An Atomic Skill Library Construction Method for Data-Efficient Embodied Manipulation

TL;DR

This paper presents a dynamic, data-driven method for constructing an atomic skill library to improve data efficiency and adaptability in embodied manipulation tasks, using vision-language planning and fine-tuning.

Contribution

It introduces a three-wheeled approach to build and expand an atomic skill library, enabling flexible task decomposition and reducing data requirements in embodied AI manipulation.

Findings

Reduces data costs significantly compared to end-to-end methods.

Maintains high task success rates with an expanding skill library.

Demonstrates effectiveness in real-world experiments.

Abstract

Embodied manipulation is a fundamental ability in the realm of embodied artificial intelligence. Although current embodied manipulation models show certain generalizations in specific settings, they struggle in new environments and tasks due to the complexity and diversity of real-world scenarios. The traditional end-to-end data collection and training manner leads to significant data demands. Decomposing end-to-end tasks into atomic skills helps reduce data requirements and improves the task success rate. However, existing methods are limited by predefined skill sets that cannot be dynamically updated. To address the issue, we introduce a three-wheeled data-driven method to build an atomic skill library. We divide tasks into subtasks using the Vision-Language-Planning (VLP). Then, atomic skill definitions are formed by abstracting the subtasks. Finally, an atomic skill library is constructed via data collection and Vision-Language-Action (VLA) fine-tuning. As the atomic skill library expands dynamically with the three-wheel update strategy, the range of tasks it can cover grows naturally. In this way, our method shifts focus from end-to-end tasks to atomic skills, significantly reducing data costs while maintaining high performance and enabling efficient adaptation to new tasks. Extensive experiments in real-world settings demonstrate the effectiveness and efficiency of our approach.