Behavior predictive control offers fast policy from demonstrations
Training-free Behavior Cloning
Robotics
Summary
Making robots or AI systems learn from examples usually requires complex training that takes a lot of time and hides how decisions are made. The authors propose Behavior Predictive Control (BPC), a way to generate policies without this slow training, by directly using stored examples to predict future actions. This allows the system to act quickly and understand which past examples guided current decisions. BPC matches or beats other learned methods in tests and works well even on small, fast devices.
What this means in practice
- •For robotics engineers: Deploy learned behaviors on robots quickly without lengthy retraining, enabling faster experimentation and adaptation.
- •For embedded systems developers: Implement real-time control policies on resource-limited devices using fast, training-free methods from demonstration data.
Authors
Maximilian Adang, Timothy Chen, Lars Osterberg, Aiden Swann, Mac Schwager
Abstract
Neural behavior cloning compresses demonstrations into large models, making individual actions difficult to trace and policy updates costly. Retrieval policies retain access to demonstrations but struggle with mismatch between recorded and live behavior. We introduce Behavior Predictive Control (BPC), which synthesizes policies without end-to-end policy training by combining an action-aware retrieval metric, a Hankel-based action-continuation prior, and a closed-form one-step residual correction. Inspired by behavioral systems theory, BPC predicts future actions by blending stored observation-action data that best reconstructs the recent runtime observation--action history. Across simulated benchmarks and real-robot deployments, BPC is competitive with learned policies such as $π_{0.5}$ (surpassing it in some cases), while reducing policy fitting from hours to seconds on consumer GPUs and supporting closed-loop control upwards of 75 Hz on a Jetson Orin Nano. The retrieved demonstration windows and their coefficients also provide an intrinsic estimate of task progress. Retaining demonstrations within the deployed policy makes its predictions traceable to supporting trajectories and enables behavior revision through the demonstration bank.