Over its lifetime, a reinforcement learning agent is often tasked with different tasks. How to efficiently adapt a previously learned control policy from one task to another, remains an open research question. In this paper, we investigate how instructions formulated in natural language can enable faster and more effective task adaptation. This can serve as the basis for developing language instructed skills, which can be used in a lifelong learning setting. Our method is capable of assessing, given a set of developed base control policies, which policy will adapt best to a new unseen task.

The psychological state of flow has been linked to optimizing human performance. A key condition of flow emergence is a match between the human abilities and complexity of the task. We propose a simple computational model of flow for Artificial Intelligence (AI) agents. The model factors the standard agent-environment state into a self-reflective set of the agent's abilities and a socially learned set of the environmental complexity. Maximizing the flow serves as a meta control for the agent. We show how to apply the meta-control policy to a broad class of AI control policies and illustrate our approach with a specific implementation. Results in a synthetic testbed are promising and open interesting directions for future work.