Towards Practical Multi-Object Manipulation using Relational Reinforcement Learning

TL;DR

This paper introduces a graph-based relational reinforcement learning framework that significantly improves data efficiency and generalization in multi-object manipulation tasks, demonstrated by a block stacking example.

Contribution

The authors propose a novel relational architecture that enables effective curriculum learning for complex multi-object tasks, outperforming existing methods in data efficiency and zero-shot generalization.

Findings

Achieved stacking of six blocks from scratch using sparse rewards.

Outperformed state-of-the-art methods with human demonstrations.

Demonstrated zero-shot generalization to taller towers and new configurations.

Abstract

Learning robotic manipulation tasks using reinforcement learning with sparse rewards is currently impractical due to the outrageous data requirements. Many practical tasks require manipulation of multiple objects, and the complexity of such tasks increases with the number of objects. Learning from a curriculum of increasingly complex tasks appears to be a natural solution, but unfortunately, does not work for many scenarios. We hypothesize that the inability of the state-of-the-art algorithms to effectively utilize a task curriculum stems from the absence of inductive biases for transferring knowledge from simpler to complex tasks. We show that graph-based relational architectures overcome this limitation and enable learning of complex tasks when provided with a simple curriculum of tasks with increasing numbers of objects. We demonstrate the utility of our framework on a simulated block stacking task. Starting from scratch, our agent learns to stack six blocks into a tower. Despite using step-wise sparse rewards, our method is orders of magnitude more data-efficient and outperforms the existing state-of-the-art method that utilizes human demonstrations. Furthermore, the learned policy exhibits zero-shot generalization, successfully stacking blocks into taller towers and previously unseen configurations such as pyramids, without any further training.