Intrinsically Motivated Open-Ended Multi-Task Learning Using Transfer Learning to Discover Task Hierarchy

TL;DR

This paper presents a hierarchical reinforcement learning approach that uses transfer learning, intrinsic motivation, and task decomposition to enable robots to learn complex tasks efficiently through online self-organized learning and knowledge transfer.

Contribution

It introduces a novel task-oriented representation called procedures and combines goal-babbling, imitation, and active learning for self-organized multi-task learning.

Findings

Effective transfer of task decomposition across different learners.

Robust learning of complex tasks with minimal demonstrations.

Successful application on both simulation and real robot.

Abstract

In open-ended continuous environments, robots need to learn multiple parameterised control tasks in hierarchical reinforcement learning. We hypothesise that the most complex tasks can be learned more easily by transferring knowledge from simpler tasks, and faster by adapting the complexity of the actions to the task. We propose a task-oriented representation of complex actions, called procedures, to learn online task relationships and unbounded sequences of action primitives to control the different observables of the environment. Combining both goal-babbling with imitation learning, and active learning with transfer of knowledge based on intrinsic motivation, our algorithm self-organises its learning process. It chooses at any given time a task to focus on; and what, how, when and from whom to transfer knowledge. We show with a simulation and a real industrial robot arm, in cross-task and cross-learner transfer settings, that task composition is key to tackle highly complex tasks. Task decomposition is also efficiently transferred across different embodied learners and by active imitation, where the robot requests just a small amount of demonstrations and the adequate type of information. The robot learns and exploits task dependencies so as to learn tasks of every complexity.