AutoIntervene: Calibrated Intervention for Action-Chunking Imitation Learning Policies
On the map: Offline RL and imitation learning from data → Robot manipulation imitation learning
Abstract (arXiv)
Action-chunking visuomotor policies learn from demonstrations and improve temporal consistency by predicting short action sequences rather than single-step commands. Yet perception errors and execution drift can move the robot outside the demonstration distribution, while the policy continues to produce smooth action chunks that are inconsistent with the observed state. We present AutoIntervene, an online framework that selectively transfers control between an action-chunking policy and an operator during deployment. AutoIntervene evaluates proposed chunks against a visual-action support memory built from successful task executions, combining visual similarity with consistency between proposed and reference actions. Phase-local support governs policy-to-operator transfer within the current task phase, whereas global support governs the return to policy control after operator recovery. We calibrate separate switching thresholds for the two directions from empirical quantiles of evaluation-level scores on held-out expert demonstrations, avoiding direct manual tuning of score cutoffs. Intervention segments retained from successful rollouts target learner-induced states and provide corrective supervision for subsequent policy updates. Experiments on real-world bimanual manipulation tasks show higher post-adaptation task success and lower operator-control time than manual intervention. Videos and additional results are available at https://aus.bot/research/autointervene/.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. Builds on 24 of its 28 references that are in the map. First posted on arXiv 7 Aug 2026.
Builds on
24 of 28 references.
Closest ancestors · 6
Same direction, newest first
- 2024FlowPolicy: Enabling Fast and Robust 3D Flow-based Policy via Consistency Flow Matching for Robot Manipulation6 Dec
- 2023Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware23 Apr
- 2021Error-Aware Imitation Learning from Teleoperation Data for Mobile Manipulation9 Dec
- 2021LazyDAgger: Reducing Context Switching in Interactive Imitation Learning31 Mar
- 2020Human-in-the-Loop Imitation Learning using Remote Teleoperation12 Dec
- 2020Learning to Generalize Across Long-Horizon Tasks from Human Demonstrations13 Mar
Other references · 18
- 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,838
- 2025DINOv3cited by 1,414
- 2018Experience Replay for Continual Learningcited by 1,095
- 2021What Matters in Learning from Offline Human Demonstrations for Robot Manipulationcited by 725
- 2017Time-Contrastive Networks: Self-Supervised Learning from Videocited by 459
- 2018An Algorithmic Perspective on Imitation Learningcited by 459
- 2018HG-DAgger: Interactive Imitation Learning with Human Expertscited by 266
- 2017DART: Noise Injection for Robust Imitation Learningcited by 191
- 2024Unpacking Failure Modes of Generative Policies: Runtime Monitoring of Consistency and Progresscited by 69
- 2025Can We Detect Failures Without Failure Data? Uncertainty-Aware Runtime Failure Detection for Imitation Learning Policiescited by 52
- 2025RaC: Robot Learning for Long-Horizon Tasks by Scaling Recovery and Correctioncited by 39
- 2026Rewind-IL: Online Failure Detection and State Respawning for Imitation Learningcited by 9
6 more
- 2024Efficient Active Imitation Learning with Random Network Distillationcited by 7
- 2025ConditionNET: Learning Preconditions and Effects for Execution Monitoringcited by 7
- 2026TriPilot-FF: Coordinated Whole-Body Teleoperation with Force Feedbackcited by 5
- 2025DOSE3 : Diffusion-based Out-of-distribution detection on SE(3) trajectoriescited by 3
- 2026PATCH: Action-Chunk-Conditioned Latent Patch Innovation Monitoring for Robot Manipulationcited by 2
- 2026TRACE: Trajectory-Routed Causal Memory for Delayed-Evidence Visuomotor Imitationcited by 1
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2026Predict in Chunks, Revise in Spans: History-conditioned Action Regeneration for Robotic Manipulation
- 2025Action Chunking and Exploratory Data Collection Yield Exponential Improvements in Behavior Cloning for Continuous Control
- 2025SViP: Sequencing Bimanual Visuomotor Policies with Object-Centric Motion Primitives
- 2026TRACT: Temporally Routed Action Chunks with Chronological Phase Authority for Contact-Rich Manipulation
- 2026BlenDAgger: Blended Shared Control for Interactive Imitation Learning
- 2026WorldToken: Time-First Sequence Modeling for Robotic Imitation Learning
- 2026Revisiting Open-Loop Execution in Robotics: Toward Reactive, Higher-Performing Policies
- 2025CDP: Towards Robust Autoregressive Visuomotor Policy Learning via Causal Diffusion