Task-Embedded Control Networks for Few-Shot Imitation Learning
This addresses the challenge of few-shot imitation learning for robots, allowing them to generalize across tasks without forgetting, though it is incremental as it builds on meta-learning and metric learning ideas.
The paper tackles the problem of enabling robots to learn new tasks quickly from few demonstrations by introducing Task-Embedded Control Networks, which use metric learning to create task embeddings, resulting in surpassing state-of-the-art performance in simulation and enabling real-world deployment with single demonstrations.
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.