Fetching the paper…

TarGF: Learning Target Gradient Field to Rearrange Objects without Explicit Goal Specification · Around