Is Diversity All You Need for Scalable Robotic Manipulation?

TL;DR

This paper investigates the impact of data diversity in robotic manipulation, revealing that task diversity is most critical, cross-embodiment transfer can be efficient with high-quality single-embodiment data, and expert diversity can introduce confounding factors.

Contribution

The study challenges the conventional belief that more diverse data is always better, providing new insights into data scaling principles for robotic manipulation.

Findings

Task diversity is more important than demonstration quantity.

Single-embodiment pre-training can transfer effectively across platforms.

Velocity multimodality from expert diversity can confound policy learning.

Abstract

Data scaling has driven remarkable success in foundation models for Natural Language Processing (NLP) and Computer Vision (CV), yet the principles of effective data scaling in robotic manipulation remain insufficiently understood. In this work, we investigate the nuanced role of data diversity in robot learning by examining three critical dimensions-task (what to do), embodiment (which robot to use), and expert (who demonstrates)-challenging the conventional intuition of "more diverse is better". Throughout extensive experiments on various robot platforms, we reveal that (1) task diversity proves more critical than per-task demonstration quantity, benefiting transfer from diverse pre-training tasks to novel downstream scenarios; (2) multi-embodiment pre-training data is optional for cross-embodiment transfer-models trained on high-quality single-embodiment data can efficiently transfer to different platforms, showing more desirable scaling property during fine-tuning than multi-embodiment pre-trained models; and (3) expert diversity, arising from individual operational preferences and stochastic variations in human demonstrations, can be confounding to policy learning, with velocity multimodality emerging as a key contributing factor. Based on this insight, we propose a distribution debiasing method to mitigate velocity ambiguity, the yielding GO-1-Pro achieves substantial performance gains of 15%, equivalent to using 2.5 times pre-training data. Collectively, these findings provide new perspectives and offer practical guidance on how to scale robotic manipulation datasets effectively.