Data Analogies Enable Efficient Cross-Embodiment Transfer

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Yang, Jonathan, Finn, Chelsea, Sadigh, Dorsa
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910061125697536
author Yang, Jonathan
Finn, Chelsea
Sadigh, Dorsa
author_facet Yang, Jonathan
Finn, Chelsea
Sadigh, Dorsa
contents Generalist robot policies are trained on demonstrations collected across a wide variety of robots, scenes, and viewpoints. Yet it remains unclear how to best organize and scale such heterogeneous data so that it genuinely improves performance in a given target setting. In this work, we ask: what form of demonstration data is most useful for enabling transfer across robot set-ups? We conduct controlled experiments that vary end-effector morphology, robot platform appearance, and camera perspective, and compare the effects of simply scaling the number of demonstrations against systematically broadening the diversity in different ways. Our simulated experiments show that while perceptual shifts such as viewpoint benefit most from broad diversity, morphology shifts benefit far less from unstructured diversity and instead see the largest gains from data analogies, i.e. paired demonstrations that align scenes, tasks, and/or trajectories across different embodiments. Informed by the simulation results, we improve real-world cross-embodiment transfer success by an average of $22.5\%$ over large-scale, unpaired datasets by changing only the composition of the data.
format Preprint
id arxiv_https___arxiv_org_abs_2603_06450
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Data Analogies Enable Efficient Cross-Embodiment Transfer
Yang, Jonathan
Finn, Chelsea
Sadigh, Dorsa
Robotics
68T40
I.2.9
Generalist robot policies are trained on demonstrations collected across a wide variety of robots, scenes, and viewpoints. Yet it remains unclear how to best organize and scale such heterogeneous data so that it genuinely improves performance in a given target setting. In this work, we ask: what form of demonstration data is most useful for enabling transfer across robot set-ups? We conduct controlled experiments that vary end-effector morphology, robot platform appearance, and camera perspective, and compare the effects of simply scaling the number of demonstrations against systematically broadening the diversity in different ways. Our simulated experiments show that while perceptual shifts such as viewpoint benefit most from broad diversity, morphology shifts benefit far less from unstructured diversity and instead see the largest gains from data analogies, i.e. paired demonstrations that align scenes, tasks, and/or trajectories across different embodiments. Informed by the simulation results, we improve real-world cross-embodiment transfer success by an average of $22.5\%$ over large-scale, unpaired datasets by changing only the composition of the data.
title Data Analogies Enable Efficient Cross-Embodiment Transfer
topic Robotics
68T40
I.2.9
url https://arxiv.org/abs/2603.06450