Article

AFFORD2ACT studies the capability ceiling of robot manipulation policies with sparse 2D keypoint input

AFFORD2ACT (Affordance-Guided Keypoints to Act) studies how much manipulation capability can be retained when only sparse 2D keypoints are used as input: an off-the-shelf language-conditioned affordance model locates task-relevant regions, selects up to 15 object keypoints plus 4 end-effector keypoints, compresses the policy input to 19 2D keypoints (38 coordinates), and then uses lightweight gating to decide which keypoints to focus on during execution.

Part of the event: AFFORD2ACT: Sparse 2D Keypoints Drive Robot Manipulation Policies

This page provides a summary and source link only.