LOTUS: Continual Imitation Learning for Robot Manipulation Through Unsupervised Skill Discovery

arXiv:2311.02058v3 Announce Type: replace-cross
Abstract: We introduce LOTUS, a continual imitation learning algorithm that empowers a physical robot to continuously and efficiently learn to solve new manipulation tasks throughout its lifespan. The core idea behind LOTUS is constructing an ever-growing skill library from a sequence of new tasks with a small number of human demonstrations. LOTUS starts with a continual skill discovery process using an open-vocabulary vision model, which extracts skills as recurring patterns presented in unsegmented demonstrations. Continual skill …

arxiv continual cs.cv cs.lg cs.ro discovery imitation learning lotus manipulation robot robot manipulation through type unsupervised