Produce a scene-grounded manipulation program
Turn a language instruction and perceived tabletop scene into symbolic manipulation steps and target sub-goals.
Desired outcome
A grounded program and sequence of object-and-pose sub-goals for a separate motion planner.
Explore from the outcome
Where this function can lead
These documented connections are starting points for selection, not a ranked list.
Model releases
Identifiable models documented for this function.
- Neuro-symbolic Robot Manipulation (ICRA 2023)
Research hybrid that grounds a natural-language instruction in an initial depth-image scene to produce a symbolic manipulation program and pose sub-goals.
Link evidence: Learning Neuro-symbolic Programs for Language Guided Robot Manipulation, Learning Neuro-symbolic Programs for Language Guided Robot Manipulation project page
Typical inputs
- Natural-language manipulation instruction
- Initial tabletop scene captured with depth information
Typical outputs
- Symbolic manipulation program grounded to scene objects
- Object-and-target-pose sub-goals for a separate low-level planner
Sources
Last reviewed 2026-09-30