SoftGPT: Learn Goal-oriented Soft Object Manipulation Skills by Generative Pre-trained Heterogeneous Graph Transformer

TL;DR

SoftGPT is a pre-trained model that enables robots to learn goal-oriented soft object manipulation skills efficiently by combining a heterogeneous graph representation with a GPT-based dynamics model, facilitating policy learning from human demonstrations.

Contribution

The paper introduces SoftGPT, a novel pre-trained model that integrates graph-based soft object representation with GPT dynamics for improved manipulation skill learning.

Findings

SoftGPT effectively learns various soft object manipulation skills.

The model can directly learn from human demonstrations.

Leveraging prior knowledge accelerates policy training.

Abstract

Soft object manipulation tasks in domestic scenes pose a significant challenge for existing robotic skill learning techniques due to their complex dynamics and variable shape characteristics. Since learning new manipulation skills from human demonstration is an effective way for robot applications, developing prior knowledge of the representation and dynamics of soft objects is necessary. In this regard, we propose a pre-trained soft object manipulation skill learning model, namely SoftGPT, that is trained using large amounts of exploration data, consisting of a three-dimensional heterogeneous graph representation and a GPT-based dynamics model. For each downstream task, a goal-oriented policy agent is trained to predict the subsequent actions, and SoftGPT generates the consequences of these actions. Integrating these two approaches establishes a thinking process in the robot's mind that provides rollout for facilitating policy learning. Our results demonstrate that leveraging prior knowledge through this thinking process can efficiently learn various soft object manipulation skills, with the potential for direct learning from human demonstrations.