Deep Object-Centric Representations for Generalizable Robot Learning

TL;DR

This paper introduces an object-centric perception method using pretrained visual models and attention mechanisms to improve generalization in robotic manipulation tasks with minimal samples.

Contribution

It presents a novel object-level attention framework that integrates pretrained visual models into robot learning, enabling better generalization and sample efficiency.

Findings

Good generalization across object instances with few samples

Effective learning of diverse manipulation tasks via reinforcement learning

Flexible adjustment of attention scope with different demonstrations

Abstract

Robotic manipulation in complex open-world scenarios requires both reliable physical manipulation skills and effective and generalizable perception. In this paper, we propose a method where general purpose pretrained visual models serve as an object-centric prior for the perception system of a learned policy. We devise an object-level attentional mechanism that can be used to determine relevant objects from a few trajectories or demonstrations, and then immediately incorporate those objects into a learned policy. A task-independent meta-attention locates possible objects in the scene, and a task-specific attention identifies which objects are predictive of the trajectories. The scope of the task-specific attention is easily adjusted by showing demonstrations with distractor objects or with diverse relevant objects. Our results indicate that this approach exhibits good generalization across object instances using very few samples, and can be used to learn a variety of manipulation tasks using reinforcement learning.