CubeDAgger: Interactive Imitation Learning for Dynamic Systems with Efficient yet Low-risk Interaction
On the map: Offline RL and imitation learning from data → Imitation learning methods
Abstract (arXiv)
Interactive imitation learning makes an agent's control policy robust by stepwise supervisions from an expert. The recent algorithms mostly employ expert-agent switching systems to reduce the expert's burden by limitedly selecting the supervision timing. However, this approach is useful only for static tasks; in dynamic tasks, timing discrepancies cause abrupt changes in actions, losing the robot's dynamic stability. This paper therefore proposes a novel method, named CubeDAgger, which improves robustness with less dynamic stability violations even for dynamic tasks. The proposed method is designed on a baseline, EnsembleDAgger, with three improvements. The first adds a regularization to explicitly activate the threshold for deciding the supervision timing. The second transforms the expert-agent switching system to an optimal consensus system of multiple action candidates. Third, autoregressive colored noise is injected to the agent's actions for time-consistent exploration. These improvements are verified by simulations, showing that the trained policies are sufficiently robust while maintaining dynamic stability during interaction. Finally, real-robot scooping experiments with a human expert demonstrate that the proposed method can learn robust policies from scratch based on just 30 minutes of interaction. https://youtu.be/kBl3SCTnVEM
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Imitation learning methods, one of 13 directions in Offline RL and imitation learning from data. Builds on 16 of its 30 references that are in the map. First posted on arXiv 8 May 2025.
Builds on
16 of 30 references.
Closest ancestors · 7
Same direction, newest first
- 2023RLIF: Interactive Imitation Learning as Reinforcement Learning21 Nov
- 2022Bayesian Disturbance Injection: Robust Imitation Learning of Flexible Policies for Robot Manipulation7 Nov
- 2022Interactive Imitation Learning in Robotics: A Survey31 Oct
- 2021ThriftyDAgger: Budget-Aware Novelty and Risk Gating for Interactive Imitation Learning17 Sep
- 2019Uncertainty-Aware Data Aggregation for Deep Imitation Learning7 May
- 2018HG-DAgger: Interactive Imitation Learning with Human Experts5 Oct
- 2018EnsembleDAgger: A Bayesian Approach to Safe Imitation Learning22 Jul
Other references · 9
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
- 2016Deep Exploration via Bootstrapped DQNcited by 527
- 2017DART: Noise Injection for Robust Imitation Learningcited by 191
- 2021LazyDAgger: Reducing Context Switching in Interactive Imitation Learningcited by 51
- 2022Fleet-DAgger: Interactive Robot Fleet Learning with Scalable Human Supervisioncited by 30
- 2022AdaTerm: Adaptive T-Distribution Estimated Robust Moments for Noise-Robust Stochastic Gradient Optimizationcited by 11
- 2024Leveraging Demonstrator-perceived Precision for Safe Interactive Imitation Learning of Clearance-limited Taskscited by 5
- 2024Revisiting Experience Replayable Conditionscited by 3
- 2024LiRA: Light-Robust Adversary for Model-based Reinforcement Learning in Real Worldcited by 3
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025The Pitfalls of Imitation Learning when Actions are Continuous
- 2023Model-Based Runtime Monitoring with Interactive Imitation Learning
- 2022Imitation Learning by Estimating Expertise of Demonstrators
- 2025TubeDAgger: Reducing the Number of Expert Interventions with Stochastic Reach-Tubes
- 2024Fusion Dynamical Systems with Machine Learning in Imitation Learning: A Comprehensive Overview
- 2023Agnostic Interactive Imitation Learning: New Theory and Practical Algorithms
- 2024Stable-BC: Controlling Covariate Shift with Stable Behavior Cloning
- 2021Learning from Imperfect Demonstrations from Agents with Varying Dynamics