A Data-Efficient Visual-Audio Representation with Intuitive Fine-tuning for Voice-Controlled Robots

TL;DR

This paper introduces a data-efficient, self-supervised visual-audio representation for voice-controlled robots that enables intuitive, continual self-improvement in new deployment domains with minimal labeled data.

Contribution

It proposes a novel representation based on contrastive learning that allows robots to self-improve in new environments without hand-crafted rewards or extensive labels.

Findings

Effective in various robotic tasks including navigation and manipulation

Achieves better performance with fewer labeled data in unseen scenarios

Works in both simulated and real-world experiments

Abstract

A command-following robot that serves people in everyday life must continually improve itself in deployment domains with minimal help from its end users, instead of engineers. Previous methods are either difficult to continuously improve after the deployment or require a large number of new labels during fine-tuning. Motivated by (self-)supervised contrastive learning, we propose a novel representation that generates an intrinsic reward function for command-following robot tasks by associating images with sound commands. After the robot is deployed in a new domain, the representation can be updated intuitively and data-efficiently by non-experts without any hand-crafted reward functions. We demonstrate our approach on various sound types and robotic tasks, including navigation and manipulation with raw sensor inputs. In simulated and real-world experiments, we show that our system can continually self-improve in previously unseen scenarios given fewer new labeled data, while still achieving better performance over previous methods.