n":10,"n_ran":2,"n_unverified":8,"n_pointer_only":0}},{"url":"/paper/webshop-towards-scalable-real-world-web","title":"WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents","date":"2022-07-04","arxiv_id":"2207.01206","repositories_listed":3,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/transfuser-imitation-with-transformer-based","title":"TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous Driving","date":"2022-05-31","arxiv_id":"2205.15997","repositories_listed":3,"syntology":{"n":5,"n_ran":1,"n_unverified":4,"n_pointer_only":0}},{"url":"/paper/end-to-end-urban-driving-by-imitating-a","title":"End-to-End Urban Driving by Imitating a Reinforcement Learning Coach","date":"2021-08-18","arxiv_id":"2108.08265","repositories_listed":3,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/of-moments-and-matching-trade-offs-and","title":"Of Moments and Matching: A Game-Theoretic Framework for Closing the Imitation Gap","date":"2021-03-04","arxiv_id":"2103.03236","repositories_listed":3,"syntology":null},{"url":"/paper/babyai-1-1","title":"BabyAI 1.1","date":"2020-07-24","arxiv_id":"2007.12770","repositories_listed":3,"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":0}},{"url":"/paper/adversarial-soft-advantage-fitting-imitation","title":"Adversarial Soft Advantage Fitting: Imitation Learning without Policy Optimization","date":"2020-06-23","arxiv_id":"2006.13258","repositories_listed":3,"syntology":null},{"url":"/paper/imitation-learning-via-off-policy-1","title":"Imitation Learning via Off-Policy Distribution Matching","date":"2019-12-10","arxiv_id":"1912.05032","repositories_listed":3,"syntology":null},{"url":"/paper/a-divergence-minimization-perspective-on","title":"A Divergence Minimization Perspective on Imitation Learning Methods","date":"2019-11-06","arxiv_id":"1911.02256","repositories_listed":3,"syntology":{"n":3,"n_ran":2,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/simitate-a-hybrid-imitation-learning","title":"Simitate: A Hybrid Imitation Learning Benchmark","date":"2019-05-15","arxiv_id":"1905.06002","repositories_listed":3,"syntology":null},{"url":"/paper/extrapolating-beyond-suboptimal","title":"Extrapolating Beyond Suboptimal Demonstrations via Inverse Reinforcement Learning from Observations","date":"2019-04-12","arxiv_id":"1904.06387","repositories_listed":3,"syntology":{"n":3,"n_ran":2,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/go-explore-a-new-approach-for-hard","title":"Go-Explore: a New Approach for Hard-Exploration Problems","date":"2019-01-30","arxiv_id":"1901.10995","repositories_listed":3,"syntology":{"n":4,"n_ran":2,"n_unverified":2,"n_pointer_only":4}},{"url":"/paper/compositional-imitation-learning-explaining","title":"CompILE: Compositional Imitation Learning and Execution","date":"2018-12-04","arxiv_id":"1812.01483","repositories_listed":3,"syntology":{"n":5,"n_ran":0,"n_unverified":5,"n_pointer_only":0}},{"url":"/paper/task-embedded-control-networks-for-few-shot","title":"Task-Embedded Control Networks for Few-Shot Imitation Learning","date":"2018-10-08","arxiv_id":"1810.03237","repositories_listed":3,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":1}}],"syntology_records":18,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}