Back to Home

GAP: Geometric Anchor Pre-training for Data-Efficient Visuomotor Learning of Manipulation Tasks

Davide Buoso1, Andrea Protopapa1, Stefano Di Carlo1, Francesca Pistilli1, Giuseppe Averta1
Politecnico di Torino1
Project Page Placeholder
Paper (Coming Soon) arXiv (Coming Soon) Code (Coming Soon)
GAP teaser

GAP introduces a geometric anchor pre-training stage to improve data-efficient visuomotor learning for robotic manipulation.

Abstract

This page is currently a placeholder. The full abstract, method details, and experiments for GAP will be published soon.

Project Status

We are preparing the complete project page materials.

  • Paper links and supplementary material will be added after public release.
  • Additional visualizations and demos are in preparation.
  • Code release details will be shared here.

Cite us


@article{gap_buoso,
  title={GAP: Geometric Anchor Pre-training for Data-Efficient Visuomotor Learning of Manipulation Tasks},
  author={Buoso, Davide and Protopapa, Andrea and Di Carlo, Stefano and Pistilli, Francesca and Averta, Giuseppe},
  year={2026},
  publisher={Under Review}
}