REA: A Routed Embodied Agent Runtime for Asynchronous Brain-Cerebellum Control
REA has been announced to asynchronously coordinate high-level reasoning via foundation models with reactive control via fast policies like VLAs.
Foundation models are becoming remarkably reliable at high-level reasoning, but they remain too slow and costly to serve as fine-grained controllers. Conversely, fast policies such as VLAs and whole-body controllers offer more efficient reflexes but struggle to generalize.
Pairing a deliberative brain (foundation models) with a reactive cerebellum (fast policies) seems natural for embodied systems. However, two obstacles remain: methods are built and scored on incompatible stacks, and their asynchronous coordination is difficult.
This research received 11 upvotes on Hugging Face Daily Papers.
This item is a summary. The full, checkable version is at the original source:
Read the original at Hugging Face Papers ↗