For robots to be useful outside labs and specialized factories we need a way to teach them new useful behaviors quickly. Current approaches lack either the generality to onboard new tasks without task-specific engineering, or else lack the data-efficiency to do so in an amount of time that enables practical use. In this work we explore dense tracking as a representational vehicle to allow faster and more general learning from demonstration. Our approach utilizes Track-Any-Point (TAP) models to isolate the relevant motion in a demonstration, and parameterize a low-level controller to reproduce this motion across changes in the scene configuration. We show this results in robust robot policies that can solve complex object-arrangement tasks such as shape-matching, stacking, and even full path-following tasks such as applying glue and sticking objects together, all from demonstrations that can be collected in minutes.
@article{arxiv.2308.15975,
title = {RoboTAP: Tracking Arbitrary Points for Few-Shot Visual Imitation},
author = {Mel Vecerik and Carl Doersch and Yi Yang and Todor Davchev and Yusuf Aytar and Guangyao Zhou and Raia Hadsell and Lourdes Agapito and Jon Scholz},
journal= {arXiv preprint arXiv:2308.15975},
year = {2023}
}