ConBaT: Control Barrier Transformer for Safe Policy Learning

*Massachusetts Institute of Technology
Microsoft Research
Scaled Foundations
TODO: alt text

(Left) An agent trained to imitate expert demonstrations may focus on the result of the task without explicit notions of safety. (Right) Our method ConBaT learns a safety critic on top of the control policy and uses this control barrier critic to optimize the policy for safe actions actively.

TODO: alt text

ConBaT architecture - a causal Transformer opearates on state and action tokens with a critic to tell safety.



Large-scale self-supervised models have recently revolutionized our ability to perform a variety of tasks within the vision and language domains. However, using such models for autonomous systems is challenging because of safety requirements: besides executing correct actions, an autonomous agent must also avoid the high cost and potentially fatal critical mistakes. Traditionally, self-supervised training mainly focuses on imitating previously observed behaviors, and the training demonstrations carry no notion of which behaviors should be explicitly avoided. In this work, we propose Control Barrier Transformer (ConBaT), an approach that learns safe behaviors from demonstrations in a self-supervised fashion. ConBaT is inspired by the concept of control barrier functions in control theory and uses a causal transformer that learns to predict safe robot actions autoregressively using a critic that requires minimal safety data labeling. During deployment, we employ a lightweight online optimization to find actions that ensure future states lie within the learned safe set. We apply our approach to different simulated control tasks and show that our method results in safer control policies compared to other classical and learning-based methods such as imitation learning, reinforcement learning, and model predictive control.

TODO: alt text

Experimental results. Compared to other baseline methods, our method ConBaT achieves the lowest collision rate.

More content coming soon!