The Pitfalls of Imitation Learning when Actions are Continuous
This is experimental HTML to improve accessibility. We invite you to report rendering errors. Use Alt+Y to toggle on accessible reporting links and Alt+Shift+Y to toggle off. Learn more about this project and help improve conversions.
The Pitfalls of Imitation Learning when Actions are Continuous The Pitfalls of Imitation Learning when Actions are Continuous Max Simchowitz 1 1 1 msimchow@andrew.cmu.edu CMU Daniel Pfrommer 2 2 2 dpfrom@mit.edu MIT Ali Jadbabaie 3 3 3 jadbabai@mit.edu MIT (July 26, 2025) Abstract We study the problem of imitating an expert demonstrator in a discrete-time, continuous state-and-action control system. We show that, even if the dynamics satisfy a control-theoretic property called exponential stability (i.e. the effects of perturbations decay exponentially quickly), and the expert is smooth and de
related reading
- Ch. 21 - Imitation Learningunderactuated.mit.edu
- [2507.09061] Action Chunking and Exploratory Data Collection Yield Exponential Improvements in Behavior Cloning for Continuous Controlarxiv.org
- 1011.0686arxiv.org
- State of Robot Learning, December 2025vedder.io
- Reinforcement Learning via Implicit Imitation Guidancearxiv.org
- You can’t imitation-learn how to continual-learn — LessWronglesswrong.com
- [2606.09758] Difference-Aware Retrieval Policies for Imitation Learningarxiv.org
- Flow Matching Policy Gradientsflowreinforce.github.io
- Generative Adversarial Imitation Learning.pdfcs.stanford.edu
- Imitating Latent Policies from Observationarxiv.org
- Imitation Bootstrapped Reinforcement Learningarxiv.org
- Reinforcement Learning in Newcomblike Problemsproceedings.neurips.cc