Saved in:
Bibliographic Details
Main Authors: Guzelkabaagac, Jed, Petrović, Boris
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.13349
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916969128656896
author Guzelkabaagac, Jed
Petrović, Boris
author_facet Guzelkabaagac, Jed
Petrović, Boris
contents We study whether 3D self-supervised pretraining with Point--JEPA enables label-efficient grasp joint-angle prediction. Meshes are sampled to point clouds and tokenized; a ShapeNet-pretrained Point--JEPA encoder feeds a $K{=}5$ multi-hypothesis head trained with winner-takes-all and evaluated by top--logit selection. On a multi-finger hand dataset with strict object-level splits, Point--JEPA improves top--logit RMSE and Coverage@15$^{\circ}$ in low-label regimes (e.g., 26% lower RMSE at 25% data) and reaches parity at full supervision, suggesting JEPA-style pretraining is a practical lever for data-efficient grasp learning.
format Preprint
id arxiv_https___arxiv_org_abs_2509_13349
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Label-Efficient Grasp Joint Prediction with Point-JEPA
Guzelkabaagac, Jed
Petrović, Boris
Robotics
Artificial Intelligence
Machine Learning
We study whether 3D self-supervised pretraining with Point--JEPA enables label-efficient grasp joint-angle prediction. Meshes are sampled to point clouds and tokenized; a ShapeNet-pretrained Point--JEPA encoder feeds a $K{=}5$ multi-hypothesis head trained with winner-takes-all and evaluated by top--logit selection. On a multi-finger hand dataset with strict object-level splits, Point--JEPA improves top--logit RMSE and Coverage@15$^{\circ}$ in low-label regimes (e.g., 26% lower RMSE at 25% data) and reaches parity at full supervision, suggesting JEPA-style pretraining is a practical lever for data-efficient grasp learning.
title Label-Efficient Grasp Joint Prediction with Point-JEPA
topic Robotics
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2509.13349