mlsanityData to 7 Oct 2026

Out-of-Distribution Recovery with Object-Centric Keypoint Inverse Policy for Visuomotor Imitation Learning

arXiv 2411.03294 ↗First posted 5 Nov 2024cs.RO cs.AIcited by 6

On the map: Offline RL and imitation learning from data → Robot manipulation imitation learning

Abstract (arXiv)

We propose an object-centric recovery (OCR) framework to address the challenges of out-of-distribution (OOD) scenarios in visuomotor policy learning. Previous behavior cloning (BC) methods rely heavily on a large amount of labeled data coverage, failing in unfamiliar spatial states. Without relying on extra data collection, our approach learns a recovery policy constructed by an inverse policy inferred from the object keypoint manifold gradient in the original training data. The recovery policy serves as a simple add-on to any base visuomotor BC policy, agnostic to a specific method, guiding the system back towards the training distribution to ensure task success even in OOD situations. We demonstrate the effectiveness of our object-centric framework in both simulation and real robot experiments, achieving an improvement of 77.7% over the base policy in OOD. Furthermore, we show OCR's capacity to autonomously collect demonstrations for continual learning. Overall, we believe this framework represents a step toward improving the robustness of visuomotor policies in real-world settings.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. Builds on 32 of its 44 references that are in the map. First posted on arXiv 5 Nov 2024.

Builds on

32 of 44 references.

Closest ancestors · 8

Same direction, newest first

  1. 2024Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation4 Jan
  2. 2023Any-point Trajectory Modeling for Policy Learning28 Dec
  3. 2023Waypoint-Based Imitation Learning for Robotic Manipulation26 Jul
  4. 2023Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware23 Apr
  5. 2023Teach a Robot to FISH: Versatile Imitation from One Minute of Demonstrations2 Mar
  6. 2022VIOLA: Imitation Learning for Vision-Based Manipulation with Object Proposal Priors20 Oct
  7. 2022K-VIL: Keypoints-based Visual Imitation Learning7 Sep
  8. 2021The Surprising Effectiveness of Representation Learning for Visual Imitation2 Dec

Other references · 24

  1. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,730
  2. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemscited by 1,873
  3. 2020Conservative Q-Learning for Offline Reinforcement Learningcited by 1,777
  4. 2021Offline Reinforcement Learning with Implicit Q-Learningcited by 1,245
  5. 2024Grounded SAM: Assembling Open-World Models for Diverse Visual Taskscited by 914
  6. 2019AugMix: A Simple Data Processing Method to Improve Robustness and Uncertaintycited by 897
  7. 2021A Minimalist Approach to Offline Reinforcement Learningcited by 882
  8. 2021What Matters in Learning from Offline Human Demonstrations for Robot Manipulationcited by 699
  9. 2018An Algorithmic Perspective on Imitation Learningcited by 455
  10. 2021Implicit Behavioral Cloningcited by 442
  11. 2023FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objectscited by 411
  12. 2021COMBO: Conservative Offline Model-Based Policy Optimizationcited by 365
12 more
  1. 2018HG-DAgger: Interactive Imitation Learning with Human Expertscited by 245
  2. 2024Flow as the Cross-Domain Manipulation Interfacecited by 126
  3. 2020Learning Dexterous Grasping with Object-Centric Visual Affordancescited by 80
  4. 2024From Imitation to Refinement -- Residual RL for Precise Assemblycited by 80
  5. 2023On the Generalization Properties of Diffusion Modelscited by 67
  6. 2017Deep Object-Centric Representations for Generalizable Robot Learningcited by 50
  7. 2024RACER: Rich Language-Guided Failure Recovery Policies for Imitation Learningcited by 46
  8. 2023ZS6D: Zero-shot 6D Object Pose Estimation using Vision Transformerscited by 45
  9. 2021An Information-theoretic Approach to Distribution Shiftscited by 25
  10. 2022Back to the Manifold: Recovering from Out-of-Distribution Statescited by 19
  11. 2023Shape-Constraint Recurrent Flow for 6D Object Pose Estimationcited by 17
  12. 2024Recover: A Neuro-Symbolic Framework for Failure Detection and Recoverycited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 3; 2026: 3.

6 papers.

  1. 2026VLA-Corrector: Lightweight Detect-and-Correct Inference for Adaptive Action Horizoncited by 14
  2. 2025Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learningcited by 5
  3. 2025Prepare Before You Act: Learning From Humans to Rearrange Initial Statescited by 3
  4. 2026Dreaming the Unseen: World Model-regularized Diffusion Policy for Out-of-Distribution Robustnesscited by 2
  5. 2026Exploration-assisted Bottleneck Transition Toward Robust and Data-efficient Deformable Object Manipulationcited by 1
  6. 2025Wavelet Policy: Imitation Learning in the Scale Domain with World Prior Memorycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Adapting by Analogy: OOD Generalization of Visuomotor Policies via Functional Correspondence
  2. 2025Latent Policy Barrier: Learning Robust Visuomotor Policies by Staying In-Distribution
  3. 2025SViP: Sequencing Bimanual Visuomotor Policies with Object-Centric Motion Primitives
  4. 2021Introspective Visuomotor Control: Exploiting Uncertainty in Deep Visuomotor Control for Failure Recovery
  5. 2026More Structure, Not More Capacity: Object-Centric Representations for Visuomotor Imitation Learning
  6. 2019Self-Supervised Correspondence in Visuomotor Policy Learning
  7. 2026It's Not Just More Demos: Counterfactual Action Sensitivity Coverage for Data-Efficient Robust Robot Imitation
  8. 2024Contractive Dynamical Imitation Policies for Efficient Out-of-Sample Recovery