single mention · not yet filed
Spatially grounded action reasoning
Spatially grounded action reasoning is a chain-of-thought approach for vision–language–action models in which each reasoning step is grounded directly in the scene.1 Each step can be decoded and visualized both on the image plane and within the 3D environment, combining depth perception tokens, visual reasoning traces, and action prediction.2
Trade-off The explicit spatial grounding improves explainability and enhances action prediction.3
How this definition was written
Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.
MolmoAct grounds every step of its reasoning chain directly in the scene
— MolmoAct Action Reasoning Models that can Reason in Spaceeach step can be decoded and visualized both on the image plane and within the 3D environment
— MolmoAct Action Reasoning Models that can Reason in SpaceThis explicit spatial grounding improves explainability and enhances action prediction
— MolmoAct Action Reasoning Models that can Reason in Space
Also called reasoning in space.