Task-Embedded Control Networks for Few-Shot Imitation Learning

TL;DR

This paper introduces Task-Embedded Control Networks that leverage metric learning to enable robots to learn new tasks from minimal demonstrations, demonstrating superior performance in simulation and real-world deployment.

Contribution

The paper presents a novel task embedding approach for few-shot imitation learning, scalable to many tasks and effective in real-world applications.

Findings

Outperforms state-of-the-art in simulation with visual data

Enables real-world task learning from a single demonstration

Effective with domain randomization for sim-to-real transfer

Abstract

Much like humans, robots should have the ability to leverage knowledge from previously learned tasks in order to learn new tasks quickly in new and unfamiliar environments. Despite this, most robot learning approaches have focused on learning a single task, from scratch, with a limited notion of generalisation, and no way of leveraging the knowledge to learn other tasks more efficiently. One possible solution is meta-learning, but many of the related approaches are limited in their ability to scale to a large number of tasks and to learn further tasks without forgetting previously learned ones. With this in mind, we introduce Task-Embedded Control Networks, which employ ideas from metric learning in order to create a task embedding that can be used by a robot to learn new tasks from one or more demonstrations. In the area of visually-guided manipulation, we present simulation results in which we surpass the performance of a state-of-the-art method when using only visual information from each demonstration. Additionally, we demonstrate that our approach can also be used in conjunction with domain randomisation to train our few-shot learning ability in simulation and then deploy in the real world without any additional training. Once deployed, the robot can learn new tasks from a single real-world demonstration.