RICL: Adding In-Context Adaptability to Pre-Trained Vision-Language-Action Models

TL;DR

This paper introduces RICL, a method to enable in-context learning in pre-trained vision-language-action models for robotics, allowing easy task adaptation with minimal demonstrations without retraining the entire model.

Contribution

The paper presents a finetuning recipe that injects in-context adaptability into VLA models, enabling effective task learning from few demonstrations without full parameter updates.

Findings

RICL enables in-context learning in VLA models with only 20 demonstrations.

RICL significantly improves performance on new manipulation tasks.

Parameter finetuning further enhances task performance.

Abstract

Multi-task ``vision-language-action'' (VLA) models have recently demonstrated increasing promise as generalist foundation models for robotics, achieving non-trivial performance out of the box on new tasks in new environments. However, for such models to be truly useful, an end user must have easy means to teach them to improve. For language and vision models, the emergent ability to perform in-context learning (ICL) has proven to be a versatile and highly useful interface to easily teach new tasks with no parameter finetuning. Unfortunately, VLAs pre-trained with imitation learning objectives do not naturally acquire ICL abilities. In this paper, we demonstrate that, with the right finetuning recipe and a small robot demonstration dataset, it is possible to inject in-context adaptability post hoc into such a VLA. After retraining for in-context learning (RICL), our system permits an end user to provide a small number (10-20) of demonstrations for a new task. RICL then fetches the most relevant portions of those demonstrations into the VLA context to exploit ICL, performing the new task and boosting task performance. We apply RICL to inject ICL into the $\pi_{0}$-FAST VLA, and show that it permits large in-context improvements for a variety of new manipulation tasks with only 20 demonstrations per task, without any parameter updates. When parameter updates on the target task demonstrations is possible, RICL finetuning further boosts performance. We release code and model weights for RICL-$\pi_{0}$-FAST alongside the paper to enable, for the first time, a simple in-context learning interface for new manipulation tasks. Website: https://ricl-vla.github.io.